wanshuiyin

wanshuiyin

Explained projects

Diffusion_Code: The One-Line Repo That Exposes the Whole Diffusion Stack
3D Generative AI
Diffusion_Code: The One-Line Repo That Exposes the Whole Diffusion Stack
A bare README turns into a map of the hidden machinery behind diffusion models, and why so many simple AI projects are anything but.
7 min read · Apr 7, 2026
DuelingUCB: Learning the Best Choice Without Ever Seeing a Score
Benchmarks & Eval
DuelingUCB: Learning the Best Choice Without Ever Seeing a Score
A compact Python repo that turns pairwise wins into an Upper Confidence Bound loop, showing why preference feedback can be more useful than numeric reward.
8 min read · Apr 7, 2026
Tencent_test: The Empty Repo That Points to AI-Written Unit Tests
Coding Agents
Tencent_test: The Empty Repo That Points to AI-Written Unit Tests
A blank GitHub shell becomes a doorway into TestAgent, a specialized model pipeline that drafts tests and completes missing assertions.
7 min read · Apr 7, 2026
wanshuiyin.github.io: The Smallest Possible Personal Site
Open-Source Replacements
wanshuiyin.github.io: The Smallest Possible Personal Site
A GitHub Pages repo that turns a reserved username into a public homepage with almost no code.
6 min read · Apr 7, 2026
Inside `wanshuiyin/AAAI-2023-The-Learnability-of-Nonlinear-SSL`: A Proof You Can Run for How Nonlinear SSL Learns
Benchmarks & Eval
Inside `wanshuiyin/AAAI-2023-The-Learnability-of-Nonlinear-SSL`: A Proof You Can Run for How Nonlinear SSL Learns
A hand-coded NumPy simulator turns an AAAI theory paper into a geometric story about basins of attraction, augmentations, and feature recovery.
8 min read · Apr 7, 2026
AISTATS_VESDE: The Diffusion Repo That Already Knows the Answer
Local LLM Inference
AISTATS_VESDE: The Diffusion Repo That Already Knows the Answer
By swapping a learned score network for a closed-form Gaussian mixture, this tiny research codebase turns sampler design into a test of discretization error, correction steps, and variance schedule choice.
8 min read · Apr 6, 2026
OpenReview-Score-Watcher: The Scoreboard Hidden in a Paper Page
AI Dev CLIs
OpenReview-Score-Watcher: The Scoreboard Hidden in a Paper Page
A Chrome extension that reads OpenReview review notes, averages the ratings, and keeps the result in a floating window that follows the reader.
6 min read · Apr 6, 2026
Auto-claude-code-research-in-sleep: The End of the Agent Framework: Unpacking ARIS
Multi-Agent Frameworks
Auto-claude-code-research-in-sleep: The End of the Agent Framework: Unpacking ARIS
How a folder of plain Markdown files and adversarial LLM psychology turned Claude Code into an autonomous ML researcher.
8 min read · Mar 31, 2026