Representations learnt through deep neural networks tend to be highly informative, but opaque in terms of what information they learn to encode. We introduce an approach to probabilistic modelling that learns to represent data with two separate deep representations: an invariant representation that encodes the information of the class from which the data belongs, and an equivariant representation that encodes the symmetry transformation defining the particular data point within the class manifold (equivariant in the sense that the representation varies naturally with symmetry transformations). This approach is based primarily on the strategic routing of data through the two latent variables, and thus is conceptually transparent, easy to implement, and in-principle generally applicable to any data comprised of discrete classes of continuous distributions (e.g. objects in images, topics in language, individuals in behavioural data). We demonstrate qualitatively compelling representation learning and competitive quantitative performance, in both supervised and semi-supervised settings, versus comparable modelling approaches in the literature with little fine tuning.
Ilya Feige (Faculty)
Ilya leads Faculty's machine learning Research & Development, and is joint appointed as a Senior Research Fellow at UCL in computer science. He works on both ML Capabilities and AI Safety research, as well as leading Faculty's development of AI Safety tooling.
Related Events (a corresponding poster, oral, or spotlight)
2019 Oral: Invariant-Equivariant Representation Learning for Multi-Class Data »
Wed Jun 12th through Thu the 13th Room Grand Ballroom