Module COMP3801: Deep Learning & Computer Vision

Department: Computer Science

COMP3801: Deep Learning & Computer Vision

Prerequisites

• COMP2261 Artificial Intelligence
• COMP2271 Data Science

Corequisites

• None

Excluded Combination of Modules

• None

Aims

• To enable students to approach complex ill-defined problems that require deep layers of learning, informed by principles of natural and artificial learning systems.
• To equip students with the practical skills to design, train, and test neural networks using modern deep learning frameworks.
• To enable students to critically assess and apply modern deep learning and computer vision techniques across established and emerging application areas.
• To provide students with the knowledge to interpret and solve real-world problems, such as in image and video understanding, using appropriate computational approaches.

Content

• Principles of deep learning: core concepts in representation learning, neural architectures, generalisation theory, and connections to natural learning systems.
• Practical deep learning: implementation and experimentation using industry-standard frameworks (e.g. PyTorch), with a focus on architecture design, training strategies, and model evaluation.
• Advanced and emerging models: topics include generative models (e.g. adversarial, diffusion, variational, and flow-based models), sequential models, Transformers, and implicit representations.
• Applied computer vision: application of deep learning techniques to vision tasks such as image classification, object detection and tracking, 3D reconstruction, scene understanding, and visual navigation.
• Contemporary issues and challenges: Exploration of real-world deployment challenges, such as dataset bias, generalisation to unseen data, and system-level integration of vision models.

Learning Outcomes

• On completion of this module, students will be able to:
• Demonstrate a broad understanding of the principles and architectures underpinning state-of-the-art deep neural networks and computer vision applications.
• Explain and critically evaluate modern generative modelling approaches and computer vision techniques, and how they are applied in real-world systems.
• Demonstrate in-depth knowledge of the contemporary deep learning and computer vision topics presented, how these are applicable to relevant industrial problems and have future potential for emerging needs in both a research and industrial setting.

• On completion of the module, students will be able to:
• Design, implement, and evaluate deep learning models using modern frameworks to solve difficult and ill-defined problems.
• Critically analyse dataset issues, model performance and deployment challenges (“in-the-wild”), identifying problems such as bias, overfitting, function design, and generalisation.

• On completion of the module, students will be able to:
• Understand and effectively communicate technical information, using appropriate terminology and referencing current research or industry practices.
• Use research and industry standard material within both existing and new system scenarios.

Modes of Teaching, Learning and Assessment and how these contribute to the learning outcomes of the module

• Lectures enable the students to learn new material relevant to deep learning and computer vision as well as their applications.
• Computer classes enable students to acquire necessary coding skills, develop a practical understanding of the underpinning theory, learn how to effectively use relevant libraries, and receive feedback on their work.
• Summative coursework assessments assess the application of methods and techniques, the knowledge of deep learning libraries, and the ability to train large neural networks on modern GPUs.

Teaching Methods and Learning Hours

Summative Assessment

Formative Assessment:

Example formative exercises are given during the course.

■ Students who do not attend monitored activities shown under Teaching Methods and Learning Hours, or who fail to complete the summative or formative assessment(s) specified above, may be subject to the Academic Progress procedures defined in the University's General Regulation V, and may be required to leave the University.