Brief

Researchers have developed a novel architecture called TIDAL-Net for physical neural networks (PNNs) to address their limited scale compared to digital neural networks. This new design reuses layers over time, effectively increasing the network's depth without a proportional increase in hardware cost. Experiments demonstrate that TIDAL-Net enhances performance on image classification and natural language processing tasks with minimal changes to existing PNN prototypes. AI

Researchers have introduced OmniLoc, a novel foundation model designed for anchor-free localization within diverse indoor environments using wireless measurements. This model addresses the challenges posed by varying building geometries and signal heterogeneities that hinder existing methods. OmniLoc employs a unified input tokenization, a geometry-aware Transformer for feature extraction, and a location estimation module that leverages geometric embeddings for consistent predictions. Evaluations on in-house and public datasets demonstrate OmniLoc's superior performance and generalization capabilities compared to current approaches. AI

Researchers have developed a new method for predicting outcomes in regression problems involving circular data, such as time of day or direction. This approach utilizes conformal prediction techniques to generate prediction sets with guaranteed coverage and adaptive arc lengths. By projecting existing linear-response regression models onto a circular space, the method can leverage high-performance models designed for linear data. AI

Researchers have developed a new method for predicting the behavior of large reasoning models (LRMs) by training specialized "Behavior Forecasters." These forecasters learn directly from a model's reasoning trajectory, bypassing the need for traditional explanations. The approach proved more accurate than existing models like GPT-5.4 and Claude Opus-4.6 in predicting answer repetition and the impact of input changes, while also being more cost-efficient. AI

Researchers have developed PRInTS, a new generative reward model designed to improve AI agents' ability to seek information over long periods. Unlike previous models that offered binary judgments on short tasks, PRInTS provides dense, multi-dimensional scoring for each step, considering factors like tool interpretation and output informativeness. It also compresses long contexts into summaries while retaining essential information for evaluation. Experiments on benchmarks like FRAMES and GAIA show that PRInTS significantly enhances information-seeking capabilities in various agents, even outperforming larger, frontier models. AI

Researchers have developed a hybrid neural network architecture, KAN-MLP-Mixer, that combines the precision of Kolmogorov-Arnold Networks (KANs) with the noise robustness and efficiency of Multi-Layer Perceptrons (MLPs). This approach strategically integrates KAN modules for input embedding and classification, while utilizing MLPs for intermediate feature mixing. Tested across eight public datasets, the KAN-MLP model demonstrated a 5.33% average improvement in macro F1 score over pure-MLP models, significantly outperforming standalone KAN and MLP baselines. AI

Researchers have developed a new weighted loss objective for neural networks to improve the detection of rare nodes in hierarchical multi-label learning. This approach combines node-wise imbalance weighting with focal weighting components, which leverage ensemble uncertainties. The method aims to address the challenge of fine-grained classifications by emphasizing rare nodes and focusing on uncertain nodes during training. Experiments on benchmark datasets showed improvements in recall by up to five times and statistically significant gains in F1 score. AI

Researchers have identified a "Latent Color Subspace" within the latent space of the FLUX.1 text-to-image model. This subspace reflects Hue, Saturation, and Lightness, offering a new understanding of how semantic information is encoded in image generation. The team demonstrated that this subspace can be used to predict and control image colors without requiring additional training, solely through manipulation of the latent space. AI

Researchers have developed LifeSentence, a novel model that adapts large language models for analyzing longitudinal human life course data. This model can encode complex life trajectories from limited panel data by treating life events as natural-language records. LifeSentence demonstrates superior performance compared to traditional statistical methods and other deep learning approaches, showing significant improvements in predicting events and their timing, as well as reconstructing chronological order. Notably, it can identify social stratification patterns like the gender wage gap and motherhood penalty without explicit supervision, offering new avenues for biographical research and counterfactual exploration. AI

Researchers have developed ProHiFlo, a new hierarchical flow matching framework for de novo protein generation. This method improves efficiency and accuracy by modeling backbone geometry before refining to all-atom coordinates. ProHiFlo also incorporates functional guidance using pretrained predictors to steer generation toward desired properties without retraining. Experiments show state-of-the-art performance, including a higher success rate in enzyme active site scaffolding compared to existing methods. AI

Researchers have introduced Vector Quantized Latent Concept (VQLC), a new framework for interpreting large language models by extracting latent concepts from their hidden states. This method aims to overcome the limitations of existing clustering techniques, which either scale poorly or produce less coherent concepts. VQLC offers a computationally efficient and scalable alternative that demonstrates competitive faithfulness and interpretability, particularly for decoder-only models. AI

