English(EN)Vision-language models for chest radiography do not always need the image
新研究增强了用于医疗、检索和机器人任务的视觉-语言模型
作者PulseAugur 编辑部·[422 个来源]·
研究人员正在开发新方法来改进各种领域的视觉-语言模型（VLM）。一篇论文介绍了CoT-Mediate，一个用于评估生成推理如何影响VLM在医疗情境下预测的框架，发现推理的位置比其声明的来源更关键。另一项研究提出了ReToken，一个可学习的单一嵌入，通过选择相关的视觉标记来增强视觉检索，在Visual Haystacks和LVBench等基准测试中显示出显著的提升。对于机器人领域，BioVLN是一个专为生物医学实验室视觉-语言导航设计的新模拟平台，解决了精确仪器交互的需求。此外，还提出了一种名为SeGP-CL的方法用于VLM的持续学习，旨在保持跨模态语义几何并对抗灾难性遗忘。最后，S-GRPO提供了一个用于大型VLM的统一的训练后框架，整合了监督微调和强化学习，以提高适应性并保持通用能力。
AI
arXiv:2608.10198v1 Announce Type: new Abstract: Latent-space communication allows heterogeneous vision-language model agents to exchange continuous representations without serializing visual and reasoning states into text. Vision Wormhole realizes this approach by translating vis…
arXiv cs.AI
TIER_1English(EN)·Meiwen Ding, Song Xia, Chenqi Kong, Xudong Jiang·
arXiv:2603.29418v2 Announce Type: replace-cross Abstract: Although multimodal large language models (MLLMs) are increasingly deployed in real-world applications, their instruction-following behavior leaves them vulnerable to prompt injection attacks. Existing prompt injection met…
arXiv:2608.10408v1 Announce Type: new Abstract: Vision-language models (VLMs) have shown strong capabilities in generating visualization code from textual or visual specifications. However, real-world visualization authoring is inherently iterative: users frequently revise existi…
arXiv:2608.11167v1 Announce Type: cross Abstract: Existing Multimodal Large Language Models (MLLMs) predominantly rely on image-text pairs for modality alignment pretraining, mapping global image representations to long textual descriptions. However, this image-level alignment su…
arXiv cs.LG
TIER_1English(EN)·Binesh Sadanandan, Vahid Behzadan·
arXiv:2604.08941v2 Announce Type: replace Abstract: Medical Vision-Language Models (VLMs) answering binary presence questions on chest radiographs can fail in two linked ways: they are confidently wrong, and they change answers when a clinically equivalent question is rephrased. …
arXiv:2608.10635v1 Announce Type: cross Abstract: Medical Vision-Language Models (Med-VLMs) excel at verbalizing visual content, yet precise visual perception, segmentation, and grounding remain challenging. Existing approaches either verbalize regions as coordinate strings or re…
arXiv:2608.10484v1 Announce Type: cross Abstract: Action verbs describe not only the physical outcomes of actions, but also how those actions are performed. Yet action representations in vision-language-action models (VLAs) are typically optimized for reconstruction under L1/L2 l…
arXiv cs.AI
TIER_1English(EN)·Foundation Model Team, XPeng Inc·
arXiv:2608.10976v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models can connect scene understanding, semantic reasoning, and trajectory generation for autonomous driving. However, verbose natural-language Chain-of-Thought (CoT) is poorly suited to real-time contro…
arXiv cs.AI
TIER_1English(EN)·Gesina Schwalbe, Mert Keser, Moritz Bayerkuhnlein, Edgar Heinert, Annika M\"utze, Marvin Keller, Sparsh Tiwari, Georgii Mikriukov, Diedrich Wolter, Jae Hee Lee, Matthias Rottmann·
arXiv:2603.26798v2 Announce Type: replace-cross Abstract: Vision-language model (VLM) encoders such as CLIP enable strong retrieval and zero-shot classification in a shared image-text embedding space, yet the semantic organization of this space is rarely inspected. We present a p…
arXiv:2608.09432v1 Announce Type: cross Abstract: Transformer-based language models rely on self-attention, whose computation is permutation-equivariant and therefore lacks an intrinsic mechanism for representing token order. Existing architectures address this limitation by expl…
arXiv cs.AI
TIER_1English(EN)·Jieyu Zhang, Ziqi Gao, Luke Zettlemoyer, Ranjay Krishna·
arXiv:2608.07886v1 Announce Type: cross Abstract: Vision-language grounding connects language to visual content, yet most existing formulations reduce grounding to a unidirectional localization problem: given a prespecified text phrase or category name, identify the corresponding…
arXiv:2608.09011v1 Announce Type: new Abstract: Uncertainty Quantification (UQ) aims to measure the reliability of model predictions, serving as a critical safeguard for deploying Vision-Language Models (VLMs) in safety-critical scenarios. Post-hoc approaches are widely adopted d…
arXiv cs.CL
TIER_1English(EN)·Zhanna Mukhametsharip (Saarland University, Germany), Vera Demberg (Saarland University, Germany, Max Planck Institute for Informatics, Germany), Varsha Suresh (Max Planck Institute for Informatics, Germany)·
arXiv:2608.09772v1 Announce Type: new Abstract: Large Vision-Language Models (LVLMs) have demonstrated strong performance on multimodal benchmarks, yet it remains unclear whether they genuinely reason about relationships between images and text or rely on superficial correlations…
Vision-language models (VLMs) have shown strong capabilities in generating visualization code from textual or visual specifications. However, real-world visualization authoring is inherently iterative: users frequently revise existing visualizations to repair flawed charts or ada…
arXiv:2608.07427v1 Announce Type: new Abstract: LLM inference accounts for over 90% of AI operational energy, scaling directly with input token count---a critical inefficiency for telecom network analytics and numerical time-series data analysis (NTSDA), where raw multivariate KP…
arXiv cs.AI
TIER_1English(EN)·Nikos Theodoridis, Reenu Mohandas, Ganesh Sistu, Anthony Scanlan, Ciar\'an Eising, Tim Brophy·
arXiv:2603.06054v2 Announce Type: replace-cross Abstract: The use of Vision-Language Models (VLMs) in automated driving applications is becoming increasingly common, with the aim of leveraging their reasoning and generalisation capabilities to handle long-tail scenarios. However,…
arXiv:2608.06756v1 Announce Type: new Abstract: Vision-language models are increasingly serving as the reasoning core of embodied agents. Robot execution is inherently iterative: each action reshapes the scene and physical state, continually renewing what must be perceived, reaso…
arXiv cs.LG
TIER_1English(EN)·Minseok Kang, Hyunwoo Kim, Chanyoung Kim, Minwoo Kim, Jaekoo Lee, Dahuin Jung·
arXiv:2608.06901v1 Announce Type: cross Abstract: Vision-language models (VLMs) have achieved remarkable generalization across diverse multimodal tasks through large-scale pre-training, yet their rapidly increasing computational and memory requirements pose significant challenges…
arXiv cs.CL
TIER_1English(EN)·Rahul Murali Shankar, Titus von der Malsburg, Sebastian Pad\'o·
arXiv:2608.07282v1 Announce Type: new Abstract: The recent advances in neural language models have also spurred much work in computational psycholinguistics, asking whether neural LMs are also promising models of human language processing. However, work has been overwhelmingly fo…
ConCor-1 treats vision-language grounding as bidirectional concept correspondence, jointly predicting text spans, image segments, and cross-modal matches without prespecified phrases.
arXiv cs.LG
TIER_1English(EN)·Rasul Khanbayov, Hasan Kurban·
arXiv:2608.05675v1 Announce Type: new Abstract: Label-free reliability for vision-language models rests on invariance: perturb the input and a faithful reader's answer should not change. This has a known blind spot, a systematic misreading survives the perturbation and gets certi…
arXiv cs.AI
TIER_1English(EN)·J. de Curt\`o, Dayani Plasencia, Diego S\'anchez, I. de Zarz\`a·
arXiv:2608.06154v1 Announce Type: cross Abstract: Vision-language models (VLMs) are increasingly used as zero-shot controllers, but successful trajectories do not necessarily show that decisions are grounded in visual input: simulator dynamics and conservative action priors can p…
arXiv:2605.16411v2 Announce Type: replace-cross Abstract: Hallucination remains a fundamental challenge in vision-language models (VLMs), where autoregressive generation may produce linguistically plausible yet physically inconsistent or visually ungrounded responses due to likel…
Vision-language models are increasingly serving as the reasoning core of embodied agents. Robot execution is inherently iterative: each action reshapes the scene and physical state, continually renewing what must be perceived, reasoned about, and verified. Meeting these demands r…
Label-free reliability for vision-language models rests on invariance: perturb the input and a faithful reader's answer should not change. This has a known blind spot, a systematic misreading survives the perturbation and gets certified wrong, which we show is computable, not jus…
arXiv:2608.04244v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) make grounded predictions in real-world scenes by combining visual and textual cues, yet existing benchmarks rarely reveal how they arbitrate between these evidence sources when they confli…
arXiv:2608.04496v1 Announce Type: cross Abstract: Visual inputs in vision-language models (VLMs) are often encoded into substantially longer token sequences than text, making visual tokens a major bottleneck for efficient inference. Abundant recent methods address this bottleneck…
arXiv:2608.04454v1 Announce Type: cross Abstract: Mixture-of-experts vision-language models (MoE-VLMs) increase model capacity with sparse expert activation, yet deployment requires storing the full expert pool. Training-free expert merging reduces this burden, and many routing-b…
arXiv cs.LG
TIER_1English(EN)·Grzegorz Gruszczynski, Pawel Olszowiec, Michal Byra, Grzegorz Stefanski, Alberto Presta·
arXiv:2608.04879v1 Announce Type: new Abstract: Vision Transformers (ViTs) achieve strong image-recognition performance, but their parameter count grows linearly with depth when each block is independently parameterized. Single-block recurrent ViTs (bViT) remove this growth by re…
arXiv:2608.04885v1 Announce Type: cross Abstract: Standard accuracy metrics for VLMs often mask significant reliability failures in sensitive domains. In this work, we utilize a histopathology-validated brain MRI dataset to systematically assess the diagnostic robustness of four …
arXiv cs.AI
TIER_1English(EN)·Cheng Yin, Yankai Lin, Wang Xu, Sikyuen Tam, Xiangrui Zeng, Zhiyuan Liu, Zhouping Yin·
arXiv:2511.15669v3 Announce Type: replace-cross Abstract: Does Chain-of-Thought (CoT) reasoning genuinely improve Vision Language Action (VLA) models, or does it merely add overhead? Existing CoT-VLA systems report limited and inconsistent gains, yet no prior work has rigorously …
arXiv:2608.04765v1 Announce Type: cross Abstract: Vision-language-action (VLA) models provide a unified paradigm for connecting visual perception, language understanding, and robotic control. However, existing VLA models still face major challenges in long-horizon tasks: sparse e…
arXiv:2608.04509v1 Announce Type: new Abstract: Vision-language systems combine images with retrieved text, but these sources can disagree or jointly fail to support an answer. Reliable models must identify the trustworthy source and abstain when neither is adequate. Existing pos…
Mixture-of-experts vision-language models (MoE-VLMs) increase model capacity with sparse expert activation, yet deployment requires storing the full expert pool. Training-free expert merging reduces this burden, and many routing-based methods aggregate routing statistics across a…
arXiv:2608.03733v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) have achieved remarkable performance across vision-language tasks, but their progress depends heavily on large-scale, high-quality multimodal data that are costly to annotate. Self-augmentati…
arXiv:2608.02830v1 Announce Type: cross Abstract: Many-shot in-context learning (ICL) lets vision-language models (VLMs) adapt from image--label demonstrations without weight updates, and is widely assumed to improve as more demonstrations are supplied. We show the opposite: as d…
arXiv cs.AI
TIER_1English(EN)·Jinquan Zhang, Dongfu Yin, Run Yang, Yufeng Yan, Zhen Tian, F. Richard Yu·
arXiv:2608.03231v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) policies promise general robotic manipulation, but their robustness against physical-world attacks remains fragile. In particular, we show that physically realizable adversarial patches can reliably in…
arXiv:2608.03204v1 Announce Type: cross Abstract: Post-training reinforcement learning (RL) algorithms are commonly used to align large vision-language models (LVLMs) with human intent and the requirements of visual reasoning tasks. However, existing RL-based alignment methods ar…
arXiv cs.AI
TIER_1English(EN)·Paribesh Regmi, Qingshuang Chen, Chi Zhang, Heba Aly, Yelin Kim, Hongda Mao·
arXiv:2608.03112v1 Announce Type: cross Abstract: Vision-language models excel at image and video understanding but suffer from high inference latency due to the need to process thousands of tokens per image, limiting their deployment on resource-constrained edge devices and in r…
arXiv cs.AI
TIER_1English(EN)·Inkyu Sa, Konstantin Stulov, Rajat Bhageria·
arXiv:2608.02958v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) policies trained by behavior cloning fail silently: from the action stream alone, a collapsing rollout looks much like one making clean progress, because imitation supplies no notion of progress. Reinf…
arXiv:2604.02486v3 Announce Type: replace-cross Abstract: Vision-language models (VLMs) have achieved impressive performance across a wide range of multimodal tasks. However, they often fail on tasks that require fine-grained visual perception, even when the required information …
arXiv:2509.23690v2 Announce Type: replace-cross Abstract: Safety hazards in the home are a leading cause of preventable domestic injuries, motivating an automated inspector that actively explores a home and reports hazards before they cause harm. We introduce HomeSafeBench, the f…
Leveraging pre-trained vision-language models (VLMs) to construct vision-language-action (VLA) models has emerged as a promising paradigm for 3D robot manipulation. However, existing 3D VLA methods remain data-hungry, exhibit limited generalization under distribution shifts, and …
Vision-language models exhibit remarkable zero-shot capabilities but suffer significant performance degradation under distribution shifts. While test-time adaptation (TTA) via Low-Rank Adaptation offers a parameter-efficient solution, we identify a fundamental bottleneck in curre…
Multimodal large language models (MLLMs) have achieved remarkable performance across vision-language tasks, but their progress depends heavily on large-scale, high-quality multimodal data that are costly to annotate. Self-augmentation offers a promising alternative by enabling mo…
Fewer visual tokens do not guarantee lower end-to-end latency. We evaluate break-even with a reproducible protocol that accounts for decision overhead, shared work, and the operators each policy can avoid. A stage-level decomposition reconciles these components with measured end-…
arXiv cs.CL
TIER_1English(EN)·Shalom Kachko, Raz Lapid, Margarita Vald, Almog Dubin, Moshe Sipper·
arXiv:2608.00561v1 Announce Type: cross Abstract: Vision-language models (VLMs) process image patches and text tokens in a shared residual stream, but the local geometry through which the two modalities interact remains poorly understood. Most interpretability methods identify gl…
arXiv cs.CL
TIER_1English(EN)·Jing Wu, Jianhua Wu, Jiayi Guan, Jiahong Chen, Jinghui Lu, Hangjun Ye, Bingzhao Gao, Long Chen·
arXiv:2608.01899v1 Announce Type: cross Abstract: Vision-Language Models (VLMs) perform well on commonsense reasoning tasks but struggle with visual spatial reasoning. Most existing solutions introduce extra 3D prior inputs or external spatial encoders, which increase complexity …
arXiv:2604.00086v2 Announce Type: replace-cross Abstract: The field of computer vision has experienced significant advancements through scalable vision encoders and multimodal pre-training frameworks. However, existing approaches often treat vision encoders and large language mod…
