arXiv 2610.28444 · cs.CV · 2026-10-20
EvaOmni: An Evidence-Grounded Benchmark for Audio-Visual Reasoning
Xiaoxing Hu, Wenxi Yao, Chaoyuan Shi, Ziqi Ye, Xiaorong Guo, Peixuan Li, Zhuoran Sun, Ziyang Gong 等 20 人
SJTU VisionXLab
在最新榜 #304 本站收录于 2026-10-24
一句话要点
上海交大 VisionXLab 的 EvaOmni 是以证据为基础的音视频推理基准：不只看选择题答对没有，还检查模型是否真的从音频和画面证据形成推理链，并防止靠文本或单一模态的捷径答对。
SJTU VisionXLab's EvaOmni is an evidence-grounded audio-visual reasoning benchmark that checks whether omni models actually build a reasoning chain from audio and visual evidence, not just answer multiple-choice questions correctly, and guards against text-only or unimodal shortcuts.
阅读 arXiv 原文 PDF Hugging Face 讨论 代码仓库 项目页
主题与学科
同主题：评测与数据
- RealCompanion: Benchmarking Human Understanding from Reasoning over Longitudinal Real-World Conversations ▲287
- PremiseCheck: Efficient Fact-Checking of Human and AI Peer Reviews ▲210
- DecisionPivot: A Diagnostic Benchmark for Decision Models in LLM Systems ▲74
- GameHorizon Suite: Multi-Horizon Data and Evaluation in Gameplay ▲73
- More Choices, Fewer Decisions: Ordinal-Scale Bias in JEV-like Direct-Decision Models ▲65
同主题：多模态
- StreamEval: Evaluating Interactive Streaming Video Generation Models 2026-10-20 · ▲649
- MJev-Spatial: System One Spatial Intelligence via Native Attention Matching 2026-10-15 · ▲556
- Omni-IO Skills: Harnessing Your Agent Omni-Native 2026-09-25 · ▲485
- LoopVL: Recurrent Visual Intelligence 2026-09-29 · ▲417
- VideoJev: System One Model for Video Understanding and Interaction 2026-10-13 · ▲375
数据说明
元数据来自 arXiv API（已核实）；点赞来自 Hugging Face 每日论文；引用来自 Semantic Scholar。最近刷新 2026-10-26。
要点由本站根据摘要自写，不转载摘要原文，请阅读原文。