Researchers have developed a new method called Cross-modal Consistency Guided Classifier-Free Guidance (CCG-CFG) to improve emotion control in auto-regressive Text-to-Speech (TTS) models. This technique dynamically adjusts guidance scales based on the conflict between textual and desired speech emotions, enhancing emotional alignment. When applied to the CosyVoice2 model, this approach led to significant improvements in emotion recognition accuracy and subjective quality scores, outperforming existing methods like HierSpeech++ and Qwen3-TTS. AI

Researchers have developed DeepRHP, a hybrid variational autoencoder designed to aid in the creation of synthetic random heteropolymers that can mimic protein functions. This model uses a semi-supervised framework, incorporating both chemical features and sequence patterns within its latent space. DeepRHP's effectiveness was demonstrated by successfully predicting monomer compositions that stabilize membrane proteins, with predictions validated against existing research. AI

Researchers have developed a new pretraining framework called Probabilistic Contrastive Pretraining (PCP) to enhance the prediction of ADME properties crucial for drug discovery. This method combines chemistry-specific self-supervision with contrastive mutual information learning, encoding molecular graphs into latent variables and reconstructing SMILES strings. The framework integrates reconstruction, contrastive discrimination, and chemistry-specific tasks into a single probabilistic objective, showing significant improvements over existing baselines on multiple datasets. AI

Researchers have developed a new method called LLM-GNN Co-Teaching to improve few-shot graph learning. This approach avoids designating one model as a "golden teacher," instead allowing a Graph Neural Network (GNN) and a Large Language Model (LLM) to learn collaboratively. The models exchange confident pseudo-labels and update each other, with supervision derived from their agreement over time. This co-teaching framework consistently outperforms previous methods on six benchmarks, showing significant gains in accuracy for tasks like node classification. AI

Researchers have developed SPADE, a novel autoregressive transformer model designed for simulating high-granularity calorimeter data in particle physics. Unlike previous methods that embed multiple features jointly, SPADE embeds them independently and introduces a delay between feature streams. This approach allows the standard self-attention mechanism to learn intra-token correlations effectively. SPADE demonstrates competitive performance against existing models for photon shower generation in the ILD detector and offers a new pathway for applying LLM-style pretraining to complex, multi-feature datasets. AI

Researchers have developed PT-WNO, a novel architecture for 3D point cloud semantic segmentation that enhances global context understanding. The model integrates a Wavelet Neural Operator (WNO) alongside a point cloud transformer backbone. This WNO branch captures multi-scale global spectral context through wavelet decomposition and reconstruction, complementing existing skip connections. Experiments show PT-WNO improves performance on benchmarks like S3DIS and DALES. AI

Researchers have developed three zero-shot auxiliary reasoning methods to improve the ability of vision-language models (VLMs) to ground themselves within graphical user interfaces (GUIs). These methods involve providing explicit spatial cues like axes, grids, and labeled intersections within the input image, enabling VLMs to better articulate their implicit spatial understanding without costly fine-tuning. Experiments across four GUI grounding benchmarks and seven VLMs demonstrated significant performance gains, with one method, Mark-Grid Scaffold, boosting Gemini-3.1-Pro's accuracy on ScreenSpot-v2 from 11.72% to 95.20% and achieving state-of-the-art results on ScreenSpot. AI

Researchers have developed a novel method to imbue large language models with specific personality traits without requiring model retraining. This approach involves identifying key neurons within the LLM that correlate with personality dimensions, based on the Big Five personality traits framework. By manipulating these identified neurons, the system can induce desired personality characteristics in the model's output, demonstrating comparable effectiveness to fine-tuned models but with greater efficiency and flexibility. AI

Researchers have introduced TouchThinker, a new framework designed to enhance tactile commonsense reasoning for embodied agents. This system addresses limitations in existing datasets and representation methods by introducing a million-scale dataset, TouchThinker-1M, covering 415 objects and various scenarios. Additionally, it incorporates an action-aware modeling mechanism to improve the efficiency and semantic expressiveness of tactile representations, enabling better open-world generalization. AI

Researchers have developed a new knowledge distillation framework that uses Large Language Models (LLMs) to extract physics principles from scientific literature. This framework creates a 'teacher' model that imbues a 'student' model with predictive capabilities for manufacturing processes, even with limited data. The resulting student model is lightweight, capable of high-frequency inference for real-time deployment, and shows robustness even when the LLM-derived physics knowledge is imperfect. AI

Researchers have developed MLaGA, a novel model designed to enhance Large Language Models' (LLMs) ability to process and reason over multimodal graphs. This system addresses the challenge of graphs containing diverse attribute types, such as text and images, which have been underexplored by existing LLM-based graph methods. MLaGA employs a structure-aware multimodal encoder and a multimodal instruction-tuning approach to integrate these varied attributes and graph structures into LLMs. AI

Researchers have introduced a novel framework for estimating precision and recall curves in generative models, moving beyond single scalar metrics. This approach frames the estimation as a binary classification problem, offering a more detailed analysis of model performance. The framework also provides a minimax upper bound on estimation risk and unifies several existing precision-recall metrics. AI