arXiv cs.LG
TIER_1English(EN)·Mayank Nautiyal, Li Ju, Andreas Hellander, Ekta Vats, Prashant Singh·
arXiv:2605.13352v2 Announce Type: replace Abstract: Standard dual-encoder vision-language models that map images and text to deterministic points on a shared unit hypersphere through $\ell_2$ normalization typically expose neither \emph{aleatoric} uncertainty (cross-modal ambigui…
arXiv:2608.01263v1 Announce Type: new Abstract: On-policy distillation (OPD) samples trajectories from the current student policy and minimizes token-level divergence between student and teacher next-token distributions at prefixes along those trajectories. This aligns the distil…
Multimodal large language models (MLLMs) make grounded predictions in real-world scenes by combining visual and textual cues, yet existing benchmarks rarely reveal how they arbitrate between these evidence sources when they conflict. We introduce SIGNPOST-Bench, a controlled coun…
Vision-language models (VLMs) achieve strong semantic understanding but remain unreliable in metric spatial reasoning, particularly when queries require comparing multiple instances of the same object category. We study this problem through the Closest-Instance Distance Query (CI…
arXiv cs.LG
TIER_1English(EN)·Mushir Akhtar, M. Tanveer·
arXiv:2607.28696v1 Announce Type: new Abstract: Medical vision-language models (VLMs) can retain high observed marginal coverage after clinical shift while substantially under-covering an individual disease class. The affected class varies with acquisition protocol and backbone g…
arXiv:2607.29613v1 Announce Type: cross Abstract: Reinforcement learning (RL) post-training of Vision-Language-Action (VLA) models has shown strong promise for robotic manipulation. Among RL methods, critic-based approaches rely on a value estimator that predominantly operates on…
arXiv cs.AI
TIER_1English(EN)·Md Ashikur Rahman, Md Arifur Rahman, Niamul Hassan Samin, Abdullah Ibne Hanif Arean, Juena Ahmed Noshin·
arXiv:2603.06828v2 Announce Type: replace-cross Abstract: We uncover a behavioral law of long-horizon vision-language models: models that maintain temporally grounded beliefs generalize better. Standard benchmarks measure only final-answer accuracy, which obscures how models use …
In video understanding, vision-language models (VLMs) must ingest massive numbers of visual tokens, causing the computational and memory cost of the prefill stage to rise sharply. Such visual sequences are highly redundant along the spatio-temporal dimension, yet a high compressi…
Vision-language MoE batches contain different numbers of image and text tokens. Image resolution, image count, tiling, and prompt length all change this token mix. We call the standard token-level Switch auxiliary loss Std-Aux. Std-Aux balances only the mixed load, so large image…
arXiv:2603.12055v3 Announce Type: replace-cross Abstract: Continual learning of pretrained vision-language models (VLMs) is prone to catastrophic forgetting, yet current approaches adapt to new tasks without explicitly preserving the cross-modal semantic geometry inherited from p…
arXiv:2607.26914v1 Announce Type: cross Abstract: Biomedical laboratory robots must navigate to instruments before performing experimental procedures. Existing embodied navigation platforms are designed for household environments and treat a target as an object center or an arbit…
arXiv cs.CL
TIER_1English(EN)·Yuming Yan, Kai Tang, Sihong Chen, Ke Xu, Dan Hu, Qun Yu, Pengfei Hu·
arXiv:2604.16557v2 Announce Type: replace-cross Abstract: Current post-training methodologies for adapting Large Vision-Language Models (LVLMs) generally fall into two paradigms: Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL). Despite their prevalence, both approach…
arXiv cs.LG
TIER_1English(EN)·Supratik Bhowal, Subhrajyoti Basu, Aritra Gir Mahanta, Anik Pal Chowdhury·
arXiv:2607.27304v1 Announce Type: new Abstract: Medical vision-language models (VLMs) generate chain-of-thought (CoT) reasoning before answering clinical questions, but whether this reasoning causally influences predictions remains unclear. We present CoT-Mediate, a behavioral fr…
arXiv:2607.28627v1 Announce Type: cross Abstract: Long visual context poses a challenge for vision-language models: performance degrades as the number of distractors grows, and processing all tokens at once is computationally infeasible under GPU memory constraints. We present Re…
Reinforcement learning (RL) post-training of Vision-Language-Action (VLA) models has shown strong promise for robotic manipulation. Among RL methods, critic-based approaches rely on a value estimator that predominantly operates on single-frame observations or single-frame VLM bac…
Vision-Language Models (VLMs) such as CLIP are now foundational to multimodal systems, yet their robustness to spurious correlations remains poorly understood at scale. We present the first large-scale empirical study of 194 publicly available VLMs, including 16 model families, c…
Long visual context poses a challenge for vision-language models: performance degrades as the number of distractors grows, and processing all tokens at once is computationally infeasible under GPU memory constraints. We present ReToken, a single learnable embedding trained as an …
Despite the impressive visuomotor capabilities enabled by Vision-Language-Action (VLA) models, their performance often degrades on challenging and out-of-domain tasks. Recent test-time steering and scaling methods improve performance without extensive data collection and retraini…
Integrating 3D medical images with vision-language models (VLMs) holds substantial promise for computer-aided diagnosis. However, volumetric images generate prohibitively long visual-token sequences with considerable spatial and inter-slice redundancy. Existing token compression …
arXiv cs.AI
TIER_1English(EN)·Minhyeok Lee, Chiyoung Kim, Chanhoe Gu, Seongrok Kim, Sanghyuk Roy Choi, Donghwan Hwang, Donghun Ryu, Seokhyun Kim·
arXiv:2607.25487v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models translate natural-language commands into robot action sequences, but leading systems on the LIBERO-Plus robustness benchmark use three- to seven-billion-parameter backbones whose memory demands ca…
arXiv cs.AI
TIER_1English(EN)·Zonghe Liu (University of Hong Kong), Shanyuan Jie (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences), Xiaoquan Sun (Huazhong University of Science and Technology), Chen Cao (University of Hong Kong), Zetian Xu (University of Hong …·
arXiv:2607.25912v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models have shown strong potential for general robot manipulation, but most existing models rely on 2D visual-language backbones and lack fine-grained 3D understanding of target objects, especially und…
Evaluating whether a vision-language model (VLM) can act through a physical body is challenging. The outcome of an action couples the VLM's decision with motor control. When a task fails, it is hard to tell whether the VLM made a bad choice or the motor controller simply failed t…
Vision-language-action (VLA) models commonly adopt an LLM-centric V to L to A pathway, where visual observations are projected into the representation space of a large language model before being decoded into robot actions. Although effective, this design incurs substantial compu…
arXiv:2607.22771v1 Announce Type: cross Abstract: Picking the frozen image encoder for a 3D~CT vision--language model (VLM), together with the token-compression scheme on top of it, is a search over many candidates. There are several encoders, several ways to compress their token…
arXiv cs.AI
TIER_1English(EN)·Ali Ansari, Yasmin Mohammadi, Farnoush Nili, Parsa Esmaeilkhani, Longin Jan Latecki, Eduard Dragut·
arXiv:2607.24707v1 Announce Type: new Abstract: Entity-Relationship Diagrams (ERDs) are central to conceptual database design, yet they are typically available only as rendered images rather than machine-readable schemas, limiting AI-assisted database engineering. We introduce ER…
arXiv:2607.22600v1 Announce Type: new Abstract: Information visualizations are widely used to communicate patterns, trends, and outliers, yet deceptive design choices-such as truncated or inverted axes, distorted aspect ratios, inappropriate encodings, and misleading color mappin…
arXiv:2607.23125v1 Announce Type: new Abstract: Post-training enables vision-language models (VLMs) to understand human instructions and perform various downstream tasks. Current post-training methods usually rely on human-annotated data, distillation from external models, reinfo…
arXiv:2210.05487v3 Announce Type: replace Abstract: Unlike most neural language models, humans learn language in a rich, multi-sensory and, often, multi-lingual environment. Current language models typically fail to fully capture the complexities of multilingual language use. We …
arXiv:2607.24440v1 Announce Type: cross Abstract: Vision-language models (VLMs) deployed on consumer hardware must decide when to answer and when to defer, and that decision depends on having a confidence signal that tracks correctness. A practitioner with a fixed memory budget f…
arXiv cs.CL
TIER_1English(EN)·Sultan Alshehri, Zhantao Yang, Han Zhang, Marios Savvides·
arXiv:2607.23052v1 Announce Type: cross Abstract: Dual-encoder vision-language models (VLMs) expose a similarity interface that enables zero-shot retrieval but fails compositional constraints: queries like "umbrella and no person" retrieve images containing both, even when concep…
arXiv cs.AI
TIER_1English(EN)·Alexander G. Ororbia, Ankur Mali, Mary Alexandria Kelly, David Reitter·
arXiv:1805.11546v3 Announce Type: replace-cross Abstract: We examine the benefits of visual context in training neural language models to perform next-word prediction. A multi-modal neural architecture is introduced that outperform its equivalent trained on language alone with a …
arXiv:2605.07544v3 Announce Type: replace Abstract: When you read a paper about a new Vision-Language Model today, it can be easy to forget how strange this idea would have sounded not so long ago. Teaching machines to see was already hard. Teaching them to read and generate lang…
arXiv:2607.22997v1 Announce Type: cross Abstract: Physical AI -- the integration of large vision-language-action (VLA) models with embodied agents that act in the real world -- has emerged as the next major frontier for AI, echoed by industry leaders such as Jensen Huang (``the n…
Multimodal Large Language Models (MLLMs) achieve strong performance by integrating visual inputs with the rich priors of pretrained language models. However, they often fail on vision-centric tasks, especially when visual evidence conflicts with pretrained knowledge. We explore t…
arXiv cs.CL
TIER_1English(EN)·Gwang Gook Lee, Kenan Emir Ak, Jay Mohta, Yan Xu, Dimitrios Dimitriadis·
arXiv:2607.21617v1 Announce Type: cross Abstract: Vision Language Models (VLMs) are increasingly used in place of traditional OCR pipelines for document understanding. In this paper, we show they do not always act as faithful transcribers: when text is imperfect, they often tend …
We introduce PerceptionBench, a benchmark specifically designed to evaluate the atomic visual perception capabilities of Multimodal Large Language Models (MLLMs). Existing benchmarks often fail to isolate perception: holistic evaluations conflate perceptual errors with failures i…
We present N_0-VTLA, a vision-tactile-language-action (VTLA) foundation model capable of (1) fine-grained contact-rich manipulation with tactile perception and tactile-feedback control, and (2) offline policy improvement from stored deployment data. Building on current vision-bas…
Large Vision-Language Models (LVLMs) remain bottlenecked by massive computational footprints, precluding their deployment on resource-constrained edge devices. While efforts to compress LVLMs focus heavily on vision token reduction or smaller language models, the vision encoder i…
arXiv:2607.21401v1 Announce Type: cross Abstract: A vision-language AI assistant returns its answer as a stream of generated tokens. Therefore, a safety guard that watches that answer has to keep up with the stream and stop a harmful reply before a user reads it. Recent vision-la…
arXiv:2607.20351v1 Announce Type: cross Abstract: We find that vision-language models are sensitive to a specific semantically irrelevant change: the order in which the image and question are presented. Across three models and three benchmarks, image first prompting consistently …
arXiv:2607.20092v1 Announce Type: cross Abstract: Contextual entrainment is the tendency of a model to let auxiliary context in its input pull its output, independently of whether that context is relevant, true, or even meaningful. Recently, it has been identified and given a mec…
We find that vision-language models are sensitive to a specific semantically irrelevant change: the order in which the image and question are presented. Across three models and three benchmarks, image first prompting consistently outperforms question-first prompting, revealing a …
arXiv cs.AI
TIER_1English(EN)·Gautam Rajendrakumar Gare, Jia Shi, Zhiqiu Lin, Deepak Pathak, John Galeotti, Deva Ramanan·
arXiv:2607.18695v1 Announce Type: cross Abstract: A popular route to interpretable zero-shot classification asks a large language model (LLM) to describe each class name and prompts CLIP with the resulting descriptors. We show that these descriptors carry little visual evidence o…
arXiv:2607.18958v1 Announce Type: cross Abstract: While Large Vision-Language Models (LVLMs), represented by LLaVA and GPT-4V, have demonstrated remarkable capabilities, their visual inputs remain vulnerable to adversarial attacks, posing significant security risks. Existing defe…
Contextual entrainment is the tendency of a model to let auxiliary context in its input pull its output, independently of whether that context is relevant, true, or even meaningful. Recently, it has been identified and given a mechanistic account in unimodal language models. Whet…
arXiv cs.CL
TIER_1English(EN)·Sudharshan Balaji, Yili Ren, Guangjing Wang, Yimin Chen, Ning Wang·
arXiv:2607.16442v1 Announce Type: cross Abstract: Machine unlearning is widely used to remove hazardous knowledge from large language models. Modern Vision-Language Models (VLMs), however, process both text and visual inputs, raising a fundamental security question: does unlearni…
arXiv:2510.00037v5 Announce Type: replace-cross Abstract: In Vision-Language-Actionf(VLA) models, robustness to real-world perturbations is critical for deployment. Existing methods target simple visual disturbances, overlooking the broader multi-modal perturbations that arise in…
arXiv:2607.17786v1 Announce Type: cross Abstract: Does adding a reasoning step make a Vision-Language-Action (VLA) model more robust to perturbation? Intuitively, a policy that reasons before acting should absorb a perturbed input better than one that maps observations directly t…
arXiv:2412.18613v2 Announce Type: replace-cross Abstract: Illusions are entertaining, but they are also a useful diagnostic tool in cognitive science, philosophy, and neuroscience. A typical illusion shows a gap between how something `really is' and how something `appears to be',…
arXiv:2512.02924v3 Announce Type: replace Abstract: While Neural Processing Units (NPUs) offer high theoretical efficiency for edge AI, state-of-the-art Vision--Language Models (VLMs) tailored for GPUs often falter on these substrates. We attribute this hardware-model mismatch to…
arXiv:2607.14499v1 Announce Type: new Abstract: Multi-modal Large Language Models (MLLMs) have made substantial advances on benchmarks, yet their real-world effectiveness remains uncertain. This gap stems from the fundamental misalignment between benchmarks in controlled, static …
