AL-503 (B) Deep Learning Notes
5 notes for Deep Learning, unit by unit.
Notes for AL-503 (B)
- Unit 1 · Short notes Introduction Introduction and History of Deep Learning · McCulloch–Pitts neuron model · Multilayer Perceptrons (MLPs) · Representation power of MLPs · Sigmoid neurons · Feed‑forward neural networks
- Unit 2 · Short notes Deep Feed‑Forward Networks and Optimization Deep feed‑forward neural networks · Gradient Descent (GD) · Momentum‑based GD · Nesterov accelerated GD · Stochastic GD · AdaGrad optimizer
- Unit 3 · Short notes Convolutional Neural Networks Introduction to Convolutional Neural Networks (CNN) and architectures · Convolutional terminology – ReLU activation, stride, padding, pooling, convolution operations · Convolutional kernels · Layer types – convolutional, pooling, fully‑connected · Visualization of CNNs · Canonical CNN models – LeNet, AlexNet, ZF‑Net, VGGNet, GoogLeNet, ResNet, R‑CNN
- Unit 4 · Short notes Recurrent Neural Networks Introduction to deep Recurrent Neural Networks (RNN) and architectures · Back‑Propagation Through Time (BPTT) · Vanishing and exploding gradient problems · Truncated BPTT · Gated Recurrent Units (GRUs) · Long Short‑Term Memory (LSTM) networks
- Unit 5 · Short notes Generative Models Introduction to deep generative models · Restricted Boltzmann Machines (RBMs) and Gibbs sampling · Deep belief networks · Markov networks and Markov chains · Autoregressive models – NADE, MADE, PixelRNN · Generative Adversarial Networks (GANs)
More for AL-503 (B)
Other 5th semester AIML subjects
- AL-501 Operating Systems
- AL-502 Database Management Systems
- AL-503 (A) Information Retrieval
- AL-503 (C) Optimization Techniques in Machine Leaning
- AL-504 (A) AI in Health Care
- AL-504 (B) Natural Language Processing
- AL-504 (C) Computational Intelligence