Home › AKTU PYQ › Machine Learning Techniques › PYQ Questions › Unit 4

Machine Learning Techniques Unit 4 – Artificial Neural Networks and Deep Learning: AKTU previous year questions

12 AKTU questions from Unit 4 (Artificial Neural Networks and Deep Learning) asked in 2025–2026, tagged by marks, year and topic. In short: about 23 marks of every paper come from this unit; the most asked topic is Derivation of Backpropagation Algorithm (8 times); 8 questions came back in a later year. The latest ones are listed below; on the page you can filter them by 2-mark or long questions, topic and repeats.

Unit 4 previous year questions

  • Define activation function in neural networks. (2 marks, 2026, Types of CNN Layers)
  • What is backpropagation in neural networks? (2 marks, 2026, Derivation of Backpropagation Algorithm)
  • Explain perceptron and multilayer perceptron with neat diagram. (7 marks, 2026, Multilayer Perceptron)
  • Derive the Backpropagation algorithm for training neural networks. (7 marks, 2026, Derivation of Backpropagation Algorithm)
  • Explain the architecture of Convolutional Neural Networks (CNN). (7 marks, 2026, Concept of Convolutional Neural Network)
  • Discuss applications of Deep Learning with suitable case studies. (7 marks, 2026, Introduction to Deep Learning)
  • Derive the Backpropagation Algorithm for Multilayer Neural Networks. (7 marks, 2026, Derivation of Backpropagation Algorithm)
  • Explain CNN architecture and different types of layers. (7 marks, 2026, Types of CNN Layers)
  • Explain the role of the activation function in a perceptron. (2 marks, 2025, Perceptrons)
  • What is gradient descent? (2 marks, 2025, Gradient Descent and the Delta Rule)
  • Discuss the Self-Organizing Map (SOM) algorithm. How does SOM work for clustering and dimensionality reduction? Explain the differences between SOM and supervised learning models. (7 marks, 2025, Unsupervised Learning: SOM Algorithm and Its Variants)
  • Describe the basic architecture of a convolution neural network (CNN) with help of diagram. (7 marks, 2025, Concept of Convolutional Neural Network)

Most asked Unit 4 topics

  • Derivation of Backpropagation Algorithm – asked 8 times
  • Types of CNN Layers – asked 6 times
  • Gradient Descent and the Delta Rule – asked 4 times

Unit 4 question pattern

  • 2-mark questions: 4 asked in 2025–2026
  • 7-mark questions: 8 asked in 2025–2026
  • About 23 marks from Unit 4 in every paper
  • 8 questions were asked again in a later year

Unit 4 questions year by year

  • 2026: 8 Unit 4 questions asked (46 marks across that year's papers)
  • 2025: 4 Unit 4 questions asked (18 marks across that year's papers)