TildAlice – Tech Blog for Developers
Python, AI, Deep Learning, Data Analysis, Computer Vision, Reinforcement Learning โ tips and tutorials
Latest Posts
-
๋ฐ์ดํฐ๋ถ์ ์ทจ์ ์ค๋น ๋ก๋๋งต: Kaggle ๋ฐ์ดํฐ์ ์ผ๋ก ์์ํ๋ ์ค๋ฌด ํ๋ก์ ํธ ์ ๋ฌธ
Data analyst roadmap: 5 Kaggle projects that replicate real job tasks (EDA, ML pipeline, time series, A/B test). Zero to portfolio in 8 weeks.
-
๋ฐ์ดํฐ๋ถ์๊ฐ ์ทจ์ค์์ ์ํ ํ์ ๋ฐ์ดํฐ๋ถ์ ๊ธฐ๋ฒ 5๊ฐ์ง์ ์ค์ ์์
5 data analysis techniques used daily by professional analysts โ with Python code you can run in 10 minutes. EDA to hypothesis testing.
-
OpenAI Whisper ๋ชจ๋ธ ํ์ธํ๋ ์๋ฒฝ ๊ฐ์ด๋: ํ๊ตญ์ด ์์ฑ์ธ์ ์ ํ๋ ๋์ด๊ธฐ
Fine-tune OpenAI Whisper for Korean speech: boost accuracy from 82% to 94% with custom datasets. Includes training code and benchmarks.
-
Advanced Game AI: Multi-Agent RL and Self-Play
Build AlphaGo-style game AI: self-play, curriculum learning, and population training. Why single-agent RL fails for competitive games.
-
PPO and A3C for Complex Game AI Environments
Why PPO replaced DQN for game AI: actor-critic variance reduction, continuous control, and 10x faster convergence on Atari benchmarks.
-
Deep Q-Networks (DQN): Training AI to Play Atari Games
Build a Deep Q-Network that beats human Atari scores. Experience replay, target networks, and Double DQN explained with training code.
-
Q-Learning for Grid Worlds: Building Your First Game AI Agent
Train your first game AI in 50 lines of Python โ no neural networks needed. Pure Q-learning with the Bellman equation for grid worlds.
-
Game AI with RL: Why It Beats Traditional Methods
Why RL beats rule-based game AI: exploration-exploitation tradeoff, credit assignment, and when traditional pathfinding still wins.