Topic: What makes a good embedding?

Speaker: Professor Patrick Rubin-Delanchy

• Biography

Professor Patrick Rubin-Delanchy is the chair of Statistical Learning at the University of Edinburgh as of January 2024. His research interests include data exploration, embedding and machine learning with a particular focus on structure discovery using AI, for example, correlations, clusters, hierarchy, trends, or manifold structure; in complex data such as large relational databases, dynamic networks, or high-dimensional data (e.g. tables with many columns, text, images). Applications of this research are wide-ranging, and he has won funding (over £7M between government & industry) for applications in biosciences, healthcare, (cyber-)security, societal resilience, environmental protection and more.

• Abstract of the talk

Embeddings are continuous vector representations of entities, such as words or nodes, perhaps most widely known for their role in modern AI systems such as large language models. In this talk I consider a different goal, which is facilitating statistical analysis, and the creation of knowledge. An embedding is an instrument which allows us to observe complex, unstructured, or otherwise intractable data, in a way that we can use. In embeddings, classical (e.g. Gaussian) statistical models are tenable; concepts like similarity, or trend, have a ‘shape’; abstract notions such as political opinion, the health of a patient, the function of a cell, can be made geometric and measurable; and we can uncover truths that could have seemed completely absent from the unprocessed data. I illustrate these points with new theory connecting statistical models, embeddings and the manifold hypothesis, and with motivating problems in science, security, and recent work with Southmead hospital at Bristol.

Rosie Wilkie, Head of External Engagement, School of Mathematics will introduce the latest data science research and expertise at the University of Edinburgh.

Agenda:

• Welcome
• Data Science at University of Edinburgh, Rosie Wilkie
• What makes a good embedding?, Patrick Rubin-Delanchy
• Audience questions
• Close

Book to attend below!

To book your place, "Register as a Guest" or use a University of Edinburgh login. For assistance, please contact us on eievents@ei.ed.ac.uk