RTU Kota B.Tech 6th Semester Natural Language Processing Question Paper 2024 (AI)

RTU Natural Language Processing 2024 Paper Review

Preparing for the Rajasthan Technical University B.Tech Natural Language Processing exam requires a firm grasp of computational linguistics, probabilistic modeling, and semantic interpretation. For Computer Science, AI, and IT students building chatbots, sentiment analysis tools, or search engines, understanding how machines parse and generate human language is foundational. You cannot build a reliable language interface without understanding morphological rules, syntax trees, and context disambiguation.

The 2024 paper tests your capability to compute N-gram probabilities, construct parse trees using context-free grammars, and apply algorithms for word sense disambiguation. This targeted preparation strategy helps you approach the exam confidently.

Understanding the Exam Pattern

The RTU theory examination is a three-hour paper worth 70 marks. The paper features three distinct sections designed to evaluate both foundational linguistic definitions and quantitative parsing algorithms.

• Part A: Ten compulsory questions, two marks each. You must define terms like lemmatization, state the difference between a bigram and a trigram model, define a context-free grammar (CFG), or explain part-of-speech (POS) tagging in under 30 words.
• Part B: Seven questions; answer five. Each is worth four marks. Your answers require explaining the Chomsky hierarchy, tracing a basic top-down parsing algorithm, or differentiating between lexical and structural ambiguity with examples.
• Part C: Five major questions; answer three. Each is worth ten marks. These require you to execute the CYK (Cocke-Younger-Kasami) parsing algorithm for a given string and grammar, calculate the smoothing probabilities for an N-gram language model, or detail the architecture of a machine translation system.

Core Topics Evaluated in the Paper

Focus your study time on these specific modules to maximize your score.

Morphology and Finite-State Transducers

This module evaluates your understanding of word formation. You must master the difference between inflectional and derivational morphology. Practice tracing finite-state automata (FSA) and finite-state transducers (FST) used for morphological parsing. Understand how regular expressions are converted into automata for tokenization and stemming.

N-Gram Language Models and Smoothing

Language models predict the probability of the next word. You must understand the Markov assumption and how to calculate bigram and trigram probabilities from a given corpus count. The paper features numerical problems requiring you to apply Laplace (Add-1) smoothing to handle zero-probability N-grams:

$$P_{Laplace}(w_i | w_{i-1}) = \frac{c(w_{i-1}, w_i) + 1}{c(w_{i-1}) + V}$$

where $V$ is the vocabulary size.

Syntax Analysis and Parsing Algorithms

Parsing maps linear text to hierarchical structures. You must master Context-Free Grammars (CFG) and treebank construction. Practice executing the CYK algorithm, which requires converting the grammar to Chomsky Normal Form (CNF) and building a dynamic programming parse table. Study the differences between top-down, bottom-up, and shift-reduce parsers.

Semantic Analysis and Disambiguation

This module focuses on extracting meaning. You must understand how to represent meaning using First-Order Logic or conceptual dependencies. Study Word Sense Disambiguation (WSD) algorithms, particularly the Lesk algorithm and Naive Bayes classifiers for determining the correct meaning of a polysemous word based on its context window.

Answer Writing Strategy for High Marks

RTU evaluators look for clean syntax trees, explicitly stated probability formulas, and clear algorithmic trace tables.

• Blue/Black Pen Usage: Use a blue pen for text explanations and calculation lines. Use a black pen and ruler for drawing parse trees, finite-state machine states, and dynamic programming grids.
• Part A Strategy: Answer directly. If a question asks for the definition of a morpheme, state clearly that it is the smallest meaning-bearing unit of language.
• Part B Strategy: Use clear graphical structures. When explaining structural ambiguity, provide a classic example sentence like "I saw the man with the telescope" and quickly draw two distinct parse trees to visually prove the two different syntactical interpretations.
• Part C Strategy: Precision in calculation is critical. When solving a ten-mark CYK parsing problem, draw the complete triangular dynamic programming table. Fill the diagonal with the lexical rules, and show the combinations of non-terminals for the upper cells. If the start symbol $S$ appears in the top-right cell, state clearly that the string is accepted by the grammar.

Time Management During the Exam

• Part A (20 minutes): Answer these first to secure your foundation marks.
• Part B (40 minutes): Spend 8 minutes per question.
• Part C (120 minutes): Reserve the remaining time for the three long-answer questions. Drawing extensive parse trees, constructing finite-state transducers, and calculating matrix cells for parsing algorithms requires steady focus and significant writing time. This distribution guarantees you 40 minutes per major question, giving you time to double-check your grammar derivations. Use the final 10 minutes to verify your question numbering and ensure all tree branches connect to the correct terminal symbols.