Neural networks assignments from University Politehnica of Bucharest
- A1: Linear regression. Logistic regression.
- A2_1: Feed-forward neural network (a FCNet) for image classification (SGD, Nesterov Momentum, Adam).
- A2_2: Update rule (SGD, Nesterov Momentum, Adagrad, RMSProp, Adam) examples on simple functions.
- A3: Convolutional Neural Networks (CNNs) for image classification (CIFAR-10). Weights visualization.
- A4: Text generation with Recurrent Neural Networks (RNNs)
- A5: Image generation with Restricted Boltzmann Machines (RBMs)
- A6: Image generation with Variational Autoencoders (VAEs)
- A7: Image generation with Generative Adversarial Networks (GANs)
- A8_1: Deep Q-Networks (DQNs) on Atari-2600 games
- A8_2: Proof of Policy Gradient Theorem (PGT)
- A9: Advantage Actor-Critic (A2C) on Atari-2600 games
- Project: Image segmentation with saliency maps