Reading Group and Research Seminar


Thursdays from 11:15 to 12:05. In person only.

This page contains information and the schedule for the Bath AI & ML reading group and research seminar. Every other week one student or member of staff will present a paper. The paper for the next session will be mailed out in advance. It would be great if attendees could have read the paper beforehand, but the speaker will explain the paper so this is not strictly necessary. This is intended to be a friendly and supportive environment where we collectively explore a new idea. Do what you can to make it comfortable for the speaker, especially if they are junior.

A list of forthcoming speakers is presented at the bottom of this page. The order is randomized. We understand that there are conference deadlines and that people go on vacations. If you can't make your week, then it is up to you to swap with someone else further down the list.

Why is attending this important? All great academic groups are more than the sum of their parts. This is a venue where you can meet and discuss ideas with people from your own and other groups. Throughout the year you will be exposed to 26 new ideas. Many of these will be interesting but not directly influence your work, but every now and then you will get an idea that you can apply to your own research. In addition, the ensuing discussions will mean that you become familiar with everyone else's expertise. Maybe knowledge that you have could be useful to someone else writing a NeurIPS paper. Co-authorship beckons. More than the sum of our parts. In addition, if you are a PhD student, you are going to need a job at the end of your degree. A broader knowledge of ML can only help with this. In short, the expected payback from this is very good even if it doesn't feel like that week to week.

Reading Group

Preparation: The organizer (currently Sophia Jones) will contact you two weeks in advance to ask for the paper that you will present. You can choose whatever you want, but please choose something general/important enough that it will be of interset to most of the group. If you don't have any ideas, a suggested list of papers is kept here. Feel free to add any interesting looking papers to the list! The room is booked for a full hour, but it's assumed that most presentations will take about 20 minutes, leaving plenty of time for discussion. Slides are encouraged, but working through the PDF of the paper is also acceptable.

Schedule for 2025


Date Presenter Paper
Future speakers
  1. Lukas Macha
  2. Professor Neill Campbell
  3. Daniel Beechey
  4. Ferdie Krammer
  5. Dr Wenbin Li
  6. Dil Jagpal
  7. Professor Mike Tipping
  8. Dr Harish Tayyar Madabushi
  9. Professor Özgür Şimşek
  10. Jack R Saunders
  11. Tom Ryder
  12. Alexandros Rotsidis
  13. Dr Andrew Barnes
  14. Doug Tilley
  15. Dr Alexander Power
  16. Emma Li
  17. Matthew Hewitt
  18. Ningchao Wang
  19. Yuchen Lu
  20. Isaac Flower
  21. Philip Lorimer
  22. Jundan Luo
  23. Dr Georgios Exarchakis
  24. James Elson
  25. Dr Xi Chen
  26. Dona Shaji

Past Talks


Date Presenter Paper
14 Dec  2023 Prof Simon Prince On the Origin of Implicit Regularization in Stochastic Gradient Descent Samuel L. Smith, Benoit Dherin, David G. T. Barrett, Soham De, ICLR 2021
18 Jan  2024 Sophia Jones Direct Preference Optimization: Your Language Model is Secretly a Reward Model Rafael Rafailov, Archit Sharma, Eric Mitchell, Stefano Ermon, Christopher D. Manning, Chelsea Finn, NeurIPS 2023
25 Jan  2024 Josh Evans Creating Multi-Level Skill Hierarchies in Reinforcement Learning Joshua Benjamin Evans and Özgür Şimşek, NeurIPS 2023
15 Feb  2024 Jack McKinlay Reinforcement Learning As a Framework for Ethical Decision Making David Abel, James MacGlashan and Michael L. Littman, AAAI Workshop: AI, Ethics, and Society 2016
29 Feb  2024 Dr Tom Fincham Haines Interventions, Where and How? Experimental Design for Causal Models at Scale Panagiotis Tigas, Yashas Annadani, Andrew Jesson, Bernhard Schölkopf, Yarin Gal, Stefan Bauer, NeurIPS 2022
14 Mar  2024 Tom Cannon On the Measure of Intelligence François Chollet
28 Mar  2024 George Fletcher TURBO: The Swiss Knife of Auto-Encoders Guillaume Quétant, Yury Belousov, Vitaliy Kinakh, Slava Voloshynovskiy
11 Apr  2024 Yifan Li Simplifying Transformer Blocks Bobby He, Thomas Hofmann, ICLR 2024
16 May  2024 Fahid Mohammed Chain-of-Thought Prompting Elicits Reasoning in Large Language Models Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Brian Ichter, Fei Xia, Ed Chi, Quoc Le, Denny Zhou, NeurIPS 2022
6 June  2024 Jessica Nicholson BYOL-Explore: Exploration by Bootstrapped Prediction Zhaohan Daniel Guo, Shantanu Thakoor, Miruna Pîslar, Bernardo Avila Pires, Florent Altché, Corentin Tallec, Alaa Saade, Daniele Calandriello, Jean-Bastien Grill, Yunhao Tang, Michal Valko, Rémi Munos, Mohammad Gheshlaghi Azar, Bilal Piot, NeurIPS 2024
27 June  2024 Dr Marina De Vos Explainable AI is Dead, Long Live Explainable AI! Tim Miller, Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency
11 July  2024 Siyuan Gou When Do Transformers Shine in RL? Decoupling Memory from Credit Assignment Tianwei Ni, Michel Ma, Benjamin Eysenbach, Pierre-Luc Bacon, NeurIPS 2023
18 July  2024 Syeda Zahra Making AI Less "Thirsty": Uncovering and Addressing the Secret Water Footprint of AI Models Pengfei Li, Jianyi Yang, Mohammad A. Islam, Shaolei Ren
26 Sep  2024 Dr Julian Padget A Framework for Understanding Sources of Harm throughout the Machine Learning Life Cycle Harini Suresh, John V. Guttag, Proceedings of the 1st ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization, 2021.
10 Oct  2024 Thomas Smith Flow Network based Generative Models for Non-Iterative Diverse Candidate Generation Emmanuel Bengio, Moksh Jain, Maksym Korablyov, Doina Precup, Yoshua Bengio, NeurIPS 2021.
24 Oct  2024 Callum Pitceathly METRA: Scalable Unsupervised RL with Metric-Aware Abstraction Seohong Park, Oleh Rybkin, Sergey Levine, ICLR 2024
7 Nov  2024 Dr Jie Zhang Learning-Augmented Dynamic Power Management with Multiple States via New Ski Rental Bounds Antonios Antoniadis, Christian Coester, Marek Eliáš, Adam Polak, Bertrand Simon, NeurIPS 2021
12 Dec  2024 Akshil Patel Learning Hierarchical Features from Generative Models Shengjia Zhao, Jiaming Song, Stefano Ermon, ICML 2017