mmatulenko Goto Github PK

BTC Ai trade bot | Work in progress..

Application of Deep Learning (SNN, CNN, LSTM) and Deep Evidential Learning (Capsa) to create an ensemble based model for stock market prediction (training data - Bitcoin daily chart with Technical Analysis, Pattern Recognition and Fundamental Analysis). Ensemble model shows:

0.9% mean absolute percentage error(mape) on test set (20%) and 2% mape on validation set.

1. T5 Transformer Summarization: T5 Seq2Seq model text summarization performance on the XSum dataset. The code is provided as a Colab notebook. https://github.com/MMatulenko/transformer_summarization

2. Machine Translation with Seq2Seq Model: This project uses a Seq2Seq model to perform machine translation from English to German. https://github.com/MMatulenko/t5_seq2seq_translation

3. Keras IMDB review classification: Simple binary classification model using the IMDB movie review dataset from Keras. https://github.com/MMatulenko/keras_imdb_classification

4. Titanic Survival Prediction Predict the survival of passengers aboard the Titanic based on various features such as gender, age, class, and fare. EDA analyze + various model testing (logreg) https://github.com/MMatulenko/Titanic

5. Wine Quality multiclass classification: Quality of wine prediction. https://github.com/MMatulenko/Wine_Quality

6. Feed Forward NN from scratch: https://github.com/MMatulenko/mm_data_science/blob/main/ffnn_from_scratch.ipynb

7. Liquid Neural Network - robustness without overparametrisation https://github.com/MMatulenko/LNN

8. RL Lunar simple ai, Pacman with Convolution layer, A3C for Kung Fu https://github.com/MMatulenko/RL/tree/main

♜ Other works: https://github.com/MMatulenko/mm_data_science/

♗ Currently: Deep learning, NLP and AGI

♘ I’m looking to collaborate on Large Language Models (LLM) and Artificial general intelligence.

♙ How to reach me: [email protected] | [email protected]