arXiv:2607.15265v1 Announce Type: cross Abstract: We present SceneBind, an omni-modal representation of realistic scenes with joint semantic and 3D spatial understanding across vision, audio and language. Existing omni-modal encoders excel at instance-level semantics (i.e., what …
arXiv cs.LG
TIER_1English(EN)·Pegah Khayatan, Sara Meziane, Jayneel Parekh, Matthieu Cord·
arXiv:2607.14280v1 Announce Type: cross Abstract: Flow-matching-based vision-language-action (VLA) models have emerged as powerful policies for robotic manipulation, yet a critical capability remains underexplored: fine-grained behavioral control, the ability to govern how a robo…
arXiv:2607.14739v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models have achieved impressive results in visuomotor policy learning, yet remain fundamentally reactive, mapping current observations and language to actions without explicit forward prediction of wor…
We present SceneBind, an omni-modal representation of realistic scenes with joint semantic and 3D spatial understanding across vision, audio and language. Existing omni-modal encoders excel at instance-level semantics (i.e., what is present), but often lack explicit spatial struc…
Vision-Language-Action (VLA) models have achieved impressive results in visuomotor policy learning, yet remain fundamentally reactive, mapping current observations and language to actions without explicit forward prediction of world dynamics. Existing visual foresight methods pre…
arXiv cs.LG
TIER_1English(EN)·Charlotte Morissette, Amin Abyaneh, Wei-Di Chang, Anas Houssaini, David Meger, Hsiu-Chin Lin, Jonathan Tremblay, Gregory Dudek·
arXiv:2603.14604v2 Announce Type: replace-cross Abstract: We propose TacFiLM, a lightweight modality-fusion approach that integrates visual-tactile signals into vision-language-action (VLA) models. While advances in VLAs have introduced robot policies that are both generalizable …
arXiv cs.AI
TIER_1English(EN)·Jose Mart\'inez-Fajardo, Pablo Pueyo, Fernando Caballero, Luis Merino·
arXiv:2607.13624v1 Announce Type: cross Abstract: Natural language interaction provides an intuitive way for non-expert users to communicate with robotic platforms. However, transforming user requests into executable navigation actions remains a challenging task, requiring the in…
arXiv:2607.13072v1 Announce Type: cross Abstract: Zero-shot object-goal navigation aims to enable an intelligent agent to explore and navigate to objects of unknown categories in an unfamiliar environment without specific target training. In zero-shot navigation tasks, pre-traine…
Natural language interaction provides an intuitive way for non-expert users to communicate with robotic platforms. However, transforming user requests into executable navigation actions remains a challenging task, requiring the integration of language understanding, environment p…
Natural language interaction provides an intuitive way for non-expert users to communicate with robotic platforms. However, transforming user requests into executable navigation actions remains a challenging task, requiring the integration of language understanding, environment p…
arXiv:2607.12815v1 Announce Type: new Abstract: Chain-of-Thought (CoT) prompting is widely used as a test-time scaling strategy for Vision-Language Models (VLMs), but it remains unclear what is extended when VLMs generate longer reasoning traces. We ask whether CoT requires conti…
arXiv:2511.09483v3 Announce Type: replace Abstract: While multimodal large language models can describe visual content, their ability to generate executable procedures remains underexplored. CrochetBench presented in this paper evaluates this shift from describing to doing throug…
arXiv cs.AI
TIER_1English(EN)·Maijunxian Wang, Yijiang Li, Bingyang Wang, Tianwei Zhao, Ran Ji, Qingying Gao, Emmy Liu, Hokin Deng, Dezhi Luo·
arXiv:2602.15892v2 Announce Type: replace-cross Abstract: Visual perspective taking--inferring how the world appears from another's viewpoint--is foundational to social cognition. We introduce FlipSet, a diagnostic benchmark for Level-2 visual perspective taking (L2 VPT) in visio…
Chain-of-Thought (CoT) prompting is widely used as a test-time scaling strategy for Vision-Language Models (VLMs), but it remains unclear what is extended when VLMs generate longer reasoning traces. We ask whether CoT requires continued access to image tokens, or whether it mainl…
Chain-of-Thought (CoT) prompting is widely used as a test-time scaling strategy for Vision-Language Models (VLMs), but it remains unclear what is extended when VLMs generate longer reasoning traces. We ask whether CoT requires continued access to image tokens, or whether it mainl…
Cross-image comparative reasoning remains challenging for vision-language models (VLMs), especially when correct prediction requires fine-grained attribute grounding and globally consistent reasoning. We present CoRe, a unified framework for this problem. CoRe includes: (i) CoRe-…
arXiv:2607.11008v1 Announce Type: cross Abstract: Open-vocabulary dense perception (OVDP) aims to localize objects unseen during training by leveraging textual knowledge. Despite the remarkable progress of recent CLIP-based approaches, we identify a critical limitation: synonym-i…
arXiv:2607.09792v1 Announce Type: cross Abstract: Navigation is a fundamental capability of autonomous systems, yet most existing approaches rely on highly structured models and strong prior assumptions, limiting their robustness in open and uncertain real-world environments. Vis…
arXiv cs.AI
TIER_1English(EN)·Byungkun Lee, Dongyoon Hwang, Dongjin Kim, Hojoon Lee, Minho Park, Jaegul Choo·
arXiv:2607.11498v1 Announce Type: cross Abstract: Vision-language-action (VLA) models predict robot actions from visual observations and language instructions. These actions are defined in the robot's own 3D coordinate frame, yet most VLAs observe the scene in the camera frame, c…
arXiv:2603.04419v2 Announce Type: replace-cross Abstract: We characterize the phenomenon of context-dependent affordance computation in vision-language models (VLMs). Our primary study uses Qwen3-VL-30B-A3B ($n = 3{,}213$ scene-context pairs from COCO-2017: 479 images under 7 age…
arXiv cs.AI
TIER_1English(EN)·Hayeon Kim, Ji Ha Jang, Junghun James Kim, Se Young Chun·
arXiv:2603.22042v3 Announce Type: replace-cross Abstract: While Vision-Language Models (VLMs) have achieved remarkable performance, their Euclidean embeddings remain limited in capturing hierarchical relationships such as part-to-whole or parent-child structures, and often face c…
arXiv cs.AI
TIER_1English(EN)·Shengzhuo Yang, Ronghao Yu, Chuanjie Lv, Linpeng Peng, Hang Yu, Jie Ren, Jiajun Lv, Yong Liu·
arXiv:2607.10383v1 Announce Type: cross Abstract: Visual Language Navigation foundation models aim to unify deep reasoning for grounded spatial decisions with broad versatility for diverse embodied tasks. Current approaches typically achieve this integration via monolithic polici…
arXiv cs.LG
TIER_1English(EN)·Finn Ferchau, Daniel Pommer, Cristian Axenie·
arXiv:2607.10172v1 Announce Type: cross Abstract: Deploying billion-parameter Vision-Language-Action (VLA) models on industrial hardware requires fine-tuning to bridge the embodiment gap. Full Fine-Tuning (FFT) provides maximal plasticity but requires data centre-grade GPUs. We p…
Vision-language-action (VLA) models predict robot actions from visual observations and language instructions. These actions are defined in the robot's own 3D coordinate frame, yet most VLAs observe the scene in the camera frame, creating a frame mismatch between where the scene i…
arXiv cs.AI
TIER_1English(EN)·Shravan Murlidaran, Miguel P. Eckstein·
arXiv:2607.09654v1 Announce Type: cross Abstract: Vision language models (VLMs) have made remarkable progress in visual reasoning during the last decade. Most evaluations have used simple scenes (MS-COCO) that do not showcase complex human interactions or behaviors, only a handfu…
arXiv:2607.09450v1 Announce Type: cross Abstract: Pre-trained Vision-Language Models (VLMs) such as CLIP achieve strong zero-shot generalization, but their performance degrades sharply under adversarial perturbations. Existing test-time adaptation methods typically rely on sample…
arXiv:2607.09657v1 Announce Type: cross Abstract: The rapid progress of large foundation models has been driven predominantly by pretraining on large-scale text corpora. However, many forms of knowledge are conveyed through visual representations, where figures, typeset equations…
Visual Language Navigation foundation models aim to unify deep reasoning for grounded spatial decisions with broad versatility for diverse embodied tasks. Current approaches typically achieve this integration via monolithic policies that map observations directly to actions, yet …
The rapid progress of large foundation models has been driven predominantly by pretraining on large-scale text corpora. However, many forms of knowledge are conveyed through visual representations, where figures, typeset equations, and page layouts carry rich information that can…
Vision language models (VLMs) have made remarkable progress in visual reasoning during the last decade. Most evaluations have used simple scenes (MS-COCO) that do not showcase complex human interactions or behaviors, only a handful of non-curated human descriptions as a benchmark…
Pre-trained Vision-Language Models (VLMs) such as CLIP achieve strong zero-shot generalization, but their performance degrades sharply under adversarial perturbations. Existing test-time adaptation methods typically rely on sample-level confidence heuristics, overlooking the intr…
arXiv:2607.08182v1 Announce Type: cross Abstract: Vision-language-action (VLA) models aim to map multimodal inputs to robot actions. However, most existing approaches struggle to cover complex dynamic scenarios due to treating all visual tokens uniformly and reasoning with human-…
arXiv:2607.07907v1 Announce Type: cross Abstract: With the growing adoption of VLMs, DMs, LLMs, and AFMs, these multimodal foundation models can inadvertently encode sensitive, copyrighted, biased, or unsafe cross-modal associations that originate from their training data. Retrai…
arXiv cs.AI
TIER_1English(EN)·Emily Jin, Joy Hsu, Yiqing Xu, Weiyu Liu, Nick Haber, Jiajun Wu·
arXiv:2607.08024v1 Announce Type: cross Abstract: Long-horizon robot planning requires jointly reasoning over semantic task structure and geometric feasibility. To successfully execute a task, a robot must decompose goals, select task-relevant objects, and sequence actions, while…
The rapid progress of large foundation models has been driven predominantly by pretraining on large-scale text corpora. However, many forms of knowledge are conveyed through visual representations, where figures, typeset equations, and page layouts carry rich information that can…
Vision-language-action (VLA) models aim to map multimodal inputs to robot actions. However, most existing approaches struggle to cover complex dynamic scenarios due to treating all visual tokens uniformly and reasoning with human-selected factors, which lack mechanisms to emphasi…
Vision-language-action (VLA) models aim to map multimodal inputs to robot actions. However, most existing approaches struggle to cover complex dynamic scenarios due to treating all visual tokens uniformly and reasoning with human-selected factors, which lack mechanisms to emphasi…
arXiv cs.AI
TIER_1English(EN)·Juyi Lin, Amir Taherin, Arash Akbari, Arman Akbari, Lei Lu, Guangyu Chen, Taskin Padir, Xiaomeng Yang, Weiwei Chen, Yiqian Li, Xue Lin, David Kaeli, Pu Zhao, Yanzhi Wang·
arXiv:2507.05116v5 Announce Type: replace-cross Abstract: Recent large-scale Vision Language Action (VLA) models have shown superior performance in robotic manipulation tasks guided by natural language. However, current VLA models suffer from two drawbacks: (i) generation of mass…
arXiv:2607.07375v1 Announce Type: cross Abstract: Adversarial vulnerability in deep neural networks (DNNs) has been studied from the perspectives of decision-boundary geometry, feature robustness, input-output Jacobians, and the instability of inverse problems. Here, we focus on …
arXiv:2607.07251v1 Announce Type: new Abstract: One of the expected abilities of vision-language models (VLMs) is spatial reasoning ability based on a given text and image. To evaluate the spatial reasoning abilities of VLMs, we focus on the use of spatial deictic expressions, wh…
arXiv cs.AI
TIER_1English(EN)·Inkyu Sa, Chanoh Park, Hea-Min Lee, Donghee Noh, Ho Seok Ahn·
arXiv:2607.06706v1 Announce Type: cross Abstract: Vision Language Action (VLA) models unify visual perception, natural-language understanding, and action generation within a single foundation model, allowing a robot to follow instructions such as fold the towel or fly to the red …
arXiv cs.AI
TIER_1English(EN)·Peter Bohm, Saimunur Rahman, Abdelwahed Khamis, Sagun Man Singh Shrestha, Chris McCool, Peyman Moghadam·
arXiv:2607.06882v1 Announce Type: cross Abstract: Visual navigation policies built on large pretrained models have so far followed a common recipe: a dedicated visual encoder, a bespoke action head, and training on thousands of hours of cross-embodiment datasets. We ask whether t…
arXiv:2512.05693v2 Announce Type: replace-cross Abstract: Generalist vision--language--action (VLA) policies are typically trained on heterogeneous mixtures of robot demonstrations spanning diverse embodiments, action spaces, and observation configurations. Modeling such heteroge…
With the growing adoption of VLMs, DMs, LLMs, and AFMs, these multimodal foundation models can inadvertently encode sensitive, copyrighted, biased, or unsafe cross-modal associations that originate from their training data. Retraining after deletion requests or policy updates is …
Adversarial vulnerability in deep neural networks (DNNs) has been studied from the perspectives of decision-boundary geometry, feature robustness, input-output Jacobians, and the instability of inverse problems. Here, we focus on the spectral structure of intermediate linear tran…
One of the expected abilities of vision-language models (VLMs) is spatial reasoning ability based on a given text and image. To evaluate the spatial reasoning abilities of VLMs, we focus on the use of spatial deictic expressions, which are defined as spatial expressions whose ref…
arXiv:2602.19710v3 Announce Type: replace-cross Abstract: Existing Vision-Language-Action (VLA) models often suffer from feature collapse and low training efficiency because they entangle high-level perception with sparse, embodiment-specific action supervision. Since these model…
arXiv cs.AI
TIER_1English(EN)·Sohwi Lim, Lee Hyoseok, Jungjoon Park, Tae-Hyun Oh·
arXiv:2604.11539v2 Announce Type: replace-cross Abstract: Human perception of visual similarity is inherently adaptive and subjective, depending on the users' interests and focus. However, most image retrieval systems fail to reflect this flexibility, relying on a fixed, monolith…
arXiv:2601.06521v2 Announce Type: replace-cross Abstract: While humans develop core visual skills long before acquiring language, contemporary Multimodal LLMs (MLLMs) still rely heavily on linguistic priors to compensate for their fragile visual understanding. We uncovered a cruc…
arXiv:2607.06370v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models have emerged as a promising approach for generalizable robotic manipulations. In particular, flow matching-based VLA models have shown remarkable success due to their capability to generate prec…
Visual navigation policies built on large pretrained models have so far followed a common recipe: a dedicated visual encoder, a bespoke action head, and training on thousands of hours of cross-embodiment datasets. We ask whether this recipe is necessary. In this paper, we introdu…
LaMem-VLA introduces a latent-memory-native framework that integrates historical experience into vision-language-action reasoning through coordinated memory components operating in the same latent space.
