VLMs hallucinate by prioritizing contradictory on-screen text over visual content, addressed via the VisualTextTrap benchmark with 6,057 human-validated samples and the VTHM-MoE dual-encoder framework using dimension-specific experts and adaptive routing.
Video-of-thought: Step-by-step video reasoning from perception to cognition
10 Pith papers cite this work. Polarity classification is still indexing.
10
Pith papers citing it
citation-role summary
background 3
citation-polarity summary
fields
cs.CV 10years
2026 10verdicts
UNVERDICTED 10roles
background 3polarities
background 3representative citing papers
Introduces AVTrack dataset for audio-visual tracking in challenging human-centric scenes, demonstrating performance drops in existing methods.
ViSRA boosts MLLM 3D spatial reasoning performance by up to 28.9% on unseen tasks via a plug-and-play video-based agent that extracts explicit spatial cues from expert models without any post-training.
Act2See trains VLMs via supervised fine-tuning on verified reasoning traces to interleave active frame calls within text CoTs, yielding SOTA results on video reasoning benchmarks.
MERIT restores temporal reasoning in VLMs via layer-selective self-attention merging guided by a TR-improving objective that penalizes TP degradation.
SCP defines a new benchmark task for predicting spatial causal outcomes beyond direct observation and shows that 23 leading models lag far behind humans on it.
CREDiT applies counterfactual reasoning via structural causal models to decompose video representations into causal and non-causal parts for more reliable VideoQA on datasets like NExT-GQA and SportsQA.
SG-PVR introduces plan-and-verify reasoning grounded in spatio-temporal scene graphs to address verification gaps and implicit evidence in existing T2V reward models.
Chain-of-Glimpse is a reinforcement-learning-based framework that iteratively grounds visual evidence regions to enable multi-step object-aware reasoning in videos.
This is a survey that frames video MLLM research via a human-view formulation of perceptual representations, memory states, reasoning traces, and predictions, then reviews methods, datasets, benchmarks, and open problems.
citing papers explorer
-
When Text Hijacks Vision: Benchmarking and Mitigating Text Overlay-Induced Hallucination in Vision Language Models
VLMs hallucinate by prioritizing contradictory on-screen text over visual content, addressed via the VisualTextTrap benchmark with 6,057 human-validated samples and the VTHM-MoE dual-encoder framework using dimension-specific experts and adaptive routing.
-
AVTrack: Audio-Visual Tracking in Human-centric Complex Scenes
Introduces AVTrack dataset for audio-visual tracking in challenging human-centric scenes, demonstrating performance drops in existing methods.
-
ViSRA: A Video-based Spatial Reasoning Agent for Multi-modal Large Language Models
ViSRA boosts MLLM 3D spatial reasoning performance by up to 28.9% on unseen tasks via a plug-and-play video-based agent that extracts explicit spatial cues from expert models without any post-training.
-
Act2See: Emergent Active Visual Perception for Video Reasoning
Act2See trains VLMs via supervised fine-tuning on verified reasoning traces to interleave active frame calls within text CoTs, yielding SOTA results on video reasoning benchmarks.
-
Reasoning Resides in Layers: Restoring Temporal Reasoning in Video-Language Models with Layer-Selective Merging
MERIT restores temporal reasoning in VLMs via layer-selective self-attention merging guided by a TR-improving objective that penalizes TP degradation.
-
SCP: Spatial Causal Prediction in Video
SCP defines a new benchmark task for predicting spatial causal outcomes beyond direct observation and shows that 23 leading models lag far behind humans on it.
-
Counterfactual Reasoning for Fine-Grained Evidence Disentanglement in VideoQA
CREDiT applies counterfactual reasoning via structural causal models to decompose video representations into causal and non-causal parts for more reliable VideoQA on datasets like NExT-GQA and SportsQA.
-
Plan-and-Verify Video Reward Reasoning with Spatio-Temporal Scene Graph Grounding
SG-PVR introduces plan-and-verify reasoning grounded in spatio-temporal scene graphs to address verification gaps and implicit evidence in existing T2V reward models.
-
Chain-of-Glimpse: Search-Guided Progressive Object-Grounded Reasoning for Video Understanding
Chain-of-Glimpse is a reinforcement-learning-based framework that iteratively grounds visual evidence regions to enable multi-step object-aware reasoning in videos.
-
Watch, Remember, Reason: Human-View Video Understanding with MLLMs
This is a survey that frames video MLLM research via a human-view formulation of perceptual representations, memory states, reasoning traces, and predictions, then reviews methods, datasets, benchmarks, and open problems.