Image: Seoul skyline
Taking place in Seoul, South Korea next month, the International Conference on Machine Learning (ICML) is one of the world’s most important gatherings of academic and industrial professionals in the field, with a reputation for presenting and publishing cutting-edge research.
Three PhD students and five members of staff from the School’s Institute for Statistical Science are named on ICML papers, as well as a Compass CDT supervisor based in the School of Biological Sciences and, in what is a rare achievement for an undergraduate, a second-year degree student:
- ‘Adaptive Momentum and Nonlinear Damping for Neural Network Training’ , Dr Katerina Karoni, Senior Research Associate
- ‘Equilibrium Selection in Multi-Agent Policy Gradients via Opponent-Aware Basin Entry’, Arina Redina, MSci Mathematics with Study Abroad student
- ‘Rethinking the roles of dimensionality in topological data analysis and manifold learning’, Hannah Sansford, PhD graduate, Professor Nick Whiteley, Director of the Institute for Statistical Science and Compass CDT
- ‘Minimum Distance Summaries for Robust Neural Posterior Estimation’, Sherman Khoo, PhD student, Dr Dennis Prangle, Associate Professor in Statistics, Dr Song Liu, Associate Professor in Data Sciences and AI, Professor Mark Beaumont, Professor of Statistics, School of Biological Sciences
- ‘Zero Flow Encoders’, Yakun Wang, PhD student, Dr Song Liu, Associate Professor in Data Sciences and AI
Commenting on this year’s accepted papers, Professor Oliver Johnson, Head of the School of Mathematics and Professor of Information Theory, said: “To see colleagues and students at various stages of their research journeys recognised by such a well-regarded conference is testament to the breadth of talent and expertise we have in the School.”
"We are enormously proud of the impactful work that colleagues and students are undertaking in the exciting field of machine learning and AI – and in related areas such as data science, statistics and across the wider mathematical sciences "
The strong presence at ICML this July follows success at last year’s event, when five students and five academic supervisors from the School’s Compass CDT had papers accepted. Alongside student, alumni, and staff involvement in other recent conferences - including ICLR, NeurIPS, ISIT and IMS-APRM - it is indicative of Bristol’s leading role within the machine learning discipline.
“ICML is an important event for us, providing the Institute for Statistical Science colleagues and students with an opportunity to present their work, and to network with leading practitioners and researchers working in machine learning and related areas” said Professor Nick Whiteley, Director of the Institute and the Compass CDT, and Heilbronn Chair in Data Science.
“It’s fantastic to have such a strong presence this year and especially rewarding to see PhD students from both Compass, and the wider Institute, submitting high-quality papers. They have already had a productive few months, with lots of conference involvement, international collaborations, and successful industry placements and internships.
Two students were finalists in the Smith Institute’s TakeAIM competition, while another reached the finals of the STEM for Britain awards. Our Compass alumni, including recent graduates, have also continued to impress and inspire us with the versatility and diversity of their chosen careers.
Our involvement in ICML builds on these successes, providing those attending with a valuable opportunity to enhance their own careers prospects, technical knowledge and research capabilities, by networking with, and learning from, leading academics and practitioners.”