Researchers have introduced OMAD, a novel framework for online multi-agent reinforcement learning (MARL) that utilizes diffusion policies to enhance agent coordination. This approach addresses the challenge of intractable likelihoods in diffusion models, which typically hinder exploration in online MARL settings. OMAD employs a relaxed policy objective that maximizes scaled joint entropy and a joint distributional value function for decentralized policy optimization, leading to significant improvements in sample efficiency. AI

Researchers have developed a theoretical framework to unify the understanding of learning dynamics and generalization in transformer models. This work formalizes transformer training as an ordinary differential equation system, approximating it to kernel behaviors. The analysis reveals a two-stage scaling law for generalization error, with an initial exponential decay followed by a power-law decay after a resource threshold is met, proving this two-stage law to be tight. AI

Researchers have developed a new technique called Neural FOXP2 to improve the performance of large language models in non-English languages. This method works by identifying and steering "language neurons" within the model, which are responsible for controlling language defaultness. The process involves localizing these neurons, defining steering directions, and then applying targeted activation shifts to make languages like Hindi or Spanish primary, thereby reducing English dominance. AI

Researchers have developed LatticeBridge, a novel method for structured sequence generation that addresses the challenge of satisfying multiple input-derived constraints within a single output. This approach frames the problem as a rare-event sequential inference task, combining a prefix language model with instance-compiled surface automata and a specialized Monte Carlo decoder. LatticeBridge aims to improve the faithfulness of generated sequences by ensuring all required anchors are jointly realized, outperforming baseline methods on benchmarks like CommonGen and WikiBio. AI

Researchers have developed VL-DINO, a new object detection model that effectively integrates knowledge from CLIP, a vision-language model. The model uses novel modules to construct better training samples and fuse visual and textual information. In zero-shot tests on the LVIS benchmark, VL-DINO achieved state-of-the-art results, outperforming previous methods. AI

Researchers have developed Gumbel-BEARD, a novel framework designed to improve the performance of speech foundation models in low-resource domains. This method automates the selection of Whisper encoder layers using a trainable Gumbel-Softmax selector and a self-supervised adaptation objective. Experiments show that Gumbel-BEARD can match fully supervised baselines with significantly less labeled data and establishes new state-of-the-art word error rates on challenging datasets like MyST and CORAAL. AI

Researchers have developed Lung-R1, a novel large language model designed for pulmonary disease diagnosis. This model is guided by LungKG, a comprehensive knowledge graph containing over 59,000 nodes and 164,000 edges related to pulmonary medicine. Lung-R1 demonstrated state-of-the-art performance in a 20-system evaluation, particularly in EMR diagnosis, outperforming previous baselines. AI

Researchers have developed LakeFM, a new foundation model designed to understand and forecast aquatic ecosystem dynamics. Unlike previous models, LakeFM can handle irregular time series data and generalize across lakes with varying characteristics. It was pre-trained on a large dataset of simulated and observed lakes, demonstrating strong forecasting performance and the ability to produce physically plausible predictions. AI

Researchers have developed GLACIER, a novel student-teacher framework designed for molecular property prediction. This model integrates multiple data types, including molecular graphs, SMILES strings, and physicochemical descriptors, to create more robust and efficient molecular embeddings. By employing a three-stage process involving pretraining student encoders, fusing modalities with a Finsler geometry-aware module, and distilling knowledge from larger teacher models, GLACIER achieves high predictive performance while reducing computational burden. AI

A new research paper explores the scaling laws of data-driven global weather models, analyzing how performance relates to model size, dataset size, and compute budget. The study found that weather models favor wider architectures over deeper ones and that increasing training data yields greater performance gains than increasing model size under fixed compute budgets. Specifically, the Aurora model showed strong data-scaling behavior, with a 10x increase in training data leading to a 3.2x reduction in validation loss. AI

Researchers have developed a Lie-algebraic framework to analyze the expressivity and error bounds of parallelizable sequence models like Transformers. Their theory establishes a direct link between a model's depth and its expressivity, showing that increasing depth exponentially reduces approximation error. This theoretical insight was validated through experiments on symbolic and continuous-valued state-tracking tasks, confirming the empirical performance of deep sequence models. AI

Researchers have developed a new adapter module called Lighting Convolutional-Attention (LCA) to improve the robustness of foundation models like SAM for instance segmentation under varied lighting conditions. LCA processes RGB features alongside contrast maps to distinguish structural changes from illumination artifacts, enhancing segmentation accuracy without needing to fine-tune the entire model. The module is trained using a pairwise strategy with a specific loss term to penalize discrepancies between clean and illuminated images, and its effectiveness is validated on existing benchmarks and a new synthetic dataset designed for complex lighting. AI