School of Physical Sciences

Wed 04 Mar 14:30: Title to be confirmed

Tue, 02/12/2025 - 17:43

Title to be confirmed

Abstract not available

Speaker: Dr Lixin Sun and Dr Gregor Simm, Microsoft
Wednesday 04 March 2026, 14:30-15:30
Venue: Unilever Lecture Theatre, Yusuf Hamied Department of Chemistry.
Series: Theory - Chemistry Research Interest Group; organiser: Lisa Masters.

Add to your calendar or Include in your list

Thu 12 Feb 14:00: Title to be confirmed

Tue, 02/12/2025 - 11:22

Title to be confirmed

Abstract not available

Speaker: Dr Jeanne Colbois (CNRS)
Thursday 12 February 2026, 14:00-15:15
Venue: Seminar Room 3, RDC.
Series: Theory of Condensed Matter; organiser: Bo Peng.

Add to your calendar or Include in your list

Thu 15 Jan 14:00: High-arity PAC learning

Tue, 02/12/2025 - 07:45

High-arity PAC learning

In this third talk of the series on high-arity learning frameworks, I will discuss the high-arity PAC learning theory, which is motivated by PAC learning of graphs, hypergraphs and relational structures and is heavily inspired by (hyper)graph limits, and is characterized by a slicewise notion of the Vapnik—Chervonenkis dimension.

I will also discuss how exchangeability theory plays a crucial role in agnostic version of learning and a phenomenon exclusive to high-arity learning: the interplay between the partite and non-partite. Time permitting, I will also talk about what part of the theory extends to learning hypergraph limits.

No background in learning theory, model theory or hypergraph limits is required for this talk.

This talk is based on joint work with Maryanthe Malliaris.

Speaker: Leonardo Coregliano (University of Chicago)
Thursday 15 January 2026, 14:00-15:00
Venue: MR2, CMS.
Series: Discrete Analysis Seminar; organiser: Julia Wolf.

Add to your calendar or Include in your list

Wed 14 Jan 14:00: High-arity learning frameworks, an overview

Tue, 02/12/2025 - 07:44

High-arity learning frameworks, an overview

Classic PAC learning theory studies when we can make an accurate guess of a set based on finitely many i.i.d. samples from it. The Fundamental Theorem of Statistical Learning characterizes when such an accurate guess can be made in terms of the Vapnik—Chervonenkis dimension. A few extensions of the PAC learning framework were made to address the case when the sample are not independent but have “reasonable” correlation. However, in these attempts, correlation is seen as an obstacle to overcome in the learning task.

In this first talk of a series of three, I will present an overview of the new framework of high-arity learning, in which structured-correlation is used to increase the learning power. I will also talk about a connection of learning theory to hypergraph regularity lemmas via Haussler packing property.

No background in learning theory or regularity lemmas is required for this talk.

This talk is based on joint works with Maryanthe Malliaris and Caroline Terry.

Speaker: Leonardo Coregliano (University of Chicago)
Wednesday 14 January 2026, 14:00-14:45
Venue: MR2, CMS.
Series: Discrete Analysis Seminar; organiser: Julia Wolf.

Add to your calendar or Include in your list

Wed 14 Jan 14:45: Sample completion, Netflix Prize Competition and k-dependence

Tue, 02/12/2025 - 07:43

Sample completion, Netflix Prize Competition and k-dependence

In the Netflix Prize Competition (2006-2009), we are given a finite set U of users, a finite set M of movies and a partial function f: U x M ⇀ R, where f(u,m) indicates how user u rates movie m and we are tasked with completing f to a total function to predict how all users U rate all movies in M. Although some algorithms did fairly well in the competition, giving a satisfactory theoretical explanation for their success has been difficult.

In this second talk of the series on high-arity learning frameworks, I will discuss how the Netflix Prize Competition can be seen as an instance of a high-arity learning framework called “sample completion learning”. I will also discuss how sample completion learning is completely characterized by the model-theoretic notion of k-dependence introduced by Shelah (which can be seen as a high-dimensional version of the Vapnik—Chervonenkis dimension). In turn, this gives a full theoretical characterization of when the Netflix problem is “solvable”.

