• Topics
• Video Tutorials
• Slides
• Links

Artificial Intelligence

• Intelligent Agents and their Environments
• The concept of a Rational Agent
• Specifying the Task environment
• (PEAS description)
• Different characteristics of environments
• Fully vs Partially observable
• Static vs Dynamic
• Episodic vs Sequential
• Different types of agents : Reflex, Goal-based, Utility-based
• Search - Formulating a search problem
• Uninformed Search strategies
• BFS, DFS, DLS, ID-DFS
• Their working principles
• Complexities
• Relative advantages and disadvantages
• Informed (heuristic) Search strategies
• Greedy Best-first search
• A* search: Working principle
• Characteristics of heuristics (admissibility and consistency)
• Proof of A*’s optimality
• Local search: Hill Climbing
• Searching with non-deterministic actions
• AND-OR search trees
• Searching with partial observability
• Belief state-space search
• Adversarial Search - Formulation of a Game tree
• The minimax algorithm
• Alpha-Beta pruning: Its rationale
• Working principle
• Move ordering and Search cut-off
• Probabilistic Reasoning
• Bayes’ rule and its uses
• Bayesian Network: Building a Bayes-net
• Making inference from it
• Markov Chains and Hidden Markov Models
• Transition and Sensor models
• Building and HMM
• Applications of HMM
• Inference in temporal models
• Filtering and Prediction
• Most Likely explanations (Viterbi algorithm)
• Particle Filters: basic working principle
• Making Decisions - Decision theory and Utility theory
• Lottery, Utility functions
• Maximum Expected Utility principle
• Constraints of Utility (Orderability, Transitivity)
• Markov Decision Processes: Policies, Rewards
• Optimal policies and the Utility of States
• Value Iteration
• Supervised learning: Training set, Test set
• Overfitting and Underfitting
• Decision trees: Basic understanding
• Learning a Decision tree through entropy calculation
• Nearest Neighbor classifier
• Basic working principle
• Relative advantages and disadvantages
• Naive Bayes classifier
• Basic working principle
• Calculating classification procedures
• Relative advantages and disadvantages
• Artificial Neural Network: Basic working principle
• Basic structure and calculation of a perceptron
• Basics of backpropagation algorithm
• Support Vector Machines: Basic working principle
• Unsupervised Learning (Clustering)
• Basic concepts and applications of Clustering
• Different types of Clustering: Partitional vs. Hierarchical
• Exclusive vs Overlapping vs Fuzzy
• Complete vs Partial
• K-means Clustering: Basic working principle
• Characteristics, advantages, disadvantages
• Agglomerative Hierarchical Clustering: Basic concepts
• Representations (Dendrograms and Nested cluster diagrams)
• Different techniques to define cluster proximity
• Single link, Complete link, Group average
• Centroid method, their relative advantages and disadvantages
• DBSCAN: Basic principle and applications
• Classification of points (Core, Border and Noise)
• Understanding basics of Reinforcement Learning
• MDPs, Policies, Rewards, Utilities
• Passive and Active Reinforcement Learning
• Exploration and Exploitation
• Adaptive Dynamic Programming
• Temporal Difference Learning
• Q-Learning