CLEVRER
PulseAugur coverage of CLEVRER — every cluster mentioning CLEVRER across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
Study: Verifiable rewards boost video-language model accuracy but not time understanding
A new study published on arXiv explores the impact of reinforcement learning from verifiable rewards (RLVR) on video-language models' understanding of time. Researchers fine-tuned four open models, including Qwen3 VL 8B…
-
New SYNCR benchmark tests cross-video reasoning in LLMs · 2 sources tracked
Researchers have introduced SYNCR, a novel framework for evaluating and training cross-video reasoning capabilities in multimodal large language models (MLLMs). Built using simulation engines like Habitat, Kubric, and C…
-
COMET framework boosts video LLMs with enhanced motion and temporal reasoning
Researchers have developed COMET, a new framework designed to enhance video multimodal large language models by improving their understanding of fine-grained motion and temporal reasoning. The framework introduces a ded…
-
MOSH-WM: New Object-Centric World Model Improves Video Prediction
Researchers have developed MOSH-WM, a novel mask-grounded soft-Hamiltonian world model designed for object-centric video prediction. This model explicitly links its position-like state to image support owned by entity s…
-
PhysMind framework boosts video physical reasoning, outperforming GPT-5.5
Researchers have developed PhysMind, a novel framework designed to enhance physical reasoning capabilities in videos. This system constructs reusable, question-agnostic executable worlds from video data, enabling more a…
-
Frontier LLMs Distill Answer Set Programming Theories with High Accuracy
A new study explored distilling Answer Set Programming (ASP) theories from large language models using a neurosymbolic approach. The research tested nine models, including frontier models like Claude Sonnet 4.6, Claude …
-
New AI model learns causal video prediction by focusing on physical interactions
Researchers have developed an Interaction-Aware JEPA (IA-JEPA) model designed to improve causal video prediction by focusing on physical interactions rather than just visual textures. This new approach uses a motion-cen…
-
New Chimera Training Method Enhances Anomaly Detection for Rare Rule Violations
Researchers have developed a novel method called chimera training for anomaly detection, particularly useful when rule violations are rare in training data. This approach uses a neural rule evaluator that compiles logic…
-
New world model approach excels at counterfactual reasoning
Researchers have introduced deterministic event-graph substrates as a novel approach to world models for counterfactual reasoning. These substrates represent agent states as logs of RDF triples and handle counterfactual…