Lectures in course CS-456: Deep reinforcement learning

Exploration versus Exploitation

Explores the balance between exploring new possibilities and exploiting known rewarding actions in reinforcement learning.

Reinforcement Learning: SARSA Algorithm

Explores the SARSA algorithm for reinforcement learning, focusing on updating Q-values and the importance of exploration in learning by rewards.

Reinforcement Learning: BackUp Diagrams

Introduces the BackUp diagram as a key graphic representation in reinforcement learning.

Markov Decision Processes: Foundations of Reinforcement Learning

Covers Markov Decision Processes, their structure, and their role in reinforcement learning.

Markov Decision Processes: Dynamic Programming Techniques

Discusses Markov Decision Processes and dynamic programming techniques for solving optimal policies in various scenarios.

Policy Iteration and Linear Programming in MDPs

Discusses policy iteration and linear programming methods for solving Markov Decision Processes.

General Introduction to Artificial Neural Networks

Covers the history and inspiration behind artificial neural networks, the structure of neurons, learning through synaptic connections, and the mathematical description of artificial neurons.

General Introduction to Artificial Neural Networks: Image classification

Introduces artificial neural networks for image classification using layered feedforward and recurrent networks.

Deep Reinforcement Learning: Policy Gradient and Value Function

Explains the implementation of the REINFORCE algorithm with a baseline using a neural network.

XOR Problem: Neural Networks

Delves into solving the XOR problem using a two-layer neural network.

Actor-Critic Architecture and Advantage-Actor-Critic

Explores actor-critic networks and the advantage actor critic method for optimizing parameters to maximize return.

Multilayer Networks: First Steps

Covers the preparation for deriving the Backprop algorithm in layered networks using multi-layer perceptrons and gradient descent.

Eligibility Traces in Policy Gradient Algorithms

Delves into the emergence of eligibility traces in policy gradient algorithms, making learning rapid and efficient.

Eligibility Traces for Policy Gradient and Actor-Critic

Explores eligibility traces in policy gradient and actor-critic architectures, leading to an elegant online learning rule.

BackProp Algorithm: Pseudocode and Processing Steps

Covers the BackProp algorithm, including initialization, signal propagation, error computation, weight updating, and complexity comparison with numerical differentiation.

Neural Networks for Action Learning: Three-Factor Rules and Dopamine

Explores neural networks learning by reward, actor-critic structures, synaptic plasticity, and the role of dopamine in synaptic changes.

Three-factor rules: DeepRL1.5A

Explains three-factor rules in policy gradient algorithms and their implementation in biological and hardware systems.

Neural Networks for Action Learning: Brain Implementation Insights

Explores neural networks for action learning and the brain's reinforcement learning implementation.

General Introduction into Artificial Neural Networks: part 3

Covers learning by rewards in deep reinforcement learning without math details.

The Problem of Overfitting

Discusses the problem of overfitting in deep networks and the importance of controlling flexibility to avoid it.