Pinned Loading

1. neural-networks_UMiSI2026 neural-networks_UMiSI2026 Public

   Neural networks & deep learning implemented from scratch in NumPy: backprop, RBMs (CD/PCD), DBNs, autoencoders, a CNN, and Word2Vec - no DL frameworks. AGH.

   Jupyter Notebook

2. reinforcement-learning_UMiSI2026 reinforcement-learning_UMiSI2026 Public

   Reinforcement learning labs: multi-armed bandits, n-step SARSA, MCTS, Actor-Critic, SAC+HER, Decision Transformer, and GRPO / reward-hacking - AGH.

   Python

3. OrbitGuide OrbitGuide Public

   AI-powered Space Law assistant — RAG + LLM (BITEhack 2026)

   Python

4. dota-prediction dota-prediction Public

   Predicting a Dota 2 match's rank tier from OpenDota statistics: logistic baseline -> LightGBM+Optuna -> SHAP/LIME explainability with a data-leakage analysis.

   Python

5. stochastic-methods-in-ML_UMiSI2026 stochastic-methods-in-ML_UMiSI2026 Public

   Stochastic methods in ML — probabilistic approaches

   Jupyter Notebook

6. DarwinWorldProject DarwinWorldProject Public

   Evolution simulation — Java, JavaFX, genetics & ecosystem

   Java