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 8, 2026

1. Compile by Training — Compiles recurring natural-language tasks into reusable local neural functions, reducing repeated model calls and latency. arxiv.org/abs/2609.04199
2. Dr. Claw — Wraps coding agents in an auditable human-in-the-loop workspace for end-to-end scientific research. arxiv.org/abs/2609.00365
3. Bilevel Coordinated Reflection — Frames multi-agent coordination, reflection, memory, and verification as a unified game-theoretic process. arxiv.org/abs/2609.02750
4. RoboTok — Turns web-scale human manipulation videos into training demonstrations for dexterous robot policies. arxiv.org/abs/2609.03199
5. Diffusion-Augmented LLMs — Uses discrete diffusion to decode multiple tokens in parallel without changing the autoregressive model distribution. arxiv.org/abs/2609.04010
6. Iris — Trains search agents on difficult multi-hop questions designed to require retrieved evidence rather than memorized answers. arxiv.org/abs/2609.04304
7. Scal3R — Improves long-video 3D reconstruction by replacing brittle fixed-anchor pose estimation with efficient multi-relative queries. arxiv.org/abs/2609.04201
8. FlowBalance — Grounds dense self-improvement signals in sparse terminal verifiers while preserving diverse reasoning paths. arxiv.org/abs/2609.03241
9. Motion-Omni — Generates conversational speech and synchronized full-body motion jointly instead of through a two-stage cascade. arxiv.org/abs/2609.04250
10. Last Translation Benchmark — Refocuses machine-translation evaluation on difficult failure cases as standard benchmarks approach saturation. arxiv.org/abs/2609.04173

Browse all paper summaries →

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.

Browse all project posts →