- Added `image_secrete.bmp` and `image_test_8bits.bmp` to the Cybersecurity directory.
- Updated README.md to include a new section for Cybersecurity, highlighting data security and analysis.
- Created README.md detailing project overview, algorithms, architecture, environment, project structure, key results, known issues, and dependencies.
- Added checkpoint files for Monte Carlo agent and updated existing checkpoints for DQN and Q-Learning agents.
- Included new training and evaluation plots for DQN, Monte Carlo, and championship matrix.
- Updated the Q-Learning model checkpoint (q_learning.pkl) to reflect recent training improvements.
- Revised the training curves plot (Q-Learning_training_curves.png) to visualize the latest performance metrics.
- Updated weight and feature storage to use float32 for reduced memory bandwidth.
- Implemented compact storage for raw observations as uint8, batch-normalized at episode end.
- Introduced vectorized return computation and chunk-based weight updates using einsum.
- Reduced weight sanitization to once per episode instead of per-step.
- Refactored action selection and return calculation for improved efficiency.
- Bump catboost from 1.2.8 to 1.2.10
- Update google-api-python-client from 2.190.0 to 2.191.0
- Upgrade langchain from 1.2.0 to 1.2.10
- Update langchain-core from 1.2.16 to 1.2.17
- Upgrade langchain-huggingface from 1.2.0 to 1.2.1
- Bump marimo from 0.19.11 to 0.20.2
- Update matplotlib from 3.10.1 to 3.10.8
- Upgrade numpy from 2.2.5 to 2.4.2
- Update opencv-python from 4.11.0.86 to 4.13.0.92
- Bump pandas from 2.2.3 to 3.0.1
- Update plotly from 6.3.0 to 6.6.0
- Upgrade polars from 1.37.0 to 1.38.1
- Bump rasterio from 1.4.4 to 1.5.0
- Update scikit-learn from 1.6.1 to 1.8.0
- Upgrade scipy from 1.15.2 to 1.17.1
- Bump shap from 0.49.1 to 0.50.0
- Adjust isort section order for better readability
- Added Q-learning model checkpoint file (q_learning.pkl) to the checkpoints directory.
- Included training curves for Q-learning in the plots directory (Q-Learning_training_curves.png).
- Added a new checkpoint file for the SARSA algorithm at checkpoints/sarsa.pkl.
- Included training curves for the SARSA algorithm in plots/SARSA_training_curves.png.
- Introduced functions to create a Tennis environment and run matches between agents.
- Implemented a round-robin tournament format excluding random agents.
- Added win-rate matrix visualization and final ranking of agents based on performance.
- Updated imports to include necessary libraries for the new functionality.
- Created a new R script 'packages.R' to manage necessary packages for the Classification and Regression module.
- Included a list of required packages and a function to install any missing packages.
- Implemented loading of all packages and added a success message upon completion.
- Added langchain-mistralai version 1.1.1 to pyproject.toml dependencies.
- Updated uv.lock to include langchain-mistralai package details and its dependencies.
- Updated pyproject.toml to include "tiktoken>=0.12.0" in dependencies.
- Modified uv.lock to reflect the addition of tiktoken in both dependencies and requires-dist sections.
- Updated execution counts for code cells to reflect changes.
- Enhanced output formatting for matched FAQs, including clearer question and answer presentation.
- Adjusted the similarity threshold in the `match_faq` function to allow for looser matching (default set to 0.5).
- Improved documentation for the threshold parameter to clarify its usage.
- Add extraction/NA-cleaning of residuals and plots + ACF for additive and multiplicative decompositions
- Clarify why multiplicative seasonality fits AirPassengers better
- Update STL plot title
- Add ACF for the final differenced series
- Expand explanation of differencing operators and show combined differencing formula
