(SEM VII) THEORY EXAMINATION 2019-20 ARTIFICIAL INTELLIGENCE

SECTION A – Explanation

Section A of the Artificial Intelligence paper is intended to test the student’s basic knowledge and understanding of core AI concepts and terminology. All questions in this section are compulsory and require short but conceptually correct answers. The examiner expects students to recall definitions, differences, and simple justifications related to artificial intelligence without going into deep mathematical or algorithmic details.

The questions in this section include the history of artificial intelligence, description of an optimal problem with example, definition of utility theory, explanation of statistical learning models, definition of Bayes classifier, justification of the use of searching in games, and difference between propositional logic and predicate logic. These questions cover a wide range of AI fundamentals, from historical development to reasoning, learning, and decision-making.

For example, the history of AI question checks awareness of important milestones such as the Dartmouth Conference and early AI programs. The optimal problem question evaluates understanding of goal-oriented problem solving. Utility theory tests knowledge of decision making under uncertainty. The Bayes classifier question checks understanding of probabilistic reasoning. Logic-related questions assess knowledge of representation and reasoning. Answers in this section must be brief, precise, and written in correct AI terminology. Long explanations are not required, but conceptual clarity is essential to score full marks.

SECTION B – Explanation

Section B evaluates the student’s conceptual clarity, analytical ability, and application of AI algorithms and reasoning techniques. Students are required to attempt any three questions, which allows them to choose topics according to their strengths. The questions in this section involve numerical computation, algorithm explanation, and logical representation.

The questions in Section B include Principal Component Analysis (PCA) with numerical computation for a given two-dimensional dataset, explanation of the hill climbing algorithm along with its drawbacks and improvements, and translation of English sentences into predicate logic and causal form. These questions test both mathematical understanding and logical reasoning.

For instance, the PCA question requires students to reduce dimensionality by computing principal components, which checks understanding of statistical learning and data preprocessing. The hill climbing algorithm question tests knowledge of local search techniques and awareness of problems such as local maxima and plateaus. The predicate logic translation question evaluates the ability to convert natural language statements into formal logical expressions, which is a core skill in knowledge representation.

 Answers in Section B should be written in a logical flow, starting with definitions or concepts, followed by explanation or calculations. Each answer typically spans about one and a half to two pages, with clarity and correctness being crucial.

SECTION C – Explanation

Section C is the most important and highest-weight section of the Artificial Intelligence paper. This section tests the student’s in-depth understanding, analytical skills, and ability to explain AI techniques in detail. Each question provides internal choices, and students must attempt only one part from each question.

The questions in Section C cover advanced and application-oriented topics such as machine learning with explanation of supervised and unsupervised learning, Naïve Bayes model, learning with hidden data using the EM algorithm, comparison between Linear Discriminant Analysis (LDA) and logistic regression, and explanation of clustering with emphasis on the k-means clustering technique. These topics represent core AI learning paradigms and statistical methods.

For example, the machine learning question tests understanding of learning types and real-world examples. The Naïve Bayes and EM algorithm questions evaluate probabilistic learning and parameter estimation with hidden variables. The LDA versus logistic regression question checks understanding of classification techniques and their assumptions. The clustering question requires explanation of unsupervised learning and iterative optimization in k-means. Answers in Section C should be detailed, well-structured, and written with clarity. Each answer generally extends over two to three pages and significantly affects the final score.

Overall Understanding of the Paper Pattern

The Artificial Intelligence (RCS-702) question paper is structured to test students progressively from basic understanding to advanced analytical and learning concepts. Section A focuses on foundational definitions and short explanations, Section B evaluates algorithmic understanding and logical reasoning, and Section C tests deep understanding of machine learning, probabilistic models, and clustering techniques. Students who understand this structure can prepare effectively by revising fundamentals for Section A, practicing algorithms and logic representation for Section B, and mastering detailed explanations of learning models for Section C.

A strong preparation strategy for this subject includes understanding search algorithms, probabilistic reasoning, machine learning models, and clustering techniques. Section C carries the maximum weight and requires special attention for scoring high marks.