natalymr/dl-notes

This repo contains my notes from my university Deep Learning course

Deep Learning

This repo contains my notes from my university Deep Learning course.

You should open the links that are located below so that the latex-formulas are correctly visible.

A list of questions for the exam and at the same time course content:

1. Question

• Loss functions: Smooth L1, Bounding Box Regression, Triplet Loss.
• Variational Autoencoders.

2. Question

• Activation functions: ReLU, LeakyReLU, PReLU, MaxOut.
• Image Classification Networks: AlexNet, VGGnet, GoogLeNet.

3. Question

• Attention. Gated Attention.
• Image Segmentation Networks: FCN, SegNet, UNet.

4. Question

• Convolutions. Causal convolutions. Dilated convolutions. Max pooling. Average pooling. Padding.
• Recurrent Neural Networks. LSTM. GRU.

5. Question

• Skip connections. ResNet. Highway connection.
• Generative Adversarial Networks.

6. Question

• Optimizers: Adagrad, Adadelta, RMSProp.
• Speech recognition. Connectionist Temporal Classification. Deep Speech. CNN Speech Recognition.

7. Question

• Word embeddings: Co-occurrence Matrix, Word2Vec, CBOW, Skip-Gram, GloVE, FastText.
• Deep Reinforcement Learning: Deep Q-Network, Deep Deterministic Policy Gradient.

8. Question

• Regularization: L2, Early Stopping, Dropout, Dropconnect, Batch Normalization.
• Adversarial attacks: White box, Black box, Targeted, Untargeted.

9. Question

• Optimizers: Stochastic Gradient Descent, Momentum, Nesterov Momentum.
• Object Detection. Faster R-CNN, Mask R-CNN.

10. Question

• Optimizers: Adam, Nadam.
• Object Detection. IoU, mAP. R-CNN. Fast R-CNN.