Vision-Language-Action (VLA) models have emerged as a promising approach for generalizable robotic manipulations. In particular, flow matching-based VLA models have shown remarkable success due to their capability to generate precise and smooth action sequences and capture multim…
Vision-Language-Action (VLA) models have emerged as a promising approach for generalizable robotic manipulations. In particular, flow matching-based VLA models have shown remarkable success due to their capability to generate precise and smooth action sequences and capture multim…
arXiv cs.AI
TIER_1English(EN)·Suhyeok Jang, Dongyoung Kim, Changyeon Kim, Youngsuk Kim, Jinwoo Shin·
arXiv:2510.05681v2 Announce Type: replace-cross Abstract: Vision-Language-Action models (VLAs) have demonstrated remarkable performance in robot control. However, they remain fundamentally limited in tasks that require high precision due to their single-inference paradigm. While …
arXiv cs.AI
TIER_1English(EN)·Qingqian Yang, Hao Wang, Sai Qian Zhang, Jian Li, Yang Hua, Miao Pan, Tao Song, Zhengwei Qi, Haibing Guan·
arXiv:2607.03358v1 Announce Type: cross Abstract: We study how visual information is routed in vision-language models (VLMs). Using causal patching on controlled synthetic and natural datasets, we find that models rely on two distinct pathways to solve visual tasks: A direct path…
arXiv:2607.04683v1 Announce Type: cross Abstract: Vision-language models (VLMs) perform well on visual question answering with high-quality images but struggle when questions require knowledge beyond what is clearly and directly visible. In such settings, uncertainty quantificati…
arXiv cs.LG
TIER_1English(EN)·Harsh Goel, S P Sharan, Sahil Shah, Minkyu Choi, Joungbin An, Kristen Grauman, Sandeep P. Chinchali·
arXiv:2607.02959v1 Announce Type: cross Abstract: We introduce VSeek, an agentic framework that transforms long-video question answering (LVQA) from a passive, single-pass perception task into a multi-turn retrieval process. VSeek utilizes a natural language-driven search to iden…
arXiv cs.LG
TIER_1English(EN)·Zelin Zhao, Min Shi, Bo Yuan, Haotian Xue, Jialuo Li, Lama Moukheiber, Humphrey Shi, Yongxin Chen·
arXiv:2607.03461v1 Announce Type: cross Abstract: World models aim to capture environment dynamics in ways that support perception, reasoning, and action, and have recently become a central direction in Vision-Language-Action-World (VLAW) modeling. Meanwhile, unified vision-langu…
arXiv cs.LG
TIER_1English(EN)·Jaeyoung Kim, Eunseok Kim, Dongsuk Jang·
arXiv:2607.05268v1 Announce Type: cross Abstract: Whether a hyperbolic representation model uses its geometry cannot be read off its curvature parameter: what matters is the dimensionless operating point $\sqrt{c}\rho$ and whether the radial and cone machinery is active there. We…
arXiv cs.AI
TIER_1English(EN)·Suhyeong Park, Junha Jung, Jungwoo Park, Jaewoo Kang·
arXiv:2607.04593v1 Announce Type: cross Abstract: Vision-Language Models (VLMs) have demonstrated impressive capabilities across different tasks, but their computational cost is dominated by the large number of visual tokens fed to the language model. Existing token reduction met…
arXiv cs.AI
TIER_1English(EN)·Xinchuan Qiu, Yi Yu·
arXiv:2607.04591v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models have demonstrated strong capabilities in robotic manipulation by integrating visual perception, language understanding, and robot action generation. Existing research has primarily focused on im…
arXiv:2607.04163v1 Announce Type: cross Abstract: Large Vision-Language Models (LVLMs) have achieved remarkable progress in visual understanding tasks such as image captioning and visual question answering. However, they remain susceptible to hallucinations, generating content th…
arXiv cs.AI
TIER_1English(EN)·Seung Il Lee, Qinqian Lei, Daguang Xu, Dong Yang, Robby T. Tan, Yixin Chen, Bo Wang·
arXiv:2607.03595v1 Announce Type: cross Abstract: Affordance grounding aims to localize image regions that support a specific action, serving as a core capability for physical intelligence and embodied perception. Previous studies have primarily relied on weakly supervised learni…
arXiv:2607.03449v1 Announce Type: cross Abstract: Current Vision-Language-Action (VLA) models excel at robotic manipulation but often struggle with non-Markovian tasks requiring long-term memory and reasoning due to their reliance on immediate observations. Existing solutions fac…
arXiv:2607.03182v1 Announce Type: cross Abstract: Autonomous driving planning requires translating navigation intent, traffic rules, dynamic interactions, and language instructions into executable continuous trajectories. Vision-Language-Action models have been introduced into dr…
arXiv cs.AI
TIER_1English(EN)·Chengzhen Yu, Canran Xiao, Siyuan Ma, Yang Liu·
arXiv:2607.03143v1 Announce Type: cross Abstract: Vision-language alignment powers open-vocabulary recognition, retrieval, and LVLM grounding, yet natural captions are often underspecified, making similarity brittle and overly confident under paraphrase and omitted details. We ai…
arXiv:2607.02995v1 Announce Type: cross Abstract: Vision-language models can exhibit visual concept-conditioned divergence: given images containing demographic features, corporate logos, or ideological symbols, some models produce unusually uniform responses that differ from what…
arXiv:2607.02819v1 Announce Type: cross Abstract: Cloud-edge Large Vision-Language Model (LVLM) inference enables efficient deployment by splitting computation between edge devices and cloud servers. In this process, intermediate vision tokens are transmitted from the edge to the…
arXiv:2607.02575v1 Announce Type: cross Abstract: Vision-language models can perform new tasks without parameter updates through in-context learning (ICL), whose core mechanism is utilizing the support set for task induction. In the standard ICL setting, once the task is induced,…
arXiv:2607.05310v1 Announce Type: new Abstract: Model editing promises a fast, targeted way to correct post-deployment mistakes in medical vision-language models (VLMs) without costly retraining. However, existing multimodal model editing benchmarks focus on general-purpose tasks…
arXiv:2508.08521v2 Announce Type: replace-cross Abstract: Vision Language Models (VLMs) are increasingly being used in a broad range of applications, bringing their security and behavioral control to the forefront. While existing approaches for behavioral control or output redire…
arXiv:2509.25533v2 Announce Type: replace-cross Abstract: As Vision Language Models (VLMs) are deployed across safety-critical applications, understanding and controlling their behavioral patterns has become increasingly important. Existing behavioral control methods face signifi…
arXiv:2512.14926v2 Announce Type: replace-cross Abstract: Focusing on low-resource languages is an essential step toward democratizing generative AI. In this work, we contribute to reducing the multimodal NLP resource gap for Romanian. We translate the widely known Flickr30K data…
Model editing promises a fast, targeted way to correct post-deployment mistakes in medical vision-language models (VLMs) without costly retraining. However, existing multimodal model editing benchmarks focus on general-purpose tasks and do not reflect realistic clinical domain re…
Whether a hyperbolic representation model uses its geometry cannot be read off its curvature parameter: what matters is the dimensionless operating point $\sqrt{c}ρ$ and whether the radial and cone machinery is active there. We develop a battery of necessary-condition diagnostics…
Existing video benchmarks evaluate action recognition on consumer videos, egocentric recordings, or simulated industrial environments. They do not test vision-language models under the visual and procedural conditions of real industrial CCTV, where workers appear as distant figur…
Vision-language models (VLMs) perform well on visual question answering with high-quality images but struggle when questions require knowledge beyond what is clearly and directly visible. In such settings, uncertainty quantification should not only indicate whether the model is l…
Multi-vector vision-language retrieval preserves fine-grained visual evidence through maximum-similarity late interaction, but dense image-side tokens make storage and scoring expensive. Existing token compression methods reduce this cost, yet they can remove or collapse object- …
Multi-vector vision-language retrieval preserves fine-grained visual evidence through maximum-similarity late interaction, but dense image-side tokens make storage and scoring expensive. Existing token compression methods reduce this cost, yet they can remove or collapse object- …
Multi-vector vision-language retrieval preserves fine-grained visual evidence through maximum-similarity late interaction, but dense image-side tokens make storage and scoring expensive. Existing token compression methods reduce this cost, yet they can remove or collapse object- …
Vision-Language-Action (VLA) models have demonstrated strong capabilities in robotic manipulation by integrating visual perception, language understanding, and robot action generation. Existing research has primarily focused on improving model architectures, training strategies, …
arXiv:2607.01754v1 Announce Type: new Abstract: On-policy exploration is a crucial component for training robust Vision-Language Navigation agents, as it exposes the policy to a broader state distribution. However, such exploration inevitably leads to trajectories that deviate fr…
arXiv cs.AI
TIER_1English(EN)·Ninghao Zhang, Bin Zhu, Shijie Zhou, Jingjing Chen·
arXiv:2603.06001v2 Announce Type: replace-cross Abstract: Vision-Language-Action (VLA) models enable robots to perform manipulation tasks directly from natural language instructions and are increasingly viewed as a foundation for generalist robotic policies. However, their reliab…
arXiv cs.AI
TIER_1English(EN)·Phillip Howard, Xin Su, Kathleen C. Fraser·
arXiv:2604.09945v2 Announce Type: replace-cross Abstract: The rapid adoption of large vision-language models (LVLMs) in recent years has been accompanied by growing fairness concerns due to their propensity to reinforce harmful societal stereotypes. While significant attention ha…
arXiv:2607.02089v1 Announce Type: cross Abstract: Vision-language models (VLMs) have achieved strong performance across diverse multimodal tasks, yet they remain vulnerable to unreliable reasoning. Existing self-correction methods mitigate these issues but typically rely on post-…
arXiv:2607.02269v1 Announce Type: cross Abstract: Vision-Language Models (VLMs) have demonstrated immense promise in Spatio-Temporal Video Grounding (STVG). However, current evaluation protocols are largely confined to zero-shot assessments on general, daily-life benchmarks. This…
Vision-Language Models (VLMs) have demonstrated immense promise in Spatio-Temporal Video Grounding (STVG). However, current evaluation protocols are largely confined to zero-shot assessments on general, daily-life benchmarks. This creates a critical disconnect from real-world app…
Large Vision-Language Models (LVLMs) have achieved remarkable progress in multimodal understanding, yet their enormous parameter scale and cross-modal computation incur substantial memory and latency overhead, severely limiting real-world deployment on resource-constrained device…
On-policy exploration is a crucial component for training robust Vision-Language Navigation agents, as it exposes the policy to a broader state distribution. However, such exploration inevitably leads to trajectories that deviate from expert demonstrations, resulting in a semanti…
arXiv cs.AI
TIER_1English(EN)·Lukas Kuhn, Giuseppe Serra, Randall Balestriero, Florian Buettner·
arXiv:2607.00784v1 Announce Type: cross Abstract: Vision-language pretraining remains dominated by contrastive objectives, whereas vision-only self-supervised learning has largely adopted non-contrastive methods. At the same time, the role of vision-language encoders has shifted:…
arXiv:2607.00684v1 Announce Type: new Abstract: The classification accuracy of pretrained Vision-Language Models (VLMs) relies on the quality of the text prompts. Handcrafted templates and Large Language Model (LLM)-generated descriptions not only make predictions more interpreta…
arXiv cs.LG
TIER_1English(EN)·Kai Hu, Akash Bharadwaj, Weichen Yu, Matt Fredrikson·
arXiv:2607.00174v1 Announce Type: cross Abstract: We present a black-box model-stealing attack that recovers private vision-tokenizer configurations of deployed vision-language models (VLMs), including the visual patch size and input preprocessing pipeline. The key idea is a task…
arXiv:2602.21397v2 Announce Type: replace-cross Abstract: Prompt learning has become a dominant paradigm for adapting vision-language models (VLMs) such as CLIP to downstream tasks without modifying pretrained weights. While extending prompts to both vision and text encoders acro…
arXiv cs.AI
TIER_1English(EN)·Shaoheng Zhang, Zhichen Li, Jie Mei·
arXiv:2607.01043v1 Announce Type: cross Abstract: Memory-based discrete vision-language navigation (VLN) agents must act under partial observability, yet even strong frozen backbones remain vulnerable at test time. Two common failure modes are stale historical evidence at memory …
arXiv cs.AI
TIER_1English(EN)·Arpita Nema, Hanwei Zhu, Xi Zhang, Weisi Lin·
arXiv:2607.01086v1 Announce Type: cross Abstract: The evaluation of long-term video quality understanding remains an open challenge for large vision-language models (LVLMs). Existing video quality benchmarks predominantly focus on short clips and isolated distortions, overlooking…
Vision-Language Models struggle with domain adaptation in specialized spatio-temporal video grounding tasks, highlighting limitations in zero-shot generalization and in-context learning capabilities.
