Skip to content
View Battam1111's full-sized avatar
🎯
Focusing
🎯
Focusing
26 followers · 50 following

Achievements

Achievement: Pull SharkAchievement: Starstruck

Achievements

Achievement: Pull SharkAchievement: Starstruck

Organizations

@polyunlp

Block or report Battam1111

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
battam1111/README.md

Yanjun Chen

Homepage  ·  Google Scholar  ·  ORCID

Hi 👋, I'm Yanjun. Third-year PhD candidate at Hong Kong PolyU, joint with EIT Ningbo. Currently a research intern at Microsoft Research Asia, Singapore.

In reinforcement learning, models are trainable. The environments that train them are not.
I'm working on closing that gap.

What I'm Building

C3 assigns exact credit in LLM agent teams: when the trace is the state, each message's worth comes from actually continuing the run, not from a prediction. Accuracy Paradox (EMNLP 2024) shows that the most accurate reward model is not the one that trains best: moderately accurate ones train better language models.

On the side: OmniSeek, a self-hosted deep-retrieval engine for AI agents, and Myco, persistent memory infrastructure. Both in the MCP ecosystem.

Also into board games 🎲, jazz 🎷, karaoke 🎤, and old-school anime that never quite went away.

Best reached by email.

Pinned Loading

  1. AccuracyParadox-RLHF AccuracyParadox-RLHF Public

    [EMNLP 2024 Main] Official implementation of the paper "The Accuracy Paradox in RLHF: When Better Reward Models Don't Yield Better Language Models".

    Python 8

  2. EIT-EAST-Lab/C3 EIT-EAST-Lab/C3 Public

    Official implementation of the paper "Contextual Counterfactual Credit Assignment for Multi-Agent Reinforcement Learning in LLM Collaboration". (by Yanjun Chen)

    Python 43

  3. Myco Myco Public

    The living armor an AI agent inhabits: eternal devouring, eternal evolution, eternal amplification.

    Rust 65 7

  4. omniseek omniseek Public

    Your agent seeks what search can't find. A self-hosted perception MCP server that transcribes speech, reads behind logins, sees images and video frames, crosses languages, and remembers.

    Python 50 5