머신러닝 · GitHub

Book_7_《机器学习》 | 鸢尾花书：从加减乘除到机器学习；欢迎批评指正

“수학은 어렵지 않다” 시리즈의 《선형대수는 어렵지 않다》 상·하권입니다. 66개 주제를 모두 다루며, 비판과 수정 의견을 환영합니다.

머신러닝 및 에이전틱 AI 자료, 실습, 연구

A comprehensive quantitative trading system with AI-powered analysis, real-time data processing, and advanced risk management

• 🔬 Advanced Factor Analysis: Multi-factor models with momentum, value, quality, and volatility factors
• 🤖 AI/LLM Integration: OpenAI GPT integration for market analysis and strategy recommendations

PyTorch implementation of UNet++ (Nested U-Net).

Approaching (Almost) Any Machine Learning Problem

MACE - Fast and accurate machine learning interatomic potentials with higher order equivariant message passing.

List of all the lessons learned, best practices, and links from my time studying machine learning

• It's also model dependent. Normalization will likely help for linear models like linear/logistic reg, but won't be as helpful for deep learning models and neural networks.
• In order to determine whether it will help or not, visualize your data to see what type of features you're dealing with. Do you have some features that have values around 1000 or 2000? Do you have some that are only 0.1 or 0.2? If so, normalizing those inputs is likely a good idea.

A modular, JAX-based framework for Physics-Informed Neural Networks, designed for scalable PDE-constrained learning with support for domain decomposition, attention mechanisms, and performance-oriented training.

Decentralized LLMs fine-tuning and inference with offloading

• Tensor offloading — reduces per-node memory consumption so each peer can host more layers, shrinking the total number of network hops.
• Speculative decoding over internet — reduces communication frequency by sending multiple draft tokens per round-trip.