The evaluation of long-term video quality understanding remains an open challenge for large vision-language models (LVLMs). Existing video quality benchmarks predominantly focus on short clips and isolated distortions, overlooking the temporal continuity, cumulative degradation, …
Memory-based discrete vision-language navigation (VLN) agents must act under partial observability, yet even strong frozen backbones remain vulnerable at test time. Two common failure modes are stale historical evidence at memory readout and inefficient local backtracking during …
Memory-based discrete vision-language navigation (VLN) agents must act under partial observability, yet even strong frozen backbones remain vulnerable at test time. Two common failure modes are stale historical evidence at memory readout and inefficient local backtracking during …
Vision-language pretraining remains dominated by contrastive objectives, whereas vision-only self-supervised learning has largely adopted non-contrastive methods. At the same time, the role of vision-language encoders has shifted: they are increasingly deployed not as zero-shot c…
Vision Language Models (VLMs) have demonstrated remarkable capabilities in multimodal reasoning tasks, yet they still suffer from recurring failures, such as skipping key visual checks, misapplying domain rules, and hallucinating unsupported concepts. Most existing solutions rely…
The classification accuracy of pretrained Vision-Language Models (VLMs) relies on the quality of the text prompts. Handcrafted templates and Large Language Model (LLM)-generated descriptions not only make predictions more interpretable, but also enable reuse of the same prompts a…
arXiv cs.AI
TIER_1English(EN)·Nan Li, Albert Gatt, Massimo Poesio·
arXiv:2606.31719v1 Announce Type: cross Abstract: In collaborative dialogue, shared perception does not guarantee shared interpretation. Mutual understanding must be established through interaction. We investigate whether vision-language models (VLMs) can distinguish what could b…
arXiv cs.AI
TIER_1English(EN)·Ta Duc Huy, Trang Nguyen, Townim Chowdhury, Ankit Yadav, Minh-Son To, Zhibin Liao, Johan W. Verjans, Vu Minh Hieu Phan·
arXiv:2606.31407v1 Announce Type: cross Abstract: Vision-language models can produce confident answers on visually ambiguous inputs, resulting in biased predictions. Common entropy-based methods, such as Semantic Entropy (SE), rely on output diversity. Yet our analysis shows that…
arXiv:2606.30696v1 Announce Type: cross Abstract: Enabling robots to follow natural language commands to complete zero-shot long-horizon tasks remains challenging. It requires extracting implicit temporal and logical constraints from natural language commands and executing multip…
arXiv cs.LG
TIER_1English(EN)·Cl\'ement Fuchs, Tim Bary, Beno\^it Macq·
arXiv:2606.31577v1 Announce Type: cross Abstract: Conformal predictions have attracted significant attention in the field of uncertainty quantification, mainly because of their strong marginal coverage guarantees. Full conditional guarantee is not an attainable goal, a well known…
arXiv:2603.18481v2 Announce Type: replace-cross Abstract: Out-of-distribution (OOD) detection remains a critical challenge in open-world learning, where models must adapt to evolving data distributions. While recent vision-language models (VLMS) like CLIP enable multimodal OOD de…
In collaborative dialogue, shared perception does not guarantee shared interpretation. Mutual understanding must be established through interaction. We investigate whether vision-language models (VLMs) can distinguish what could be shared from what has been shared between dialogu…
Vision-language models can produce confident answers on visually ambiguous inputs, resulting in biased predictions. Common entropy-based methods, such as Semantic Entropy (SE), rely on output diversity. Yet our analysis shows that overconfident visual embeddings suppress output d…
arXiv:2606.30319v1 Announce Type: cross Abstract: Modeling the bidirectional correspondence between external sensory stimuli and internal neural activity has emerged as a critical frontier in neuroscience. However, existing approaches predominantly treat brain encoding and decodi…
arXiv cs.CL
TIER_1English(EN)·Hee-Seon Kim, Minbeom Kim, Seokil Ham, Changick Kim·
arXiv:2412.08108v3 Announce Type: replace-cross Abstract: Large Vision-Language Models (LVLMs) have achieved remarkable performance on multimodal tasks but remain highly vulnerable to small adversarial perturbations in input images. Existing attacks typically target the vision en…
arXiv cs.CL
TIER_1English(EN)·Matteo Farina, Vishaal Udandarao, Thao Nguyen, Selim Kuzucu, Maximilian B\"other, Andreas Hochlehnert, Adhiraj Ghosh, Marianna Nezhurina, Karsten Roth, Joschka Struber, Yuhui Zhang, Sebastian Dziadzio, Elaine Sui, Soumya Jahagirdar, Dhruba Ghosh, Hasan H…·
arXiv:2606.28551v1 Announce Type: cross Abstract: Building performant Vision-Language Models (VLMs) requires carefully curating large-scale training datasets, yet the community lacks systematic benchmarks for evaluating such curation strategies. We introduce DataComp for VLMs (DC…
arXiv:2602.23353v2 Announce Type: replace-cross Abstract: The Platonic Representation Hypothesis posits that neural networks trained on different modalities converge toward a shared statistical model of the world. Recent work exploits this convergence by aligning frozen pretraine…
arXiv cs.AI
TIER_1English(EN)·Chen Yang, Yuhao Wei, Ze Xu, Ziheng Zou, Shuang Liang, Delin Ouyang, Lingfeng Qi, Jie Li, Guofa Li·
arXiv:2606.29879v1 Announce Type: cross Abstract: Vision-Language Models (VLMs) provide powerful semantic understanding and commonsense reasoning for End-to-End Autonomous Driving (E2E-AD) planning. However, trajectories directly generated by VLMs often encode only coarse driving…
arXiv cs.AI
TIER_1English(EN)·Rahul Chowdhury, Timothy A Rupprecht, Xuan Shen, Pu Zhao, Yanzhi Wang·
arXiv:2606.29579v1 Announce Type: cross Abstract: Spatial reasoning remains a persistent challenge for many vision language models (VLMs), and improving it typically requires fine-tuning with substantial additional parameters. Our preliminary analysis reveals that rescaling activ…
arXiv:2606.29464v1 Announce Type: cross Abstract: Vision-language dataset distillation (VLDD) compresses a large image-text paired dataset into a small set of synthetic pairs that can efficiently train contrastive vision-language models under strict data and compute budgets. Most…
arXiv cs.AI
TIER_1English(EN)·Xiao Wang, Liye Jin, Dan Xu, Yuehang Li, Lan Chen, Yaowei Wang, Yonghong Tian, Jin Tang·
arXiv:2606.29357v1 Announce Type: cross Abstract: Vision-language tracking guided by natural language specifications leverages high-level semantic cues of target objects to substantially boost tracking accuracy and robustness. Existing studies have verified that adaptively optimi…
arXiv cs.AI
TIER_1English(EN)·Minh-Loi Nguyen, Nghiem Tuong Diep, Hung Khang Nguyen, Minh Le, Doanh Le Thien, Hoang H. Tran, Dung D. Le, Vu N. Duong, Daniel Sonntag, An Thai Le, Duy Minh Ho Nguyen, Vien Anh Ngo, Tran Van Nhiem·
arXiv:2606.28385v1 Announce Type: cross Abstract: Recent advances in robot world models enable synthetic video generation for embodied prediction and planning. However, evaluating these videos is challenging: visually realistic outputs often violate physical laws, temporal consis…
arXiv cs.AI
TIER_1English(EN)·Guanglong Sun, Shuang Cui, Bo Lei, Liyuan Wang, Zihan Zhai, Hongwei Yan, Hang Su, Jun Zhu, Yi Zhong·
arXiv:2606.28719v1 Announce Type: new Abstract: Test-time adaptation (TTA) of vision-language models (VLMs) is essential for their robust deployment in dynamic, real-world environments. However, existing TTA methods often adapt locally without accumulating knowledge over time, or…
arXiv cs.AI
TIER_1English(EN)·Yichen Guo, Kai Tang, Fenglai Lin, Yiding Sun, Dongshuo Zhang, Wenya Wang, Lin William Cong, Shanghang Zhang·
arXiv:2606.29431v1 Announce Type: new Abstract: Despite the impressive capabilities of Large Vision-Language Models (LVLMs), they remain susceptible to hallucination, generating content inconsistent with the input image. Recent studies attribute this to the dominance of language …
arXiv cs.LG
TIER_1English(EN)·Kwun Ho Ngan, Saman Sadeghi Afgeh, Joe Townsend, Artur d'Avila Garcez·
arXiv:2511.16527v2 Announce Type: replace-cross Abstract: Contrastive vision-language models continue to be the dominant approach for image-text retrieval. Contrastive Language-Image Pre-training (CLIP) trains two neural networks to align their image and text embeddings in a shar…
3D HAMSTER framework enhances robot manipulation by integrating a vision-language model with depth encoding to generate metrically accurate 3D trajectories for point cloud-based control policies.
Vision-language models struggle to distinguish between shared and interpreted visual information in dialogue, relying on static map cues rather than dynamic grounding processes.
Modeling the bidirectional correspondence between external sensory stimuli and internal neural activity has emerged as a critical frontier in neuroscience. However, existing approaches predominantly treat brain encoding and decoding as isolated tasks, relying heavily on unimodal …
Improving vision-language models (VLMs) on visual reasoning typically requires retraining or hand-designed prompts and tools. We present Dynamo, a training-free framework that adapts a frozen VLM without any weight updates. On a small labeled training subset, the agent inspects i…
arXiv:2604.00757v2 Announce Type: replace-cross Abstract: Large Vision Language Models show impressive performance across image and video understanding tasks, yet their computational cost grows rapidly with the number of visual tokens. Existing token pruning methods mitigate this…
arXiv:2606.27500v1 Announce Type: cross Abstract: Large Vision-Language Models (LVLMs) specialized in healthcare are emerging as a promising research direction due to their potential impact in clinical and biomedical applications. However, progress is constrained by the scarcity …
arXiv cs.CL
TIER_1English(EN)·Niclas Lietzow, Danielle Bitterman, Carsten Eickhoff, William Rudman, Michal Golovanevsky·
arXiv:2606.28273v1 Announce Type: new Abstract: Vision-language models must reconcile visual evidence with memorized world knowledge when the two conflict. How they resolve this conflict shapes the reliability of multimodal systems, yet prior work characterizes it behaviorally wi…
BrainJanus represents the first unified brain model integrating brain, vision, and language through a shared Omni space, enabling bidirectional mapping between neural activity and sensory stimuli via a tokenized representation and autoregressive architecture.
Vision-language dataset distillation method using rank-aware hyperbolic alignment to optimize synthetic image-text pairs for efficient contrastive model training while preserving modality-specific diversity.
Vision-language models must reconcile visual evidence with memorized world knowledge when the two conflict. How they resolve this conflict shapes the reliability of multimodal systems, yet prior work characterizes it behaviorally without a component-level causal account. We combi…
arXiv:2606.26891v1 Announce Type: cross Abstract: Concept Bottleneck Models (CBMs) promise transparent reasoning by predicting through human-interpretable concepts, yet their effectiveness fundamentally depends on how well visual and textual representations are aligned or matched…
arXiv cs.AI
TIER_1English(EN)·Haoxiang Sun, Tao Wang, Li Yuan, Jian Zhao, Jiancheng Lv·
arXiv:2606.26196v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) have recently made remarkable progress in unifying vision-language understanding and reasoning, especially following the introduction of models such as OpenAI's O-series and DeepSeek's R-se…
arXiv:2505.12343v2 Announce Type: replace-cross Abstract: Despite the impressive capabilities of Large Vision-Language Models (LVLMs), they remain susceptible to hallucinations, where generated content is inconsistent with the input image. Existing training-free hallucination mit…
arXiv cs.CL
TIER_1English(EN)·Byung-Kwan Lee, Ryo Hachiuma, Yong Man Ro, Yu-Chiang Frank Wang, Yueh-Hua Wu·
arXiv:2506.15681v4 Announce Type: replace Abstract: Recent advancements in vision-language models (VLMs) have leveraged large language models (LLMs) to achieve performance on par with closed-source systems like GPT-4V. However, deploying these models in real-world scenarios, part…
DataComp for VLMs (DCVLM) establishes a comprehensive benchmark for evaluating data curation strategies in vision-language models, demonstrating that data mixing rather than filtering significantly improves model performance at scale.
Large Vision-Language Models (LVLMs) specialized in healthcare are emerging as a promising research direction due to their potential impact in clinical and biomedical applications. However, progress is constrained by the scarcity of high-quality medical multimodal data, concerns …
Sparse Autoencoders (SAEs) have shown promise for analyzing language models, but applying them to vision-language models (VLMs) often yields representations that are difficult to use as controllable cross-modal steering directions. We introduce the Joint Sparse Autoencoder (JSAE)…
arXiv cs.AI
TIER_1English(EN)·Ahmad Algadhi, Ahmed Alzuhair, Omar Alkhulaif, Muzammil Behzad·
arXiv:2606.23897v1 Announce Type: cross Abstract: Prompt distillation compresses large vision-language models (VLMs) such as CLIP into lightweight student models by matching teacher predictions on unlabeled domain images. PromptKD (CVPR 2024) established this paradigm with a sing…
4D dynamic scene understanding requires grounding language to a persistent worldline that binds identity, metric 3D motion, and synchronized multi-view 2D projections. Existing paradigms capture only part of this structure: large multimodal models reason over rich visual evidence…
arXiv cs.CL
TIER_1English(EN)·Yusuf Salcan (Computer Vision Group, University of Freiburg, Germany, CRIION-AI Lab, Freiburg, Germany), Simon Ging (Computer Vision Group, University of Freiburg, Germany, Adaptive & Agentic AI), Robin Schirrmeister (Department of Radiology, Medical Cen…·
arXiv:2606.20477v1 Announce Type: cross Abstract: We study how to train visually grounded vision-language models (VLMs) for radiology without manual spatial annotations. We introduce RefRad2D, a large-scale bilingual (German/English) dataset of 1.2M CT and MR image-text pairs der…
arXiv cs.AI
TIER_1English(EN)·Mahshad Lotfinia, Sebastian Ziegelmayer, Lisa Adams, Daniel Truhn, Andreas Maier, Soroosh Tayebi Arasteh·
arXiv:2606.17710v1 Announce Type: cross Abstract: Medical vision-language models report strong chest radiograph accuracy, and this is increasingly read as evidence that they use the image. That inference is unsafe: a model exploiting finding-name priors scores like one that reads…
Medical vision-language models report strong chest radiograph accuracy, and this is increasingly read as evidence that they use the image. That inference is unsafe: a model exploiting finding-name priors scores like one that reads the scan, and no standard benchmark separates the…
arXiv cs.CV
TIER_1English(EN)·Parsa Esmaeilkhani, Longin Jan Latecki·
arXiv:2602.01530v2 Announce Type: replace Abstract: Modern autoregressive Vision-Language Models (VLMs) can generate fluent answers while their visual-token representations become weakly tied to the image regions from which they originate. This limits patch-level explainability: …
arXiv cs.CV
TIER_1English(EN)·Hunter Schofield, Mohammed Elmahgiubi, Mohammad Mahdavian, Richard Shi, Jinjun Shan, Amir Rasouli, Dongfeng Bai·
arXiv:2608.10278v1 Announce Type: new Abstract: Spatial understanding is fundamental to embodied intelligence, underpinning applications such as robotic manipulation, embodied navigation, and autonomous driving. Although recent vision-language models (VLMs) have achieved impressi…
arXiv:2608.10758v1 Announce Type: new Abstract: Vision-language models can describe an image with remarkable accuracy, yet a more fundamental question remains unanswered: what visual information actually drives their answers? In this work, we investigate this question through cau…
arXiv cs.CV
TIER_1English(EN)·Kiet T. Nguyen, Hanbo Shim, Jinwoo Kim, Seunghoon Hong·
