KTU 2024 Scheme · Semester 6 · Program Core — Artificial Intelligence
Introduction to Deep Learning (PBCAT604) Syllabus
Official KTU 2024 Scheme syllabus for Introduction to Deep Learning, Semester 6, Program Core — Artificial Intelligence (Artificial Intelligence).
Course Code
PBCAT604
Credits
4
Teaching Hours
3:0:0:1 (L:T:P:R)
CIE Marks
60
ESE Marks
40
Exam Duration
2 Hrs 30 Min
Prerequisites
PCCST503 (Machine Learning)
Semester
Semester 6
Course Objective
To teach the basics of neural networks along with advanced topics including recurrent neural networks, long short-term memory cells and convolutional neural networks, and to enable students to complete programming solutions for real-world problems.
Module-wise Syllabus
Module 1
10 contact hoursNeural Networks — introduction to human and artificial intelligence: history of AI, forms of learning — supervised and unsupervised, perceptron learning rule, bio-inspired learning, artificial neural networks, backpropagation, multi-layer perceptron model, activation functions, loss functions, optimization. Training neural networks — gradient descent, stochastic gradient descent, momentum, weight initialization, batch normalization, hyperparameter optimization, parameter updates, model ensembles.
Module 2
12 contact hoursOverview of deep learning: deep feedforward networks and training deep models, including optimization techniques — Gradient Descent (GD), GD with momentum, Stochastic GD, AdaGrad, RMSProp and Adam. Regularization methods — L1 and L2, early stopping, dataset augmentation, parameter sharing and tying, input noise injection, ensemble methods, dropout and parameter initialization.
Module 3
12 contact hoursConvolutional Neural Networks: convolution layer, pooling layer, fully connected layer, ConvNet; case study of the ImageNet challenge — LeNet, AlexNet, VGG, GoogLeNet, ResNet, Inception Net, EfficientNet. Regularization techniques, data augmentation — zooming, rotation, cropping, blurring, noise addition; self-supervision techniques, semi-supervised and weakly supervised learning, adversarial training. Transfer learning — freezing the input layers, fine-tuning output layers.
Module 4
10 contact hoursDeep Unsupervised Learning and Recent Trends: autoencoders (standard, sparse, denoising, contractive, etc.), variational autoencoders, Generative Adversarial Networks, autoencoders and DBM, multi-task deep learning, multi-view deep learning. Suggested project topics: applications of deep learning to computer vision and to NLP.
Course Outcomes
- CO1Explain the fundamental principles of neural networks and discuss the practical challenges associated with them.
- CO2Explain the common regularization and optimization methods used in deep neural networks.
- CO3Use deep neural networks with CNNs and RNNs for tasks such as object detection, image segmentation and text-related problems.
- CO4Determine which deep learning algorithms are best suited for different types of learning tasks across domains.
- CO5Implement deep learning algorithms and solve real-world problems.
Assessment Pattern (CIE: 60 marks, ESE: 40 marks)
Continuous Internal Evaluation (CIE)
| Attendance | 5 |
| Project | 30 |
| Internal Examination 1 (Written) | 12.5 |
| Internal Examination 2 (Written) | 12.5 |
End Semester Examination (ESE)
Total 40 marks, 2 Hrs 30 Min. See the official KTU syllabus document for the exact Part A / Part B question pattern for this course.
Textbooks & Reference Books
Textbooks
- Deep Learning — Ian Goodfellow, Yoshua Bengio, Aaron Courville (MIT Press, 1st edition, 2016)
- Neural Networks and Deep Learning — Michael A. Nielsen (Determination Press, 1st edition, 2015)
- Pattern Recognition and Machine Learning — C. M. Bishop (Springer, 1st edition, 2006)
- Deep Learning with Python — Francois Chollet (Manning Publications, 2nd edition, 2017)
Reference Books
- Neural Networks: A Classroom Approach — Satish Kumar (Tata McGraw-Hill Education, 1st edition, 2005)
- Artificial Neural Networks — B. Yegnanarayana (PHI Learning, 1st edition, 2009)
Frequently Asked Questions
How many credits is KTU Introduction to Deep Learning (PBCAT604)?
4 credits, with 3:0:0:1 (L:T:P:R) teaching hours per week, under the KTU 2024 Scheme.
How many modules are in the PBCAT604 syllabus?
4 modules, 44 total contact hours.
What is the CIE and ESE mark split for this course?
CIE (Continuous Internal Evaluation): 60 marks. ESE (End Semester Examination): 40 marks, 2 Hrs 30 Min. Total: 100 marks.
What are the recommended textbooks for PBCAT604?
Deep Learning (Ian Goodfellow, Yoshua Bengio, Aaron Courville); Neural Networks and Deep Learning (Michael A. Nielsen); Pattern Recognition and Machine Learning (C. M. Bishop); Deep Learning with Python (Francois Chollet).
Is this syllabus specific to one branch, or common to others too?
This Semester 6 course is listed under Program Core — Artificial Intelligence at KTU under the 2024 Scheme — check the course header above for which branches it's common to.
Need help with KTU Introduction to Deep Learning?
Learnizo offers live, 1-on-1 online tuition for KTU AI subjects — matched to your exact module and semester.