This course offers a probabilistic perspective to machine learning, applied to visual computing applications. A central theme is the probabilistic inference and probabilistic modeling of images, human pose and motion, and 3D shapes and appearances. Topics include
* classification (recap),
* probabilistic regression,
* factorized representations for human pose estimation, and
* generative models, such as variational autoencoders (VAE), generative adversarial networks (GANs), and diffusion models (DDIM & DDPM, classifier free and classifier-based guidance).
The students will present recent papers on computer graphics and computer vision that have a strong probabilistic component. Practical experiences for applying probabilistic models in PyTorch are gained in the assignments and tutorials.
The course is largely based on the Visual AI course I gave at UBC: https://www.cs.ubc.ca/~rhodin/2022_2023_CPSC_533R/
The following slides provide additional details and context to the existing visual computing modules at Bielefeld University:
https://docs.google.com/presentation/d/e/2PACX-1vRGyE0j_cU324qp-vOEE4hEBPn3HdwliZQqeg_vT0w7Yt9EUXmgFB6NamL_ynqeBMYRgD-cdvfvXobA/pub?start=false&loop=false&delayms=3000
Required skills: Excellent Python programming skills (>1 year), fundamental mathematics, and basics of machine learning or neural networks.
Optional skills: PyTorch or TensorFlow experience, computer graphics, computer vision (lectures or projects).
Students need not be experts in all aspects of visual computing but have a strong background in one of graphics, vision, and machine learning. These skills are typically acquired through lectures or by private projects, hackathons, and coding competitions.
The papers that will be discussed are from the CVPR, SIGGRAPH, and NeurIPS conferences, or other respected computer vision and graphics venues.
The (online) book Probabilistic Machine Learning - An Introduction by Kevin P. Murphy serves as the basis for the lectures:
https://probml.github.io/pml-book/book1.html
A few chapters from the follow-up book Probabilistic Machine Learning - Advanced topics by Kevin P. Murphy are relevant too:
https://probml.github.io/pml-book/book2.html
Frequency | Weekday | Time | Format / Place | Period |
---|
Module | Course | Requirements | |
---|---|---|---|
39-M-Inf-AI-app Applied Artificial Intelligence | Applied Artificial Intelligence: Seminar | Student information | |
- | Graded examination | Student information | |
39-M-Inf-AI-app-foc Applied Artificial Intelligence (focus) | Applied Artificial Intelligence (focus): Seminar | Student information | |
- | Ungraded examination Graded examination | Student information | |
39-M-Inf-INT-app Applied Interaction Technology | Applied Interaction Technology: Seminar | Student information | |
- | Graded examination | Student information | |
39-M-Inf-INT-app-foc Applied Interaction Technology (focus) | Applied Interaction Technology (focus): Seminar | Student information | |
- | Ungraded examination Graded examination | Student information |
The binding module descriptions contain further information, including specifications on the "types of assignments" students need to complete. In cases where a module description mentions more than one kind of assignment, the respective member of the teaching staff will decide which task(s) they assign the students.
Degree programme/academic programme | Validity | Variant | Subdivision | Status | Semester | LP | |
---|---|---|---|---|---|---|---|
Studieren ab 50 |
The focus of this seminar is on reading, presenting, and discussing seminal papers in machine learning (ML), computer vision (CV), and computer graphics (CG). Topics and the papers to present are picked to give a broad overview of the current state of the art and how their components have been developed in the past.
The instructors will give a few lectures during the first half of the term, to teach the essentials to students of all backgrounds (stronger in ML, CV, or CG).
Students present one paper, lead the discussion of one paper, and engage in the discussion of the papers presented by others, including the writing style, strengths, limitations, and ethical implications.
A report and small assignments written in PyTorch support the practical understanding of probabilistic models, as typical for seminars+tutorial.
A corresponding course offer for this course already exists in the e-learning system. Teaching staff can store materials relating to teaching courses there: