Large Multimodal Models
PulseAugur coverage of Large Multimodal Models — every cluster mentioning Large Multimodal Models across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
Quicksviewer LMM achieves 45x video compression with Gumbel Softmax
Researchers have introduced Quicksviewer, a novel Large Multimodal Model (LMM) designed for more efficient video understanding. Quicksviewer employs a unique perceiving paradigm that partitions videos into varying cubes…
-
New tuning methods teach LMMs skills without forgetting
Researchers have developed novel methods for fine-tuning large multimodal models (LMMs) to acquire new skills without degrading existing capabilities. By analyzing the output token distribution and identifying a 'counti…
-
New ROAR benchmark tests LMM agents in real-time multiplayer games
Researchers have introduced ROAR, a new benchmark designed to evaluate Large Multimodal Models (LMMs) as agents in real-time, multiplayer action-shooter game environments. Unlike previous turn-based benchmarks, ROAR add…
-
New ROAR benchmark tests LMM agents in real-time multiplayer games
Researchers have introduced ROAR, a new benchmark designed to evaluate Large Multimodal Models (LMMs) as agents in real-time, multiplayer action-shooter game environments. Unlike existing turn-based benchmarks, ROAR add…
-
CORE framework uses LMMs for anomaly synthesis in industrial settings
Researchers have developed CORE, a novel framework that leverages Large Multimodal Models (LMMs) for training-free anomaly synthesis in industrial settings. This approach addresses the scarcity of anomalous samples by u…
-
New benchmark UnifiedAttack targets LMM safety in harmful image-text generation
Researchers have developed UnifiedAttack, a new benchmark to evaluate the safety of Large Multimodal Models (LMMs) in generating harmful content by coordinating text and image modalities. This approach aims to identify …
-
RegRet framework enhances region-level retrieval in Large Multimodal Models
Researchers have introduced RegRet, a new framework designed to improve region-level retrieval in Large Multimodal Models (LMMs). This approach focuses on enhancing the capture of detailed regional features within image…
-
New LogiScope-VQA benchmark reveals LMMs lag human performance in industrial hazard identification
A new benchmark dataset called LogiScope-VQA has been developed to evaluate the capabilities of large multimodal models (LMMs) in identifying logistics hazards within industrial settings. The dataset, comprising images,…
-
New benchmark tests AI's social reasoning in counterfactual videos
Researchers have introduced SocialReasonBench, a new video question-answering benchmark designed to evaluate the social reasoning capabilities of Large Multimodal Models (LMMs). This benchmark utilizes counterfactual na…
-
New ImgCoder framework generates scientifically accurate images for AI reasoning
A new research paper introduces ImgCoder, a framework designed to generate scientifically accurate images, addressing the limitations of current text-to-image models that often produce visually plausible but logically i…
-
New MedReaMM Benchmark Reveals LMMs Struggle with Clinical Diagnosis
Researchers have introduced MedReaMM, a new benchmark designed to evaluate the diagnostic synthesis capabilities of Large Multimodal Models (LMMs) in clinical settings. Unlike previous benchmarks that focused on isolate…
-
ArmorOCR framework enhances adversarial OCR perception with new AdvSpot benchmark
Researchers have introduced ArmorOCR, a novel two-stage training framework designed to enhance the robustness of optical character recognition (OCR) against adversarial attacks. This framework addresses the limitations …
-
New RL framework advances 3D point cloud quality assessment
Researchers have introduced PCQA-R1, a novel reinforcement learning framework designed for no-reference 3D point cloud quality assessment. This system utilizes a chain-of-thought dataset and a Gaussian proximity reward …
-
New Latent-OPD method enhances LMMs for frame-efficient video reasoning
Researchers have introduced Latent-OPD, a novel method for improving the efficiency of Large Multimodal Models (LMMs) in video reasoning. This technique enhances On-Policy Distillation (OPD) by incorporating trajectory-…
-
New framework enhances AI reasoning for dense sports video analysis
Researchers have developed SportsGrounder, a new framework designed to improve the reasoning capabilities of Large Multimodal Models (LMMs) when analyzing dense sports videos. The framework addresses the challenge of di…
-
New ReGraph framework enables LMMs to generate structured recipe graphs from food images
Researchers have introduced ReGraph, a novel dataset and framework designed to enable Large Multimodal Models (LMMs) to generate structured recipe graphs from food images. This approach aims to explicitly represent ingr…
-
Open-source multimodal search agent POINTS-Seeker tackles context limits
Researchers have introduced POINTS-Seeker, an open-source multimodal search agent designed to overcome limitations in current Large Multimodal Models (LMMs). The system addresses the challenges of cultivating search age…
-
New framework uses LMMs to correct visual species recognition errors
A new research paper proposes a framework called Post-hoc Correction (POC) to improve visual species recognition (VSR) accuracy. The study found that while Large Multimodal Models (LMMs) underperform expert few-shot lea…
-
New VVM-Tuning framework enhances LMMs for unseen visual modalities
Researchers have developed a new training framework called VVM-Tuning to enhance the generalization capabilities of Large Multimodal Models (LMMs) across various visual modalities. This method synthesizes diverse visual…
-
New HART technique enables LMMs to reason with high-resolution images without annotations
Researchers have developed a new technique called HART (High-resolution Annotation-free Reasoning Technique) to improve how Large Multimodal Models (LMMs) handle high-resolution images. Current LMMs struggle with the la…