Interactive-Machine-Learning

Exercises from the course Praktikum Interactive Machine Learning at University of Augsburg

Commit rules:

1. Exercise #: the work done - what to do next / what is needed
2. README.md : update

---

All(mostly all) work is done in exercise_notebook.ipynb files

Model or Weights files are not pushed here, because they take a lot of memory. More practical would be to run model urself and obtain needed weights.

imgs-folder can be found in root directory. Paths to this folder in notebooks are outdated

---

02_ML_Basics

Working on Pima Indians diabetes dataset from the UCI Machine Learning repository.

1. Normalization, balancing data
2. Support Vector Machines Classifier.
3. Basic Keras implementation.

03_Neural_Networks

Classifying Pokemons with CNN

1. OpenCV
2. Building own Sequential model with keras
3. Transfer Learning - VGG16
4. Tensorboard

04_Cooperative_ML

Classifying Pokemons with Neural Networks, pretraining the model, Building GUI

1. GUI with ipywidgets
2. Building own Functional model with keras
3. Transfer Learning - VGG16

05_eXplainable_AI

Classifying Pokemons with Neural Networks, Building GUI to visualise convolutional blocks and filters, Lime framework

1. GUI with ipywidgets to visualise convolutional blocks and filters
2. Lime for visualizing model-predicting explanations

btc_eth_LSTM

The final Project of the Course. My part part the task was to build LSTM-network for Bitcoin and Ethereum Prediction

1. Bitcoin prediction for one day ahead using 4 features and a sequence of 1/3/7 days
2. Ethereum prediction for one day ahead using one feature