arXiv:2608.10864v1 Announce Type: new Abstract: Vision-language models (VLMs) have achieved strong image and video understanding, yet their visual-spatial representations remain geometrically fragile, leading to failures in spatial reasoning needed for embodied AI, robotics, and …
arXiv:2608.10824v1 Announce Type: cross Abstract: Vision-Language-Action(VLA) models map camera images and language instructions directly to motor commands through a single autoregressive transformer. In real-time control, they still spend substantial compute recomputing key-valu…
arXiv:2604.02048v2 Announce Type: replace Abstract: Developing vision-language models (VLMs) that generalize across diverse tasks requires large-scale training datasets with diverse content. In English, such datasets are typically constructed by aggregating and curating numerous …
arXiv:2608.07584v1 Announce Type: new Abstract: Vision-language models (VLMs) have made rapid progress in visual perception and increasingly support real-world tasks that depend on images. Many such tasks, however, require more than rec- ognizing what an image contains: a model m…
arXiv:2608.08060v1 Announce Type: new Abstract: Fine-tuning vision-language models such as CLIP typically requires backpropagation (BP) through the full model, which is infeasible when only forward-pass access is available, as is common for memory-constrained edge devices and pro…
arXiv:2608.07548v1 Announce Type: cross Abstract: Vision-and-Language Navigation in Continuous Environments (VLN-CE) requires agents to make fine-grained navigation decisions under partial observability. However, most existing methods rely on open-loop execution, lacking mechanis…
arXiv:2608.08663v1 Announce Type: cross Abstract: Humans converge on shared names for novel, hard-to-describe objects through repeated interaction, a process psycholinguists call lexical entrainment. Leading vision-language models fail at this: recent empirical work documents tha…
arXiv cs.CV
TIER_1English(EN)·Hongjin Ji, Guoyang Xia, Luoyang Sun, Fangxiang Feng, Lei Ren·
arXiv:2608.09448v1 Announce Type: cross Abstract: Test-time training (TTT) offers a lightweight way to adapt vision--language--action (VLA) policies from unlabeled deployment streams, but it remains difficult to use reliably in closed-loop manipulation. A shared adaptation space …
arXiv:2604.11240v2 Announce Type: replace Abstract: Token pruning has emerged as an effective approach to reduce the substantial computational overhead of Large Vision-Language Models (LVLMs) by discarding less informative visual tokens while preserving performance. However, exis…
arXiv:2608.06729v1 Announce Type: cross Abstract: While Vision-Language-Action (VLA) models have advanced embodied AI, their fundamentally reactive paradigm severely limits performance in partially observable and long-horizon tasks. When restricted to a single wrist-mounted camer…
arXiv:2608.07314v1 Announce Type: cross Abstract: Vision-language-action (VLA) models are commonly adapted to downstream manipulation tasks via supervised fine-tuning (SFT) or online reinforcement learning (RL) post-training. SFT is prone to distribution mismatch, and existing RL…
arXiv cs.CV
TIER_1English(EN)·Harisankar Babu, Benjamin Coors, Christopher Lang, Hendrik Berkemeyer, Tamim Asfour, Simon Foell·
arXiv:2608.07361v1 Announce Type: cross Abstract: Vision-language-action (VLA) models route driving decisions through a deep language model, but it is unclear how much of that depth the action itself requires. We study a representative driving VLA whose entire plan is carried by …
arXiv:2608.05505v1 Announce Type: new Abstract: Acting in a physical scene requires knowing its real later state, not a plausible one. Current evaluations often accept words or a realistic-looking image, so the predicted state is never checked against the true one. We introduce D…
arXiv:2608.05215v1 Announce Type: cross Abstract: Learning manipulation skills from human videos is promising for scalable robot learning. However, the embodiment mismatch between humans and robots makes this challenging. One promising solution is to learn object-centric actionab…
arXiv:2608.05260v1 Announce Type: new Abstract: Contrastive vision-language models such as CLIP and BLIP are typically trained on short image captions, limiting their ability to retrieve images from detailed textual descriptions. While methods such as Long-CLIP extend the token l…
arXiv:2608.05945v1 Announce Type: new Abstract: Test-time adaptation (TTA) can improve the recognition accuracy of vision-language models under distribution shift, but often degrades calibration, making predictive confidence unreliable for downstream decision-making. Many existin…
arXiv cs.CV
TIER_1English(EN)·Xi Xiao, Xingjian Li, Cheng Han, Tianyang Wang, Lin Zhao, Yunbei Zhang, Guosheng Hu, Runmin Jiang, Xi Li, Xiao Wang, Min Xu·
arXiv:2608.05393v1 Announce Type: new Abstract: Prompt tuning, a leading parameter-efficient adaptation paradigm in NLP, has recently been extended to computer vision. Visual prompt tuning (VPT) adapts pre-trained vision transformers (ViTs) by updating a small set of additional p…
arXiv:2512.22969v2 Announce Type: replace Abstract: Conventional object detectors rely on cross-entropy classification, which can be vulnerable to class imbalance and label noise. We propose CLIP-Joint-Detect, a simple and detector-agnostic framework that integrates CLIP-style co…
arXiv:2608.04935v1 Announce Type: new Abstract: Recent work has shown that a simple linear probe on frozen representations from modern vision foundation models (VFMs) can achieve state-of-the-art AIGI detection performance, substantially outperforming specialized detectors in cha…
arXiv:2608.04396v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models have driven significant progress in robotic manipulation, yet they fundamentally struggle with the vision-override phenomenon. Driven by the severe modality imbalance between dense visual streams …
arXiv:2608.03885v1 Announce Type: new Abstract: Vision-language models exhibit remarkable zero-shot capabilities but suffer significant performance degradation under distribution shifts. While test-time adaptation (TTA) via Low-Rank Adaptation offers a parameter-efficient solutio…
arXiv:2608.03649v1 Announce Type: new Abstract: Fewer visual tokens do not guarantee lower end-to-end latency. We evaluate break-even with a reproducible protocol that accounts for decision overhead, shared work, and the operators each policy can avoid. A stage-level decompositio…
arXiv:2608.02790v1 Announce Type: new Abstract: Vision-language models (VLMs), including medical specialists, are increasingly proposed for medical imaging, yet their stated confidence is rarely evaluated separately from correctness. We use brain MRI as a controlled, high-stakes …
arXiv:2608.02980v1 Announce Type: new Abstract: Large Multimodal Models (LMMs) have achieved remarkable success on images and short videos, yet scaling them to long videos remains challenging due to frame-centric tokenization and limited context windows. 3D geometry provides a na…
arXiv cs.CV
TIER_1English(EN)·Zihan Wang, Tong Liu, Zhiwei Wang, Tao Huang, Wentao Jiang, Sihan Ma, Shanshan Ye, Xiaohui Yang, Jing Zhang·
arXiv:2608.03322v1 Announce Type: new Abstract: Medical visual grounding connects free-form clinical queries to spatial evidence in medical images and is an important component of interpretable medical artificial intelligence. However, general-purpose grounding models are predomi…
arXiv:2608.03143v1 Announce Type: new Abstract: Vision-and-Language Navigation (VLN) requires an agent to follow a route-level instruction by executing its constituent steps from egocentric visual observations. Existing VLM-based navigators typically supervise both capabilities t…
arXiv:2608.03580v1 Announce Type: new Abstract: While Vision-Language Models (VLMs) have demonstrated remarkable performance in processing and understanding both text and images, their large parameter sizes lead to significant computational overhead, limiting their deployment on …
arXiv:2608.01985v1 Announce Type: new Abstract: Visual token pruning reduces the computational cost of Vision-Language Models (VLMs) by removing redundant visual tokens. The key is to learn a score that measures whether a token is useful. Existing methods typically rely on Gumbel…
arXiv cs.CV
TIER_1English(EN)·Yanbin Hu, Jin Cui, Jun Ye, Jiepeng Zhou, Jiangcheng Song, Boran Zhao, Pengju Ren·
arXiv:2608.00110v1 Announce Type: new Abstract: 3D scene understanding requires reasoning about entity existence, spatial layout, and object relations, yet RGB images alone often provide insufficient 3D cues. Existing 3D-VLMs commonly rely on depth or 3D-position-aware inputs at …
arXiv:2608.00235v1 Announce Type: new Abstract: Vision-language models (VLMs) have emerged as a powerful framework for multimodal video understanding. However, they remain limited in the sign language translation task, where we identify a key failure mode of existing VLMbased tra…
arXiv:2608.00239v1 Announce Type: new Abstract: Learning transferable representations from CT-report pairs requires combining whole-volume context with anatomy-specific evidence. Existing methods typically emphasize either global CT-report alignment or fine-grained anatomy-level …
arXiv:2608.00574v1 Announce Type: new Abstract: Vision-language MoE batches contain different numbers of image and text tokens. Image resolution, image count, tiling, and prompt length all change this token mix. We call the standard token-level Switch auxiliary loss Std-Aux. Std-…
arXiv:2608.01067v1 Announce Type: new Abstract: Training-free test-time defenses offer a practical way to improve the adversarial robustness of CLIP-style vision--language models without modifying the pretrained model. However, their correction strength is typically fixed for a n…
arXiv cs.CV
TIER_1English(EN)·Puzhuo Zheng, Hasan Kurban·
arXiv:2608.01207v1 Announce Type: new Abstract: Test-time scaling lifts large language model reasoning by sampling many candidate solutions and selecting among them, yet the same recipe transfers poorly to vision-language models (VLMs): recent work shows that simple majority voti…
arXiv:2608.01534v1 Announce Type: new Abstract: Hard symbolic-reasoning tasks such as Sudoku, maze pathfinding, and ARC remain challenging for LLMs due to their fixed-depth autoregressive reasoning, which limits systematic search, refinement, and backtracking. While recursive mod…
arXiv:2608.01614v1 Announce Type: new Abstract: Vision-Language Models (VLMs) face a critical computational bottleneck when processing high-resolution imagery due to the $O(N^2)$ memory complexity of Softmax Multi-Head Attention (MHA). While substituting MHA with independent Mult…
arXiv:2608.01644v1 Announce Type: new Abstract: In video understanding, vision-language models (VLMs) must ingest massive numbers of visual tokens, causing the computational and memory cost of the prefill stage to rise sharply. Such visual sequences are highly redundant along the…
arXiv:2608.01709v1 Announce Type: new Abstract: Vision-language models (VLMs) achieve strong semantic understanding but remain unreliable in metric spatial reasoning, particularly when queries require comparing multiple instances of the same object category. We study this problem…
arXiv cs.CV
TIER_1English(EN)·Hongjie Zhou, Shiqin Wang, Haoyang Chen, Haonan Guo, Di Wang, Juhua Liu, Fu Lin, Yong Luo·
arXiv:2608.02039v1 Announce Type: new Abstract: Remote-sensing videos enable real-time observation of changes in target attributes, short-term activities, and scene evolution. They record motion, actions, interactions, and scene changes that cannot be captured by isolated images.…
arXiv:2608.02137v1 Announce Type: new Abstract: Vision-language models (VLMs) exhibit strong generalization across multimodal tasks but remain vulnerable to adversarial perturbations. Existing attacks typically follow single-trajectory gradient optimization or task-specific objec…
arXiv:2608.02216v1 Announce Type: new Abstract: Vision-language models (VLMs) such as CLIP exhibit remarkable zero-shot capabilities, yet their performance frequently degrades sharply under unexpected test-time distribution shifts. While Test-Time Adaptation (TTA) offers a promis…
arXiv cs.CV
TIER_1English(EN)·Brian Song, Michael A. Lepori, Ellie Pavlick·
arXiv:2608.00035v1 Announce Type: cross Abstract: Goal-directed visual processing is a hallmark of human visual intelligence, resulting in representations that support downstream tasks such as categorization or search. Though vision-language models (VLMs) are often faced with the…
arXiv:2608.01265v1 Announce Type: cross Abstract: Despite recent progress in Vision-Language-Action (VLA) models for robotic manipulation, the action chunk remains a weakly structured interface. Existing work typically flatten each chunk into per-timestep controls, relying on imp…
arXiv:2607.29596v1 Announce Type: cross Abstract: Vision-language-action models (VLAs), which leverage the cognition of multimodal information to infer physical-world actions, provide a generalized solution for embodied AI applications. Conventional VLAs usually concentrate on cu…
arXiv cs.CV
TIER_1English(EN)·Jiasheng Li, Zhong Ji, Yan Zhang, Huihui Li·
arXiv:2607.27700v1 Announce Type: new Abstract: Large Vision-Language Models (VLMs) suffer from prohibitive inference overhead due to long sequences of visual tokens. However, existing visual token reduction methods mainly improve efficiency by pruning or compressing redundant to…
arXiv:2607.27897v1 Announce Type: new Abstract: Vision-language models (VLMs), such as CLIP, are vulnerable to adversarial attacks, posing a serious problem for real-life applications and deployment. Adversarial fine-tuning emerges as a prominent defense method; however, differen…
arXiv cs.CV
TIER_1English(EN)·Ioannis Sarridis, Ioannis Kompatsiaris, Symeon Papadopoulos·
arXiv:2607.28211v1 Announce Type: new Abstract: Vision-Language Models (VLMs) such as CLIP are now foundational to multimodal systems, yet their robustness to spurious correlations remains poorly understood at scale. We present the first large-scale empirical study of 194 publicl…
arXiv:2603.05950v2 Announce Type: replace Abstract: Visual token reduction is critical for accelerating Vision-Language Models (VLMs), since visual inputs are represented as token sequences that introduce substantial computational overhead in the LLM backbone. However, most pruni…
arXiv:2604.08881v2 Announce Type: replace Abstract: With the widespread deployment of vision-language large models (VLLMs), their safety alignment faces dual challenges across languages and modalities. Existing methods model multilingual and multimodal safety separately, overlook…
arXiv:2607.28464v1 Announce Type: new Abstract: Detection and localization of AI-tampered images are critical for trustworthy AI, yet modern generative models have made such manipulations increasingly difficult to identify. While traditional binary classifiers can detect image ta…
arXiv:2607.26554v1 Announce Type: new Abstract: Integrating 3D medical images with vision-language models (VLMs) holds substantial promise for computer-aided diagnosis. However, volumetric images generate prohibitively long visual-token sequences with considerable spatial and int…
arXiv cs.CV
TIER_1English(EN)·Jinkun Zhao, Kui Zhang, Wenjun Wu·
arXiv:2607.26536v1 Announce Type: new Abstract: High-pressure prompts can push vision-language models (VLMs) into unsupported commitments, such as reading illegible text, reporting indeterminate times, or affirming absent objects. This paper asks whether the pressure-induced dist…
arXiv:2512.01008v2 Announce Type: replace Abstract: Text-driven 3D reconstruction requires masks that understand free-form instructions and remain stable under viewpoint changes. We present LISA-3D, a two-stage framework that adapts the instruction-following segmenter LISA with g…
arXiv:2607.27205v1 Announce Type: new Abstract: Vision-language-action (VLA) models commonly adopt an LLM-centric $V \to L \to A$ pathway, where visual observations are projected into the representation space of a large language model before being decoded into robot actions. Alth…
arXiv:2607.27180v1 Announce Type: new Abstract: Evaluating whether a vision-language model (VLM) can act through a physical body is challenging. The outcome of an action couples the VLM's decision with motor control. When a task fails, it is hard to tell whether the VLM made a ba…
