Deep learning.

Neural Networks and Deep Learning.

• Understanding, implementing the math behind forward/backward propagation; differentiating and understanding why/when we use activation functions (tanh, ReLU, sigmoid).

Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization.

• Understanding why/how we use methods to optimize algorithm and implementing: Regularization; Momemtum; RMSprop; Adam; Mini-batch size; Batch normalization; tips in tuning hyperparameters process.

Structuring Machine Learning Projects.

• Strategy to choose the right direction to improve model; Detect the avoidable bias, variance, mismatch data distribution, overfitting; Do error analysis.

Convolutional Neural Networks.

• Why we use CNN, understand the function of components inside CNN. Understanding the architecture and the target of LeNet‑5, AlexNet, ResNet, Inception Network, Mobile Net, YOLO, U-net, Siamese Network.
• Projects:

1. Implement basic building blocks and skip connection of ResNets.
2. Build Alpaca/Not Alpaca classifier using pre-train MobileNetv2.
3. Implement YOLO (IoU, non-max suppression, handle anchor box) to detect objects.
4. Build U-net to do Image segmentation.
5. Build Face recognition system using Siamese network.
6. Generate art using Neural Transfer Style.

Natural language processing.

• Understanding vector spaces, probabilities models, sequence models (GRU, LSTM, Bidirectional RNN), attention model, transformer network.
• Projects

1. Build Language model to generate dinosaur names.
2. Build LSTM model to generate music (Jazz).
3. Load and modify pre-trained word embeddings.
4. Generate emoji for a sentence.
5. Build Machine translation using Attention model to translate "25th of June, 2009" into "2009-06-25".
6. Implement an algorithm for trigger word detection.