Beijing Academy of Artificial Intelligence · 11 records

dateidtitleauthorsvenue↑
2026-10-162610.16784InfoVis: Unified Slide Image Generation and EditingJiahao Wang et al.arXiv32
2026-10-012609.38288AREX-2: Advancing Self-Improving Agents through Long-Horizon Reflective TasksHongjin Qian et al.arXiv92
2026-09-292609.32712MassAlloc Attention: Let Attention Allocate Its Own ComputeJingze Shi et al.arXiv42
2026-09-292609.32704CoWindow Attention: Full Causal Coverage Is a Collective PropertyJingze Shi et al.arXiv32
2026-09-032609.02749Repo-To-Skill: Distilling GitHub Repositories Into AI4AI SkillsJianlyu Chen et al.arXiv137
2026-08-212608.20281Inject, Align, Recover: Staged Post-Training for Retrieval-Free Document Knowledge InternalizationQian Kou et al.arXiv10
2026-08-042608.01862Wnuan: Staged Post-Training for Question Answering over Proprietary Enterprise KnowledgeXiaofeng Shi et al.arXiv6
2026-07-242607.21461AREX: Towards a Recursively Self-Improving Agent for Deep ResearchShuqi Lu et al.arXiv95
2026-06-252606.20023When Lower Privileges Suffice: Investigating Over-Privileged Tool Selection in LLM AgentsKaiyue Yang et al.arXiv3
2026-06-242606.23997ChartWalker: Benchmarking the Cross-Chart RAG TaskNing Tang et al.arXiv4
2026-05-012604.27711ExoActor: Exocentric Video Generation as Generalizable Interactive Humanoid ControlYanghao Zhou et al.arXiv39