用分阶段验证与知识图谱两种方法，解决长时多视角自拍视频的复杂问答挑战。
CuriosAI Submission to the CASTLE Challenge at EgoVis 2026
- 分三阶段搜索-验证-回答，通过规则抑制幻觉，提升答案可信度。
- 在600+小时数据上，最佳方案准确率达50%，超越基线35%。
- 适合研究多模态推理、视频理解与抗幻觉问答系统的开发者。
CASTLE 2026 要求对超过600小时的同步多视角自拍视频回答185个多项选择题。我们基于共享的多模态预处理层，探索两种方法：方法A（SVA）为搜索-验证-回答三阶段流程，通过四条防幻觉规则验证子窗口，利用大模型按证据优先级融合结果；方法B（TMKG）构建时序多模态知识图，通过图搜索定位核心节点，由单一受控视觉语言模型生成答案。SVA在排行榜上取得0.50的准确率，作为最终提交方案；TMKG准确率为0.35。
原文摘要 · Abstract (English)
CASTLE 2026 asks 185 multiple-choice questions over 600+ hours of synchronized multi-view egocentric video. We explore two approaches on top of a shared multimodal preprocessing layer, including per-person timelines, speaker-resolved transcripts, and multi-VLM caption ensembles. Approach A, SVA: Search-Verify-Answer, is a three-stage pipeline that hierarchically narrows to a primary window, verifies sub-windows with a VLM under four anti-confabulation rules, and fuses evidence with an LLM judge under an evidence-priority hierarchy. Approach B, TMKG: Temporal-Multimodal-Knowledge-Graph, is the contrast: it builds a temporal multimodal knowledge graph, locates a primary cell via graph search, and produces the final answer with a single grounded VLM. SVA reaches a leaderboard accuracy of 0.50 and is our final challenge submission; TMKG reaches 0.35.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。