NTUA-Neural-Networks

💭 NTUA ECE Neural Network and Course Source Codes in Python

This repository hosts the exercises for Neural Networks course of ECE NTUA.

Collaborators: Konstantinos Tsiggelis Konstantinos Spathis Panagiotis Skrempos

Lab1 - Supervised Learning

The whole lab consists of two parts:

1. We apply various classification algorithms like NaiveBayes,kNN and Logistic Regression to the UCI dataset and check and analyze their performance.
2. We construct and optimize MLP and SVM networks and we apply them in a much bigger dataset in Kaggle.

Lab2 - Unsupervised Learning

We create a movie recommendation system basedn on both TFIDF-Vectorizer and Transfer Learning (finding the embeddings through pretainedl NLP models) and we compare their performance. Lastly we use Self organizing Maps(SOM) for the topological and semnatic represenation of the movies.

Lab3 - Deep Learning

We use a classic encoder - decoder architecture in order to find the captions of images. We use CNN as an encoder and for the decoder we apply LSTM and attention layers.