Book3

COMPUTER SCIENCE INTERVIEW PREPARATION SERIES MONOGRAPH 03 / 06 BOOK 3 · ACADEMIC & TECHNICAL RECRUITMENT CSE Lecturer Interview AI, Machine Learning, and Advanced CS Search and Reasoning, Statistical Machine Learning, Deep Learning, Data Mining, Computer Vision, and Bioinformatics CORE SUBJECT ARCHITECTURE 01 Classical Artificial Intelligence Heuristic search, A* admissibility, adversarial games, constraint satisfaction, and probabilistic logic. 02 Statistical Machine Learning Bias-variance decomposition, linear and kernel meth- ods, ensemble trees, and evaluation metrics. 03 Deep Learning Architectures Backpropagation, optimization dynamics, CNNs, sequence models, attention, and transformers. 04 Data Mining & Information Retrieval Clustering, association rules, dimensionality reduction, inverted indices, and ranking models. 05 Computer Vision & Bioinformatics Image filtering, feature descriptors, object detection, sequence alignment, and phylogenetic trees. 06 Advanced Algorithms & Research Viva NP-completeness reductions, approximation bounds, randomized algorithms, and thesis defence. VOLUME SCOPE AND METHODOLOGY This monograph targets advanced undergraduate and graduate topics

Contents How to Use This Book iv 0.1 The Five-Rung Follow-Up Ladder . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv 0.2 Panel Evaluation Rubric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv 0.3 What changed in AI/ML interviewing (2024 to 2026) . . . . . . . . . . . . . . . . . v 0.4 Notation used throughout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi I Artificial Intelligence 1 1 Artificial Intelligence: Search, Heuristics and Reinforcement Learning 2 1.1 State-Space Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.2 Uninformed search: the comparison table you must be able to redraw . . . . . . . 3 1.3 BFS versus Best-First versus A-star . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.4 Admissibility, consistency and the optimality proof . . . . . . . . . . . . . . . . . . 4 1.5 Designing heuristics: relaxation, dominance, pattern databases . . . . . . . . . . . 5 1.6 Adversarial search: minimax, alpha–beta, and what replaced them . . . . . . . . . 6 1.7 Constraint satisfaction: search plus inference . . . . . . . . . . . . . . . . . . . . . 8 1.8 Reinforcement learning . . . . . . . .

Contents Book 3: AI, ML & Advanced CS 3.1 The multilayer perceptron, drawn and written . . . . . . . . . . . . . . . . . . . . . 23 3.2 Backpropagation as a computation graph . . . . . . . . . . . . . . . . . . . . . . . 23 3.3 Convolution: the arithmetic, drawn . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.4 Attention and the transformer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 3.5 Why adding more layers does not always improve a model . . . . . . . . . . . . . . 29 3.6 Optimisers and training diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 3.7 Parameter-efficient fine-tuning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 IV Data Mining, Information Retrieval and Computer Vision 32 4 Data Mining and Information Retrieval 33 4.1 KDD: mining is one box out of seven . . . . . . . . . . . . . . . . . . . . . . . . . . 33 4.2 Association rule mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 4.3 Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 4.4 Information retrieval: from the inverted index to RAG . . . . . . . . . . . . . . . . 35 5 Comput