Recent & Upcoming Talks

An overview of text-and-audio methods that supports the Multimodal Learning module in R255.

A live demo of our ongoing work on Music AI tools; more info here.

A talk on our recently-published general-purpose architecture, Perceiver AR, and why long context matters when processing real-world …

An overview of why few-shot learning on structured data is important and some examples of works I have contributed to that address …

An overview of several graph generation and probabilistic approaches, part of the R250 MPhil course at the Cambridge CS department.

A reading group-style session focused around MetaFun (Xu et al. ICML 2020).

Introducing a novel framework for learning graph representations while incorporating uncertainty modelling.

A novel method based on topology and GNNs for graph visualisation and pooling.

A short intro to graph generation methods in recent ML research.

Challenges and approaches to QA in realistic environments; an introduction to VideoNavQA, an alternative take on EQA.

Introducing the recently published VideoNavQA benchmark and study (BMVC 2019).

An overview of challenging QA tasks and an alternative take on EQA.