arXiv:2607.26885v1 Announce Type: new Abstract: Vision-language pre-training (VLP) serves as a cornerstone for medical multimodal representation learning. However, existing medical VLP frameworks are often constrained by the limited context windows and shallow representational ca…
arXiv cs.CV
TIER_1English(EN)·Jiaang Li, Chengzu Li, Zhaochong An, Yifei Yuan, Xi Liu, Serge Belongie, V\'esteinn Sn{\ae}bjarnarson·
arXiv:2607.26326v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) achieve strong performance by integrating visual inputs with the rich priors of pretrained language models. However, they often fail on vision-centric tasks, especially when visual evidence c…
arXiv:2607.26596v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) have demonstrated remarkable capabilities by integrating visual and textual understanding within a unified transformer architecture. However, fine-tuning all parameters of these models for vi…
arXiv:2607.24957v1 Announce Type: new Abstract: We introduce PerceptionBench, a benchmark specifically designed to evaluate the atomic visual perception capabilities of Multimodal Large Language Models (MLLMs). Existing benchmarks often fail to isolate perception: holistic evalua…
arXiv cs.CV
TIER_1English(EN)·Riccardo Andrea Izzo, Gianluca Bardaro, Matteo Matteucci·
arXiv:2603.05147v2 Announce Type: replace Abstract: Current research on Vision-Language-Action (VLA) models predominantly focuses on enhancing generalization through reasoning techniques. While effective, these improvements increase computational complexity and inference latency.…
arXiv cs.CV
TIER_1English(EN)·Daojie Peng, Fulong Ma, Jun Ma·
arXiv:2603.27577v2 Announce Type: replace Abstract: Vision-Language Navigation (VLN) requires an embodied agent to navigate complex environments by following natural language instructions, which typically demands tight fusion of visual and language modalities. Existing VLN method…
arXiv:2607.24424v1 Announce Type: new Abstract: Vision-language models commonly project all tokens produced by a pretrained vision encoder into a large language model. However, final-layer features can discard text, local attributes, and spatial relationships, while high-resoluti…
arXiv:2607.23504v1 Announce Type: new Abstract: Vision-and-Language Navigation in Continuous Environments (VLN-CE) requires agents to maintain long-horizon visual history for trajectory consistency while executing actions with low latency. Existing video-based VLN approaches typi…
arXiv cs.CV
TIER_1English(EN)·Ioannis Maniadis Metaxas, Adrian Bulat, Alberto Baldrati, Anestis Zaganidis, Yassine Ouali, Hyeonuk Kim, Georgios Tzimiropoulos·
arXiv:2607.23373v1 Announce Type: new Abstract: Large Vision-Language Models (LVLMs) remain bottlenecked by massive computational footprints, precluding their deployment on resource-constrained edge devices. While efforts to compress LVLMs focus heavily on vision token reduction …
arXiv:2607.23335v1 Announce Type: new Abstract: Auxiliary signal pathways in VLMs are routinely fitted with learnable gates so the optimiser can decide how much of the signal to admit. We find that the optimiser almost always decides on zero: across five injection designs, every …
arXiv cs.CV
TIER_1English(EN)·Yihao Wu, Chenyi Xu, Liqi Yan, Chenhuan Cai, Geyong Min, Bin Lin, Fangli Guan, Jianhui Zhang, Pan Li·
arXiv:2607.23181v1 Announce Type: new Abstract: Vision-and-Language Navigation in continuous environments (VLN-CE) requires an agent to ground language in egocentric observations and plan in unseen scenes. Although recent multimodal large models and world-model-based methods have…
arXiv cs.CV
TIER_1English(EN)·Futa Waseda, Saku Sugawara, Isao Echizen·
arXiv:2507.16257v2 Announce Type: replace Abstract: Defending pre-trained vision-language models (VLMs), such as CLIP, against adversarial attacks is crucial, as these models are widely used in diverse zero-shot tasks, including image classification. However, existing adversarial…
arXiv cs.CV
TIER_1English(EN)·Pau de Jorge, C\'esar Roberto de Souza, Bj\"orn Michele, Mert B\"ulent Sar{\i}y{\i}ld{\i}z, Philippe Weinzaepfel, Florent Perronnin, Diane Larlus, Yannis Kalantidis·
arXiv:2604.12935v2 Announce Type: replace Abstract: Efficiently merging several models fine-tuned for different tasks, but stemming from the same pretrained base model, is of great practical interest. Despite extensive prior work, most evaluations of model merging in computer vis…
arXiv:2512.11899v2 Announce Type: replace Abstract: Large vision-language models (LVLMs) are vulnerable to typographic attacks, where misleading text inserted into an image can override visual understanding. However, existing evaluation protocols and defenses are largely focused …
arXiv:2607.18673v1 Announce Type: new Abstract: Vision Language Models (VLMs) are well known for hallucinating non-existent objects in images. Objects with missing parts present a unique challenge for VLMs, stemming from both real-world knowledge bias and the scarcity of such ima…
arXiv:2509.25699v4 Announce Type: replace Abstract: Interleaved-Modal Chain-of-Thought (I-MCoT) advances vision-language reasoning, such as Visual Question Answering (VQA). This paradigm integrates specially selected visual evidence from the input image into the context of Vision…
arXiv:2412.11475v3 Announce Type: replace Abstract: We present OmniVLM, a sub-billion-parameter vision-language model for efficient on-device inference. OmniVLM introduces a token compression mechanism that reduces visual token sequence length from 729 to 81 tokens, significantly…
arXiv:2607.17052v1 Announce Type: new Abstract: Vision-language models normally execute the same complete vision encoder for every question, even when OCR, counting, object, attribute, and spatial queries may not require identical computation. We study whether fixed-budget combin…
arXiv cs.CV
TIER_1English(EN)·Zhifang Zhang, Qiqi Tao, Jiaqi Lv, Na Zhao, Lei Feng, Joey Tianyi Zhou·
arXiv:2509.24566v2 Announce Type: replace Abstract: Large vision-language models (LVLMs) have achieved impressive performance across a wide range of vision-language tasks, while they remain vulnerable to backdoor attacks. Existing backdoor attacks on LVLMs aim to force the victim…
arXiv:2607.14966v1 Announce Type: new Abstract: The prompt learning paradigm for vision-language models is effective yet faces a granularity dilemma: global prompts lack fine-grained semantic awareness, while local prompts ignore contextual associations, limiting cross-task gener…
The prompt learning paradigm for vision-language models is effective yet faces a granularity dilemma: global prompts lack fine-grained semantic awareness, while local prompts ignore contextual associations, limiting cross-task generalization. This dilemma exists in dense predicti…
arXiv:2602.17659v2 Announce Type: replace Abstract: Vision-Language-Action models (VLAs) promise to ground language instructions in robot control, yet in practice often fail to faithfully follow language. When presented with instructions that lack strong scene-specific supervisio…
arXiv:2607.13500v1 Announce Type: new Abstract: Vision-Language Models (VLMs) have achieved strong performance in multimodal understanding, yet remain challenging to deploy on resource-constrained edge devices due to the substantial computational overhead of processing numerous v…
Vision-Language Models (VLMs) have achieved strong performance in multimodal understanding, yet remain challenging to deploy on resource-constrained edge devices due to the substantial computational overhead of processing numerous visual tokens. Token reduction is a promising dir…
arXiv:2607.12429v1 Announce Type: new Abstract: Consistent cross-view understanding under extreme viewpoint changes is essential for spatial intelligence, as it enables models to recognize the same scene across extreme viewpoint gaps. Cross-view localization naturally provides a …
arXiv:2607.12786v1 Announce Type: new Abstract: Cross-image comparative reasoning remains challenging for vision-language models (VLMs), especially when correct prediction requires fine-grained attribute grounding and globally consistent reasoning. We present CoRe, a unified fram…
arXiv cs.CV
TIER_1English(EN)·Xinyue Xu, Zheng Zhang, Kunyang Ma, Ge Zhu, Lianshuai Cao, Lei Wang, Zixuan Li, Yi Cheng·
arXiv:2607.12319v1 Announce Type: new Abstract: As vision-language models (VLMs) are increasingly deployed in geospatial question answering and visual scene understanding, improving their spatial cognition capability on street view imagery for complex logical reasoning has emerge…
arXiv:2607.12680v1 Announce Type: new Abstract: Existing vision-language navigation methods often couple a VLM with waypoint decoders to produce multi-step action plans, but they typically lack an explicit closed-loop mechanism for tracking semantic progress, diagnosing execution…
arXiv cs.CV
TIER_1English(EN)·Sania Waheed, Michael Milford, Sarvapali D. Ramchurn, Shoaib Ehsan·
arXiv:2607.12818v1 Announce Type: new Abstract: Visual place recognition (VPR) is a key enabler of accurate localization and long-term autonomous navigation in robotics applications, such as loop closure detection for simultaneous localisation and mapping (SLAM). However, real-wo…
arXiv cs.CV
TIER_1English(EN)·Jiho Hong, Eunae Kang, Sanghyun Kim, Young-Sik Shin·
arXiv:2607.12630v1 Announce Type: cross Abstract: Visual Language Navigation (VLN) aims to enable an embodied agent to navigate complex environments by following natural language instructions. Recent approaches build semantic spatial maps and leverage Large Language Models (LLMs)…
Visual place recognition (VPR) is a key enabler of accurate localization and long-term autonomous navigation in robotics applications, such as loop closure detection for simultaneous localisation and mapping (SLAM). However, real-world VPR deployment relies on selecting an image …
Cross-image comparative reasoning remains challenging for vision-language models (VLMs), especially when correct prediction requires fine-grained attribute grounding and globally consistent reasoning. We present CoRe, a unified framework for this problem. CoRe includes: (i) CoRe-…
Existing vision-language navigation methods often couple a VLM with waypoint decoders to produce multi-step action plans, but they typically lack an explicit closed-loop mechanism for tracking semantic progress, diagnosing execution failures, and recovering from error accumulatio…
Visual Language Navigation (VLN) aims to enable an embodied agent to navigate complex environments by following natural language instructions. Recent approaches build semantic spatial maps and leverage Large Language Models (LLMs) for reasoning and decision making. Despite these …
Consistent cross-view understanding under extreme viewpoint changes is essential for spatial intelligence, as it enables models to recognize the same scene across extreme viewpoint gaps. Cross-view localization naturally provides a promising pathway toward this ability, as it req…
arXiv:2607.10797v1 Announce Type: new Abstract: Assembly action understanding is a key enabler for effective human-robot collaborative assembly, yet it remains challenging due to subtle motions and fine-grained hand-object interactions. We adapt vision-language models (VLMs) to t…
arXiv:2607.10640v1 Announce Type: new Abstract: Vision-Language Models (VLMs) are costly at inference time because they must process long sequences of visual tokens. Existing token pruning methods often degrade under high compression by blindly discarding information, breaking sp…
arXiv:2607.11173v1 Announce Type: new Abstract: Vision-language models (VLMs) are expected to reason about physical space -- which object is closer, what lies behind what, and how objects are arranged in 3D -- yet they still struggle with such spatial judgments. A natural remedy …
arXiv:2607.10796v1 Announce Type: new Abstract: Large Vision Language Models (LVLMs) require strong reasoning over both visual and textual input. Recent work suggests that cognitive elements, especially diverse representations and metacognition, correlate with better performance.…
As vision-language models (VLMs) are increasingly deployed in geospatial question answering and visual scene understanding, improving their spatial cognition capability on street view imagery for complex logical reasoning has emerged as a key research priority. However, existing …
Vision-language models (VLMs) are expected to reason about physical space -- which object is closer, what lies behind what, and how objects are arranged in 3D -- yet they still struggle with such spatial judgments. A natural remedy is to show the model a depth map, but we find th…
arXiv cs.CV
TIER_1English(EN)·Yuncheng Yang, Feiyang Ye, Shixian Luo, Yinna Zhu, Lianlei Shan, Wangcai Zhao, Kuo Zhang, Yan Chen, Yong Wu, Yan Xie·
arXiv:2607.09029v1 Announce Type: new Abstract: Vision-Language Models (VLMs) have achieved success using homogeneous Transformers to process multimedia data. Recent studies show that heterogeneous structures interleaving efficient mechanisms, like linear attention, improve both …
arXiv:2607.09008v1 Announce Type: new Abstract: Safety-critical perception systems must reliably detect rare object classes within small label spaces, a setting that long-tailed detection methods, designed for hundreds of classes with dense annotation, fundamentally do not addres…
arXiv:2507.05240v2 Announce Type: replace-cross Abstract: Vision-and-Language Navigation (VLN) in real-world settings requires agents to process continuous visual streams and generate actions with low latency grounded in language instructions. While Video-based Large Language Mod…
arXiv cs.CV
TIER_1English(EN)·Zuntao Liu, Yi Du, Taimeng Fu, Shaoshu Su, Cherie Ho, Chen Wang·
arXiv:2511.20644v2 Announce Type: replace Abstract: Spatial reasoning is a critical capability for intelligent robots, yet current vision-language models (VLMs) still fall short of human-level performance in video-based spatial reasoning. This gap mainly stems from two challenges…
arXiv:2607.08194v1 Announce Type: new Abstract: Vision-language alignment, the stage that bridges pretrained vision encoders and large language models, is widely treated as a form of pretraining requiring full-parameter updates. We challenge this view and investigate what happens…
Vision-Language Models (VLMs) have achieved success using homogeneous Transformers to process multimedia data. Recent studies show that heterogeneous structures interleaving efficient mechanisms, like linear attention, improve both performance and inference latency over homogeneo…
Safety-critical perception systems must reliably detect rare object classes within small label spaces, a setting that long-tailed detection methods, designed for hundreds of classes with dense annotation, fundamentally do not address. Open-vocabulary detectors offer a promising a…
Vision-language alignment, the stage that bridges pretrained vision encoders and large language models, is widely treated as a form of pretraining requiring full-parameter updates. We challenge this view and investigate what happens when low-rank adaptation is applied to the LLM …
Long-horizon robot planning requires jointly reasoning over semantic task structure and geometric feasibility. To successfully execute a task, a robot must decompose goals, select task-relevant objects, and sequence actions, while ensuring that plans satisfy spatial constraints s…
Mainstream Vision-Language-Action (VLA) models predict actions primarily from the current observation under a Markovian assumption, thus struggling with long-horizon, temporally dependent tasks. Existing memory-augmented VLAs either expand the observation window or retrieve histo…
arXiv cs.CV
TIER_1English(EN)·Jiho Choi, Jaemin Kim, Sanghwan Kim, Seunghoon Hong, Jin-Hwi Park·
arXiv:2604.03316v3 Announce Type: replace Abstract: Attention sinks are defined as tokens that attract disproportionate attention. While these have been studied in single modality transformers, their cross-modal impact in Large Vision-Language Models (LVLM) remains largely unexpl…
arXiv:2510.13808v2 Announce Type: replace Abstract: Large Vision Language Models (VLMs) excel at general visual reasoning but experience significant performance degradation when deployed in novel domains that exhibit substantial distribution shifts from their pretraining data. Ex…
