Entity · paper

Evidence-Backed Video Question Answering

paperactiveprovisionalevidence-backed-video-question-answering-eee2f28f·1 events·first seen 5d ago

Aliases: Evidence-Backed Video Question Answering

Co-occurring entities

ST-Evidence-Instruct Salesforce Research UniPixel ST-Evidence

More like this (12)

Visual Question Answering Document Visual Question Answering Multi-hop Question Answering Beyond Tokenization: Direct Timestep Embedding and Contrastive Alignment for Time-Series Question Answering Bias Benchmark for Question Answering Trace Only What You Need: Structure-Aware On-Demand Hypergraph Memory for Long-Document Question Answering Where Does the Answer Come From? Benchmarking View-Level Visual Evidence Identification in Multi-View MLLMs for Autonomous Driving Google-Proof Q&A ICML 2026 Workshop on Efficient Multimodal Question Answering Observation-Guided Video-Context Routing Knowledge-Graph Grounding Helps LLMs Only for Out-of-Training Knowledge: A Controlled Study on Clinical Question Answering Table Question Answering

Recent events (1)

5arXiv · cs.AI·5d ago·source ↗

E-VQA: Evidence-Backed Video Question Answering with Spatio-Temporal Grounding

Researchers from Salesforce AI Research propose Evidence-Backed Video Question Answering (E-VQA), a task requiring Video LLMs to jointly produce semantic answers and precise spatio-temporal evidence via dense tracked object segmentation masklets. They introduce ST-Evidence, the first human-verified pixel-level grounding benchmark for video QA, and ST-Evidence-Instruct, a 160k-scale training dataset. Fine-tuning on this data yields substantial gains over size-matched baselines (+27.2 t-mean, +13.8 J&F on a 7B model), revealing a critical gap between QA accuracy and true visual perception that scaling alone does not close.

Evaluation and Benchmarking Multimodal Progress ST-Evidence-Instruct Evidence-Backed Video Question Answering Salesforce Research +2 more