Get Started
Research questionHow can evolving-context summarization repair outdated claims without damaging supported content?As event contexts change, existing summaries may contain claims that are no longer supported. Rewriting the entire summary can unnecessarily alter accurate content and make the changes harder to track.
Diffusion Models
Evaluation & Benchmarks
Natural Language Processing
Latest papersRecent research connected to this question, newest first.Detect, Remask, Repair: Diffusion Editing for Faithful Summarization of Evolving ContextsApplies to evolving-context summarization of real-world events using an existing draft and updated context. Evidence covers DialogSum, the synthetic StreamSum benchmark, and post-hoc correction of autoregressive summaries; broader datasets and deployment conditions are not established.research paper · Sep 3, 2026
Related questions
How can multi-document summaries make each claim traceable to supporting source spans?How can aspect-based summarizers adapt user-provided aspects when they are incomplete or irrelevant?How can LLMs propagate local conversational revisions across dependent artifact components without excessive test-time computation?How can language models compress conversational and document context without reconstruction while preserving task-relevant information?
Home
Topics
Search
Library