SPPU · Final Year Information Technology · 2019 Course
Deep Learning
414443 · Semester VII · 3 credits · In-semester 30 + end-semester 70
Unit 1: Fundamentals of Deep Learning
6 hoursMultilayer perceptrons; feed-forward networks and backpropagation; gradient descent; vanishing gradients; ReLU variants; optimisers; hyperparameters: layer size, momentum and learning rate; dropout, DropConnect, L1 and L2.
Marks: 15 in the in-semester exam
- Multilayer perceptron and backpropagation
- Gradient descent and optimisers
- Vanishing gradient problem
- Activation functions
- Regularisation: dropout, L1 and L2
Unit 2: Convolutional Neural Networks
6 hoursConvolution; parameter sharing and equivariance; pooling; convolution variants; basic CNN architecture; AlexNet.
Marks: 15 in the in-semester exam
- Convolution and parameter sharing
- Pooling
- CNN architecture
- AlexNet
Unit 3: Recurrent Neural Networks
6 hoursTypes of RNNs; feed-forward versus recurrent networks; LSTMs; encoder–decoder architectures; recursive neural networks.
Marks: 17–18 in the end-semester exam
- Recurrent neural networks
- LSTM networks
- Encoder-decoder architectures
- Recursive neural networks
Unit 4: Autoencoders
6 hoursUndercomplete, sparse, denoising and contractive autoencoders; stochastic encoders and decoders; applications.
Marks: 17–18 in the end-semester exam
- Undercomplete autoencoders
- Sparse and denoising autoencoders
- Contractive autoencoders
- Applications of autoencoders
Unit 5: Representation Learning
6 hoursGreedy layer-wise pre-training; transfer learning and domain adaptation; distributed representations; DenseNet.
Marks: 17–18 in the end-semester exam
- Greedy layer-wise pre-training
- Transfer learning and domain adaptation
- DenseNet
Unit 6: Applications
6 hoursImage classification; social network analysis; speech recognition; recommender systems; NLP.
Marks: 17–18 in the end-semester exam
- Deep learning for image classification
- Deep learning for speech and NLP
- Deep learning recommender systems
Unit contents are summarised from the official SPPU syllabus . AiCanCode.org is not affiliated with Savitribai Phule Pune University.