Tue, 23 May 2023

11:00 - 12:00
L3

SDEs and rough paths on manifolds

Emilio Rossiferrucci
Abstract

I will begin by speaking about Ito SDEs on manifolds, how their meaning depends on the choice of a connection, and an example in which the Ito formulation is preferable to the more common Stratonovich one. SDEs are naturally generalised to the case of more irregular driving signals by rough differential equations (RDEs), i.e. equations driven by rough paths. I will explain how it is possible to give a coordinate-invariant definition of rough integral and rough differential equation on a manifold, even in the case of arbitrarily low regularity and when the rough path is not geometric, i.e. it does not satisfy a classical integration by parts rule. If time permits, I will end on a more recent algebraic result that makes it possible to canonically convert non-geometric RDEs to geometric ones.

Tue, 23 May 2023

16:00 - 17:00
L6

Moments of the high order derivatives of CUE characteristic polynomials

Fei Wei
(University of Oxford)
Abstract

In this talk, I will firstly give asymptotic formulas for the moments of the n-th derivative of the characteristic polynomials from the CUE. Secondly, I will talk about the connections between them and a solution of certain Painleve differential equation. This is joint work with Jonathan P. Keating.
 

Major adverse cardiovascular events (MACE) in patients with severe COVID-19 registered in the ISARIC WHO clinical characterization protocol: a prospective, multinational, observational study
Reyes, L Garcia-Gallo, E Murthy, S Fuentes, Y Serrano, C Ibáñez-Prada, E Lee, J Rojek, A Citarella, B Gonçalves, B Dunning, J Rätsep, I Viñan-Garces, A Kartsonaki, C Rello, J Martin-Loeches, I Shankar-Hari, M Olliaro, P Merson, L Journal of Critical Care volume 77 (09 May 2023)
Polynomial bounds for chromatic number VII. Disjoint holes
Chudnovsky, M Scott, A Seymour, P Spirkl, S Journal of Graph Theory volume 104 issue 3 499-515 (14 May 2023)
A fast randomized algorithm for computing an approximate null space
Park, T Nakatsukasa, Y BIT Numerical Mathematics volume 63
Automatic Alignment of Local Representations
Teh, Y Roweis, S NIPS 2002: Proceedings of the 15th International Conference on Neural Information Processing Systems 841-848 (01 Jan 2002)
MULTIPLICATIVE INTERACTIONS AND WHERE TO FIND THEM
Jayakumar, S Czarnecki, W Menick, J Schwarz, J Rae, J Osidnero, S Teh, Y Harley, T Pascanu, R 8th International Conference on Learning Representations, ICLR 2020 (01 Jan 2020)
FUNCTIONAL REGULARISATION FOR CONTINUAL LEARNING WITH GAUSSIAN PROCESSES
Titsias, M Schwarz, J de Matthews, A Pascanu, R Teh, Y 8th International Conference on Learning Representations, ICLR 2020 (01 Jan 2020)
Photo of Dewi

Symmetry underpins all physics research. We look for fundamental and beautiful patterns to describe and explain the laws of nature. One way of explaining symmetry is to ask: "what is the full set of operations I can do to my real-world experiment or abstract theory written on paper that doesn't change any physical measurements or predictions?'' There are simple symmetries we are perhaps already familiar with. For example, lab-based physics experiments usually don't care if you wait an hour to do the experiment or if you rotate your apparatus by 90 degrees.

Tue, 16 May 2023

11:00 - 12:00
L3

DLA and related models, part II

Dmitry Belyaev
Abstract

This will be a continuation of the talk from last week (9 May). 

Subscribe to