arXiv:2607.06185v1 Announce Type: new Abstract: Fine-grained image recognition poses a significant challenge due to the substantial expertise and effort required for manual annotation. Vision-language models (VLMs) like CLIP provide a compelling zero-shot alternative, reducing re…
arXiv cs.CV
TIER_1English(EN)·Zhiwei Yang, Yuanchen Wu, Nan Zhang, Yucong Meng, Ke Yan, Shouhong Ding·
arXiv:2607.05716v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) have demonstrated strong perception and reasoning capabilities. However, most existing models focus on isolated objects and neglect structured relationships for efficient target navigation, l…
Fine-grained image recognition poses a significant challenge due to the substantial expertise and effort required for manual annotation. Vision-language models (VLMs) like CLIP provide a compelling zero-shot alternative, reducing reliance on extensive labeled data. However, their…
arXiv cs.CV
TIER_1English(EN)·Zaiyu Cheng, Khai-Nguyen Nguyen, Antonio Mastropaolo·
arXiv:2607.02853v1 Announce Type: new Abstract: Vision Language Models (VLMs) are increasingly applied to software engineering artifacts, especially UML class diagrams whose meaning depends on visual notation. Yet, it is unclear whether VLMs actually read such diagrams or instead…
arXiv:2602.04356v2 Announce Type: replace Abstract: Targeted adversarial attacks on Large Vision-Language Models (LVLMs) test whether small image perturbations can steer model responses toward attacker-specified content. Under the standard L-infinity constraint, targeted attacks …
arXiv cs.CV
TIER_1English(EN)·Zhaoxu Li, Chenqi Kong, Yi Yu, Qiangqiang Wu, Xinghao Jiang, Ngai-Man Cheung, Bihan Wen, Alex Kot, Xudong Jiang·
arXiv:2508.03177v2 Announce Type: replace Abstract: Large Vision-Language Models (LVLMs) recently achieve significant breakthroughs in understanding complex visual-textual contexts. However, hallucination issues still limit their real-world applicability. Although previous mitiga…
arXiv:2607.03792v1 Announce Type: cross Abstract: Vision-and-Language Navigation (VLN) agents may satisfy conventional success criteria while still failing to establish reliable object-level grounding, because current evaluation protocols mainly reward stopping within a 3-meter r…
arXiv:2607.05264v1 Announce Type: new Abstract: Existing video benchmarks evaluate action recognition on consumer videos, egocentric recordings, or simulated industrial environments. They do not test vision-language models under the visual and procedural conditions of real indust…
arXiv:2607.04637v1 Announce Type: new Abstract: Vision-Language-Action Models (VLAs), which leverage the advanced reasoning capabilities of Vision-Language Models (VLMs), show promising generalization in complex autonomous driving scenarios. Existing VLAs typically predict and op…
arXiv:2607.04057v1 Announce Type: new Abstract: Robots deployed over long periods must reason about environments that change over time. Existing long-term perception systems often address object change reactively, updating their maps only after revisiting a scene and observing th…
arXiv:2607.03647v1 Announce Type: new Abstract: Large vision language models (VLMs) report strong accuracy on medical question-answering, yet it remains unclear whether they reason from visual evidence or exploit textual shortcuts. We introduce a counterfactual evaluation framewo…
arXiv:2607.03184v1 Announce Type: new Abstract: While Multimodal Large Language Models (MLLMs) demonstrate impressive general capabilities, they struggle with fine-grained perception in ultra-high-resolution (UHR) images, particularly for tiny objects in cluttered scenes. Existin…
arXiv:2607.02961v1 Announce Type: new Abstract: Large Vision-Language Models (LVLMs) exhibit remarkable vision-language capabilities and are increasingly deployed in real-world applications such as personal assistants, document analysis systems, and embodied agents. However, thei…
arXiv cs.CV
TIER_1English(EN)·Marwane Hariat, David Filliat, Antoine Manzanera·
arXiv:2607.02707v1 Announce Type: new Abstract: Feed-forward 3D models are commonly trained using either expensive geometric supervision or self-supervised photometric objectives, both of which provide incomplete learning signals. We introduce Vision-Language Reprojection Consist…
Existing video benchmarks evaluate action recognition on consumer videos, egocentric recordings, or simulated industrial environments. They do not test vision-language models under the visual and procedural conditions of real industrial CCTV, where workers appear as distant figur…
arXiv:2601.10710v2 Announce Type: replace Abstract: Vision-Language Models (VLMs) create a severe visual feature bottleneck by using a crude, asymmetric connection that links only the output of the vision encoder to the input of the large language model (LLM). This static archite…
arXiv cs.CV
TIER_1English(EN)·Yiqian Liu, Iuliia Kotseruba, John K. Tsotsos·
arXiv:2607.01503v1 Announce Type: new Abstract: In this paper, we study depth perception of vision-language models (VLMs) to isolate the effects of pictorial depth cues and disentangle vision and language influences on model performance. To this end, we combine depth-ordering and…
arXiv cs.CV
TIER_1English(EN)·Lev Sorokin, Chen Yang, Ken E. Friedl, Andrea Stocco·
arXiv:2607.02300v1 Announce Type: new Abstract: In the automotive domain, in-car scene understanding (ISU) enables the detection of safety-critical events, such as driver distraction, and supports drivers or passengers by analyzing the in-car scene and adapting the environment (e…
arXiv:2508.04101v2 Announce Type: replace Abstract: Computer-aided medical image analysis is crucial for disease diagnosis and treatment planning. While vision-language models (VLMs) such as CLIP exhibit strong generalization ability, their direct application to medical imaging r…
In the automotive domain, in-car scene understanding (ISU) enables the detection of safety-critical events, such as driver distraction, and supports drivers or passengers by analyzing the in-car scene and adapting the environment (e.g., ambient lighting). The industry is increasi…
arXiv cs.CV
TIER_1English(EN)·Sangyun Chung, Youngjoon Yu, Se Yeon Kim, Youngchae Chee, Yong Man Ro·
arXiv:2412.20750v3 Announce Type: replace Abstract: Large-scale Vision-Language Models (VLMs) have achieved notable progress in aligning visual inputs with text. However, their ability to deeply understand the unique physical properties of non-RGB vision sensor images remains lim…
arXiv cs.CV
TIER_1English(EN)·Diogo Gl\'oria-Silva, Jo\~ao Cardeira, Manuel Letras da Luz, Afonso Simpl\'icio, Gon\c{c}alo Vinagre, Diogo Tavares, Rafael Ferreira, In\^es Calvo, In\^es Vieira, David Semedo, Jo\~ao Magalh\~aes·
arXiv:2606.19100v3 Announce Type: replace Abstract: Large Vision and Language Models (LVLMs) have advanced rapidly, yet European Portuguese (pt-PT) remains systematically underserved by existing open-source multimodal models, which either conflate it with Brazilian Portuguese or …
arXiv:2607.00338v1 Announce Type: new Abstract: Object detection for Unmanned Aerial Vehicles (UAVs) working in open and dynamic environments is a highly challenging task. While Vision-Language Models (VLMs) have offered a powerful solution for universal object detection, adaptin…
arXiv:2607.00357v1 Announce Type: new Abstract: Personalized object localization (POL) localizes an object instance in a query image based on a few reference images with bounding-box annotations and a target object label. The pioneering method, IPLoc, solves this task through in-…
arXiv cs.CV
TIER_1English(EN)·Kyan Mahajan, Mohammad Saqlain·
arXiv:2607.00780v1 Announce Type: new Abstract: Most adaptive-inference techniques for foundation models change what the model does - early exit, MoE routing, KV-cache compression, dynamic attention sparsity. The input that hits the backbone, however, remains a fixed-grid tokenis…
Most adaptive-inference techniques for foundation models change what the model does - early exit, MoE routing, KV-cache compression, dynamic attention sparsity. The input that hits the backbone, however, remains a fixed-grid tokenisation indifferent to image content. We argue tha…
Conformal predictions have attracted significant attention in the field of uncertainty quantification, mainly because of their strong marginal coverage guarantees. Full conditional guarantee is not an attainable goal, a well known fact in conformal predictions literature. As a re…
Open-vocabulary 3D scene graph methods typically operate in two stages: first reconstruct, then enrich with vision-language models, leaving the graph unqueryable during exploration. We argue that this sequential coupling is unnecessary and propose an asynchronous architecture in …
arXiv:2606.28864v1 Announce Type: new Abstract: Test-time scaling is a paradigm where large models use additional compute at inference to achieve better performance, without changing model weights. While it has been widely studied for Large Language Models (LLMs), its applicabili…
arXiv cs.CV
TIER_1English(EN)·Enshen Zhou, Yibo Li, Jingkun An, Jiayuan Zhang, Shanyu Rong, Mengzhen Liu, Yi Han, Yuheng Ji, Huajie Tan, Jiawei He, Pengwei Wang, Zhongyuan Wang, Cheng Chi, Lu Sheng, Shanghang Zhang·
arXiv:2512.13660v3 Announce Type: replace-cross Abstract: Spatial tracing, as a fundamental embodied interaction ability for robots, is inherently challenging as it requires multi-step metric-grounded reasoning compounded with complex spatial referring and real-world metric measu…
arXiv cs.CV
TIER_1English(EN)·Wenhui Liao, Hongliang Li, Pengyu Xie, Xinyu Cai, Yufan Shen, Yi Xin, Qi Qin, Shenglong Ye, Tianbin Li, Ming Hu, Junjun He, Yihao Liu, Wenhai Wang, Min Dou, Bin Fu, Botian Shi, Yu Qiao, Lianwen Jin·
arXiv:2602.12957v3 Announce Type: replace Abstract: Document parsing is a fundamental task in multimodal understanding, supporting a wide range of downstream applications such as information extraction and intelligent document analysis. Benefiting from strong semantic modeling an…
arXiv:2512.10310v2 Announce Type: replace Abstract: While Multimodal Large Language Models (MLLMs) have demonstrated significant promise in Vision-Language Navigation (VLN), existing agents remain heavily constrained by systemic bottlenecks across inference, training, and data co…
arXiv cs.CV
TIER_1English(EN)·Atif Belal, Heitor R. Medeiros, Marco Pedersoli, Eric Granger·
arXiv:2510.00458v3 Announce Type: replace Abstract: Vision-language object detectors (VLODs) such as YOLO-World and Grounding DINO exhibit strong zero-shot generalization, but their performance degrades under distribution shift. Test-time adaptation (TTA) offers a practical way t…
arXiv:2606.30168v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) often fail in fine-grained visual reasoning, as question-relevant visual cues are diluted by dense and redundant image tokens. Recent multimodal reasoning methods usually extend chain-of-thou…
arXiv cs.CV
TIER_1English(EN)·Hong-Han Wang, Yuntao Wang, Hu Ding·
arXiv:2606.29462v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) inherit rich relational priors from their language backbones, yet often fail when asked to apply these relationships in visual contexts. We trace this failure to a structural blind spot: proj…
Multimodal large language models (MLLMs) often fail in fine-grained visual reasoning, as question-relevant visual cues are diluted by dense and redundant image tokens. Recent multimodal reasoning methods usually extend chain-of-thought from language models into visual or latent s…
arXiv cs.CV
TIER_1English(EN)·Jiajun Cheng, Xianwu Zhao, Sainan Liu, Xiaofan Yu, Ravi Prakash, Patrick J. Codd, Jonathan Elliott Katz, Shan Lin·
arXiv:2505.10764v4 Announce Type: replace Abstract: Innovations in digital intelligence are transforming robotic surgery with more informed decision-making. Real-time awareness of surgical instrument presence and actions (e.g., cutting tissue) is essential for such systems. Yet, …
arXiv:2606.27646v1 Announce Type: new Abstract: Conventional machine-vision pipelines typically rely on high-quality optics that produce clean, human-interpretable images, and optical design has therefore been driven by image-level criteria such as resolution, aberration correcti…
arXiv:2606.27660v1 Announce Type: new Abstract: Vision-Language Models (VLMs) improve generalization and interpretability in autonomous driving but suffer from efficiency issues due to long visual token sequences, particularly in standard multi-view settings. Existing token pruni…
arXiv cs.CV
TIER_1English(EN)·Chaoxiang Cai, Minghe Weng, Jie Li, Yibo Jiang, Longrong Yang, Zequn Qin, Xi Li·
arXiv:2606.26287v1 Announce Type: new Abstract: With the increase in model parameters and training data, the instruction following and generalization capabilities of Large VisionLanguage Models (LVLMs) have been significantly improved. Based on the Mixture of Experts (MoE) archit…
Vision-Language Models (VLMs) improve generalization and interpretability in autonomous driving but suffer from efficiency issues due to long visual token sequences, particularly in standard multi-view settings. Existing token pruning methods employ fixed pruning rate allocation …
Conventional machine-vision pipelines typically rely on high-quality optics that produce clean, human-interpretable images, and optical design has therefore been driven by image-level criteria such as resolution, aberration correction, and pixel fidelity. However, such optics are…
Concept Bottleneck Models (CBMs) promise transparent reasoning by predicting through human-interpretable concepts, yet their effectiveness fundamentally depends on how well visual and textual representations are aligned or matched. Existing vision-language CBMs often rely on pre-…
arXiv:2606.25657v1 Announce Type: new Abstract: Sparse Autoencoders (SAEs) have shown promise for analyzing language models, but applying them to vision-language models (VLMs) often yields representations that are difficult to use as controllable cross-modal steering directions. …
arXiv cs.CV
TIER_1English(EN)·Qitong Wang, Fan Du, Pranav Maneriker, Jihui Jin, Christopher Rasmussen·
arXiv:2606.25160v1 Announce Type: cross Abstract: The rapid rise of Vision-Language Models (VLMs) in egocentric visual understanding has made low-latency inference in human-robot collaborative (HRC) tasks increasingly critical. Weight pruning techniques developed for VLMs to shri…
The rapid deployment of Vision-Language Models (VLMs) in dynamic environments necessitates the ability to learn continuously without forgetting. However, traditional continual learning (CL) settings often rely on white-box paradigms, which is increasingly invalidated by the shift…
We study how to train visually grounded vision-language models (VLMs) for radiology without manual spatial annotations. We introduce RefRad2D, a large-scale bilingual (German/English) dataset of 1.2M CT and MR image-text pairs derived from clinical practice, with task-specific VQ…
<h4><em>A research grounded deep dive</em></h4><h3>1. The Hook</h3><p>Picture an ordinary photo like this one. A vision language model will confidently caption it “a dog in a park,” and it will be right, almost every time [1].</p><figure><img alt="A golden retriever sits on green…
<p><em>A daily deep dive into llm topics, coding problems, and platform features from <a href="https://pixelbank.dev" rel="noopener noreferrer">PixelBank</a>.</em></p> <h2> Topic Deep Dive: Vision-Language Models </h2> <p><em>From the Multimodal LLMs chapter</em></p> <h2> Introdu…
dev.to — LLM tag
TIER_1English(EN)·Vahid Aghajani·
<blockquote> <p>Originally published on <a href="https://software-engineer-blog.com/content/vision-language-models-when-ai-learns-to-see-and-talk-part-3-of-3?id=51" rel="noopener noreferrer">my blog</a>. Cross-posted here with a canonical link.</p> </blockquote> <p> </p> <p>This …
2/ Generative AI: Powered by Chery's integration with large language and vision models, allowing the robot to understand natural speech, process visual environments, and carry on human-like conversations. Combined with Ai (Artificial Intelligence), the name essentially translates…
VAORA aligns vision-language model reasoning with physical actions VAORA, a new reward design on arXiv, targets hallucinated reasoning and reasoning-action misalignment in vision-language models on physical tasks. https://www. notatechguy.com/vaora-aligns-v ision-language-model-r…
