Guanghan Ning
I obtained a Ph.D. degree from Mizzou, under Prof. Henry He. Before the end of 2022, I enjoyed research on vision & deep learning, loved participating competitions and challenges. My first popular project was ROLO for object tracking; another project worth mentioning, called LightTrack, ranked 1st place on PoseTrack Challenge.
Since the beginning of 2023, I have switched my focus to developing foundation models, especially Code LLMs. As a believer in singularity, I am dedicated to contributing my limited part. In my spare time, I am also a blogger writing articles about AGI.
Numero 10 in an amateur futbol club located in the Bay Area of California. Devoted fan of nostalgic music. My Piano Covers (click to play or pause): Eyes on Me, Aerith's Theme.
Articles
Singularity Memoirs
- From Weeping Angels to Godlike Machines: AGI at the Crossroads
- The Weeping Angels of AI: LLMs, Time, and Human-Like Cognition
- The Quantum Labyrinth: Is Reality an Interface?
Intelligence Cartography
- Enter the Scaling of RL Environments
- From The Witness to ARC-AGI-3: 'Teaching' Fluid Intelligence via 1,872 Puzzles
- Re-visiting Mid-training Stage: for & with Agentic RL
- Inside the Agentic RL Training Loop
- RL Infra for Large-Scale Agentic Training
- more →
Recent Papers
Seed1.5-Thinking: Advancing Superb Reasoning Models with Reinforcement Learning
Seed-Coder: Let the Code Model Curate Data for Itself
FullStack Bench: Evaluating LLMs as Full Stack Coder
Unconstrained Model Merging for Enhanced LLM Reasoning
InfiBench: Evaluating the Question-Answering Capabilities of Code Large Language Models
Learning Stackable and Skippable LEGO Bricks for Efficient, Reconfigurable, and Variable-Resolution Diffusion Modeling
|
Here is the full list of publications.
Selected Projects
Open-sourced at Github: ROLO / LightTrack / FractalNet / OpenSVAI / mxnet-centernet
Patents
10+ patents have been granted.
Posts
Previous site (Since 2015) is accidentally down. All posts are lost. But worry not.
Here is a list of back-up posts saved on WeChat. Most of them were written in 2015 (while I was young, bold but fun).
2025 blog is where I currently share fresh ideas about AGI & ASI. Hope you enjoy:)
Competitions
1st place on PoseTrack Challenge Benchmark (non-public test set).
Runner-up on Human Pose Estimation Benchmarks, MPII (non-public test set) & LSP.
Professional Services
50+ verified reviews from Publons.
Conference reviewer for CVPR, ICCV, NeurIPS, etc.
Journal reviewer for Pattern Recognition, TMM, TPAMI, TCSVT, etc.