MLMI 17: Advanced Computer Vision

Term: Lent 2026
Time: 12-1pm, Thursdays
Location: JDB Seminar Room
Office Hours: Fridays, ~~1-2PM (BE4-57)~~ 3-4PM (BE4-54)

Module Leaders

Elliott Wu
Assistant Professor

Ayush Tewari
Assistant Professor

Aims

This course explores recent advances in computer vision, focusing on the foundational concepts that drive modern research. The aim is to equip students with both the theoretical grounding and practical intuition needed to understand cutting-edge research and begin contributing to novel research projects in the field.

Syllabus

Session	Reading Material	Slides
Part I: Representation Learning and Generative Models (Ayush Tewari)
Overview of Modern Computer Vision Supervised Learning	1. Foundations of Computer Vision (Chapter 1, 9, 24) 2. An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale 3. Segment Anything	[Slides]
Representation Learning	1. Foundations of Computer Vision (Chapter 30) 2. Masked Autoencoders Are Scalable Vision Learners 3. Emerging Properties in Self-Supervised Vision Transformers 4. [Video Explanation] Further Readings: 5. Deep Clustering for Unsupervised Learning of Visual Features 6. Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning	[Slides1], [Slides2]
Generative Models in Computer Vision	1. Foundations of Computer Vision (Chapter 32) 2. Denoising Diffusion Probabilistic Models 3. Classifier-Free Diffusion Guidance 4. Diffusion Transformers with Representation Autoencoders	[Slides1]
Part II: 3D Vision & Foundation Models (Elliott Wu)
3D Computer Vision	1. Foundations of Computer Vision (Chapter 39, 40, 44) 1+. The Fundamental Matrix Song 2. VGGT: Visual Geometry Grounded Transformer 3. NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis	[Slides]
Vision Foundation Models and Applications	[TBD]	[TBD]

Assessment

Coursework	Weight	Release	Deadline
Coursework 1: Representation Learning and Generative Models	50%	[PDF]	23 February 2026
Coursework 2: 3D Vision	50%	[PDF]	24 March 2026