RTU Kota B.Tech AI 5th Semester Information Theory and Coding Question Paper 2025

RTU Artificial Intelligence Information Theory and Coding 2025 Paper Review

Preparing for the Rajasthan Technical University B.Tech Information Theory and Coding exam requires a strict understanding of probability, data compression, and mathematical channel capacities. For Artificial Intelligence students, this subject provides the absolute theoretical basis for machine learning loss functions (like cross-entropy), data dimensionality reduction, and efficient feature encoding. You cannot fully understand the mathematical limits of neural network data processing without a firm grasp of Shannon's theorems and mutual information. The 2025 paper tests your capability to calculate entropy bounds, draw Huffman coding trees, and construct matrix equations for error-correcting codes. Reviewing this specific branch paper shows you exactly how examiners frame the questions and allocate marks across the mathematical modules. This systematic preparation helps you approach your fifth-semester exam confidently.

Understanding the AI Branch Exam Pattern

The RTU theory examination is a three-hour paper worth 70 marks. The paper features three distinct sections designed to evaluate both basic definitions and quantitative algorithmic problems.

• Part A: This section contains ten compulsory questions worth two marks each. You must state specific definitions like mutual information, differentiate between source coding and channel coding, define Hamming distance, or write the formula for channel capacity under 30 words.
• Part B: You will find seven questions here. You must answer five of them. Each question is worth four marks. Your answers require proving simple entropy properties, calculating the capacity of a Binary Symmetric Channel (BSC), or generating the parity check matrix for a simple block code.
• Part C: This section offers five major questions. You need to answer three. Each question carries ten marks. These require you to execute multi-step Huffman or Lempel-Ziv coding algorithms, calculate the syndrome for linear block codes, or draw the state diagram and trellis for a convolutional encoder.

Core Topics Evaluated in the AI Paper

The 2025 question paper covers several critical modules that establish the mathematical rules for data compression and transmission. Focus your study time on these specific areas to maximize your score.

Information Measure and Entropy

This module evaluates your understanding of fundamental probability in data. You must master the calculations for self-information, Shannon entropy, joint entropy, conditional entropy, and mutual information. Practice proving the mathematical properties of entropy. Examiners heavily test your ability to calculate these values given a joint probability matrix for a transmitter and receiver.

Source Coding Algorithms

Source coding focuses on removing data redundancy. You must be able to calculate the coding efficiency and redundancy for any given algorithm. Practice executing the Shannon-Fano coding and Huffman coding algorithms step by step for a given set of message probabilities. The 2025 paper heavily features ten-mark questions asking you to build the coding tree, list the final binary codewords, and prove the average codeword length.

Channel Capacity and Continuous Channels

This module bridges discrete probability with physical channel limits. Study the transition probability matrices for the Binary Symmetric Channel (BSC) and Binary Erasure Channel (BEC). You must understand Shannon's Channel Coding Theorem. Practice numerical problems calculating the maximum theoretical capacity using the Shannon-Hartley theorem for continuous channels with additive white Gaussian noise:

$$C = B \log_2\left(1 + \frac{S}{N}\right)$$

Linear Block Codes

Error control coding is heavily tested for logical matrix execution. You must fully understand how to construct a Generator Matrix ($G$) and a Parity Check Matrix ($H$) for an $(n, k)$ block code. Practice encoding a message vector into a codeword and executing syndrome decoding to detect and correct single-bit errors at the receiver.

Cyclic and Convolutional Codes

These represent advanced error correction methods. For cyclic codes, you must know how to generate codewords using generator polynomials instead of matrices. Practice polynomial division to find the remainder. For convolutional codes, study the shift-register implementation. Expect questions asking you to draw the state transition diagram, tree diagram, or trellis diagram for a given encoder rate and constraint length.

Answer Writing Strategy for High Marks

RTU evaluators look for clean probability matrices, properly structured coding trees, and logical step-by-step matrix operations. Use a blue pen for your general text and mathematical steps, and use a black pen and ruler for drawing Huffman trees, channel diagrams, and block matrices.

In Part A, answer directly. If a question asks for the definition of channel capacity, state clearly that it is the tightest upper bound on the amount of information that can be reliably transmitted over a communications channel, and write its mathematical notation.

In Part B, use clear probability tables. When proving properties of mutual information, explicitly write down the conditional probability formulas before substituting variables to make the logic visually scannable for the checker.

In Part C, precision in calculation is critical. When solving a ten-mark linear block code problem, clearly separate your matrix multiplication steps. Write down the message vector, show the multiplication with the generator matrix explicitly, and state the final codeword clearly. When drawing a Huffman coding tree, label every branch with its correct binary assignment (0 for left/up, 1 for right/down) and list the final codes in a summary table. Draw a clean box around your final efficiencies and matrix results.

Time Management During the Exam

Allocate 20 minutes to Part A. Spend 40 minutes on Part B. Reserve the remaining 120 minutes for the three long-answer questions in Part C. Executing polynomial divisions for cyclic codes, computing logarithmic entropy values, and drawing large trellis diagrams requires steady focus and significant time to prevent arithmetic mistakes. This plan guarantees you 40 minutes per major question, giving you time to cross-verify your matrix additions (modulo-2 arithmetic) and probability summations. Use the final 10 minutes to verify your question numbering, ensure all matrix rows and columns align correctly, and check that you have not skipped any intermediate steps in your probability proofs.