Blog
Welcome to my blog! Here I share insights from my daily readings of AI research papers and write about interesting open-source projects I’m working on. I hope these posts are helpful to the community and spark interesting discussions.
AI Daily Paper Summary
I read AI and ML research papers daily and post concise summaries here. Each entry covers the key contributions, methods, and takeaways from recent papers across topics like deep learning, reinforcement learning, NLP, computer vision, and more.
Latest — September 29, 2026
- YuE2 — Unifies symbolic composition and full-song audio generation in one editable music model. arxiv.org/abs/2609.33757
- Behavioral Shadows — Shows post-training capabilities can transfer through task-unrelated one-word outputs. arxiv.org/abs/2609.29233
- Domain-Normalized MOPD — Balances specialist-teacher signals so one domain does not dominate shared-student training. arxiv.org/abs/2609.35347
- VisionHOPE — Turns visual backbones into stable self-modifying memory systems within each image. arxiv.org/abs/2609.33325
- Physical Coding — Extends coding-agent workflows to robots with explicit state, tools, verification, and recovery. arxiv.org/abs/2609.35432
- Duplex-MPE — Tests whether speech assistants know when to answer, stay silent, or stop in multi-party conversations. arxiv.org/abs/2609.31948
- GAGAR — Replaces binary test rewards with groupwise code-quality credit for more stable agent RL. arxiv.org/abs/2609.32577
- MassAlloc Attention — Lets attention skip negligible post-score computation while preserving full causal access. arxiv.org/abs/2609.32712
- CoWindow Attention — Splits distant context across heads to retain collective coverage at far lower long-context cost. arxiv.org/abs/2609.32704
- TraceDance — Converts real agent failures into targeted benchmarks and continuation-based evaluations. arxiv.org/abs/2609.33295
Interesting Projects
In this section, I share outcomes, results, and lessons learned from open-source projects I’m working on. The goal is to give back to the community and document the journey of building and experimenting with different ideas.