No background in learning theory or model theory is required for this talk.

This talk is based on joint work with Maryanthe Malliaris.

Speaker: Leonardo Coregliano (University of Chicago)
Wednesday 14 January 2026, 14:45-15:30
Venue: MR2, CMS.
Series: Discrete Analysis Seminar; organiser: Julia Wolf.

Add to your calendar or Include in your list

Wed 04 Feb 13:30: Short character sums evaluated at homogeneous polynomials (date tbc)

Tue, 02/12/2025 - 07:37

Short character sums evaluated at homogeneous polynomials (date tbc)

Let p be a prime. Bounding short Dirichlet character sums is a classical problem in analytic number theory, and the celebrated work of Burgess provides nontrivial bounds for sums as short as p1/4+ε for all ε>0. In this talk, we will first survey known bounds in the original and generalized settings. Then we discuss the so-called ``Burgess method’’ and present new results that rely on bounds on the multiplicative energy of certain sets in products of finite fields.

Speaker: Rena Chu (University of Göttingen)
Wednesday 04 February 2026, 13:30-14:30
Venue: MR4, CMS.
Series: Discrete Analysis Seminar; organiser: Julia Wolf.

Add to your calendar or Include in your list

Wed 03 Dec 13:30: Short character sums evaluated at homogeneous polynomials

Mon, 01/12/2025 - 18:24

Short character sums evaluated at homogeneous polynomials

Let p be a prime. Bounding short Dirichlet character sums is a classical problem in analytic number theory, and the celebrated work of Burgess provides nontrivial bounds for sums as short as p1/4+ε for all ε>0. In this talk, we will first survey known bounds in the original and generalized settings. Then we discuss the so-called ``Burgess method’’ and present new results that rely on bounds on the multiplicative energy of certain sets in products of finite fields.

Speaker: Rena Chu (University of Göttingen)
Wednesday 03 December 2025, 13:30-14:30
Venue: MR4, CMS.
Series: Discrete Analysis Seminar; organiser: Julia Wolf.

Add to your calendar or Include in your list

Thu 29 Jan 12:00: Public opinion and the art of asking questions TGM156 - Listening and Responding to Public Interest: A Mathsci-comm Network Event

Mon, 01/12/2025 - 15:30

Public opinion and the art of asking questions

TGM156 - Listening and Responding to Public Interest: A Mathsci-comm Network Event

Speaker: Gillian Prior (National Centre for Social Research)
Thursday 29 January 2026, 12:00-12:20
Venue: Seminar Room 1, Newton Institute.
Series: Isaac Newton Institute Seminar Series; organiser: nobody.

Add to your calendar or Include in your list

Thu 11 Dec 11:30: Anomalies and non-linear realization of conformal and global symmetries in DCFT BIDW03 - The Many Faces of Boundaries, Impurities, and Defects

Mon, 01/12/2025 - 14:30

Anomalies and non-linear realization of conformal and global symmetries in DCFT

Conformal defects break a part of the bulk conformal symmetry and, sometimes, a part of the bulk global symmetry. Perhaps more appropriately, these symmetries can be seen as realized non-linearly and lead to identities for integrated correlation functions of displacement and tilt operators. In this talk I will discuss a systematic classification of such identities and explain how they incorporate Weyl and global symmetry anomalies associated with the defect.

BIDW03 - The Many Faces of Boundaries, Impurities, and Defects

Speaker: Petr Kravchuk (King's College London)
Thursday 11 December 2025, 11:30-12:30
Venue: Seminar Room 1, Newton Institute.
Series: Isaac Newton Institute Seminar Series; organiser: nobody.

Add to your calendar or Include in your list

Mon 08 Dec 12:00: Exact solutions in superconductivity BIDW03 - The Many Faces of Boundaries, Impurities, and Defects

Mon, 01/12/2025 - 14:30

Exact solutions in superconductivity

I will show a new model, called Berry trashcan, which is relevant to both superconductivity in rhombohedral graphene, and has an exactly solvable limit in two dimensions.

BIDW03 - The Many Faces of Boundaries, Impurities, and Defects