skip to content

Microstructural Kinetics Group

Department of Materials Science & Metallurgy
 

Tue 11 Feb 16:00: Spacetime extensions in low regularity

School of Physical Sciences - Mon, 27/01/2025 - 20:31
Spacetime extensions in low regularity

Abstract not available

Add to your calendar or Include in your list

Tue 11 Feb 16:00: Spacetime extensions in low regularity

School of Physical Sciences - Mon, 27/01/2025 - 20:31
Spacetime extensions in low regularity

Abstract not available

Add to your calendar or Include in your list

Wed 05 Feb 14:00: Generalization and Informativeness of Conformal Prediction

School of Physical Sciences - Mon, 27/01/2025 - 15:28
Generalization and Informativeness of Conformal Prediction

A popular technique for uncertainty quantification is conformal prediction, which converts point predictions into set predictions that are guaranteed to contain the true label of a test input with a user-defined probability. However, the size of the predicted set—-and thus the informativeness of the prediction—-is not controlled. In this talk, we present a theoretical connection between the informativeness of conformal prediction sets and generalization properties of the underlying model. Furthermore, we extend this analysis to conformal risk control and covariate shifts. The results provide insight into the effect of task-specific quantities and algorithmic hyperparameters, which we also illustrate via experiments.

Add to your calendar or Include in your list

Wed 12 Mar 16:00: Title to be confirmed

School of Physical Sciences - Mon, 27/01/2025 - 11:03
Title to be confirmed

Abstract not available

Add to your calendar or Include in your list

Fri 31 Jan 13:00: The Maxwell equations on the full Kerr black hole family

School of Physical Sciences - Mon, 27/01/2025 - 10:27
The Maxwell equations on the full Kerr black hole family

We discuss a proof of uniform boundedness and decay statements for solutions to the Maxwell equations on Kerr black holes. The proof is unconditional in the full subextremal |a| less than M family, in view of earlier joint work with Yakov Shlapentokh-Rothman. For extremal |a|=M Kerr, it is conditional on a conjecture for the spin ±1 Teukolsky equations motivated by work of Gajic and Casals—Gralla—Zimmerman. This is joint work with Gabriele Benomio (GSSI).

Add to your calendar or Include in your list

Thu 13 Feb 14:30: A new algebraic approach to the wreath conjecture

School of Physical Sciences - Mon, 27/01/2025 - 09:36
A new algebraic approach to the wreath conjecture

In 1970s, Baranyai proved that the hyperedges of the k-uniform complete hypergraph on n vertices can be decomposed into perfect matchings whenever k divides n. In the same paper, he posed a more general conjecture. Katona, who later rephrased this conjecture as decomposing [n]^(k) into so-called wreaths, wrote “Baranyai’s brilliant idea was to use matrices and flows in networks. This conjecture, however, seems to be too algebraic. One does not expect to solve it without algebra. (Unless it is not true.)”. In this talk, we will discuss a new algebraic approach to the wreath conjecture, defining a matrix encoding the problem and studying its properties. (based on joint work with Pavel Turek)

Add to your calendar or Include in your list

Thu 13 Feb 14:30: A new algebraic approach to the wreath conjecture

School of Physical Sciences - Mon, 27/01/2025 - 09:35
A new algebraic approach to the wreath conjecture

In 1970s, Baranyai proved that the hyperedges of the k-uniform complete hypergraph on n vertices can be decomposed into perfect matchings whenever k divides n. In the same paper, he posed a more general conjecture. Katona, who later rephrased this conjecture as decomposing [n]^(k) into so-called wreaths, wrote “Baranyai’s brilliant idea was to use matrices and flows in networks. This conjecture, however, seems to be too algebraic. One does not expect to solve it without algebra. (Unless it is not true.)”. In this talk, we will discuss a new algebraic approach to the wreath conjecture, defining a matrix encoding the problem and studying its properties. (based on joint work with Pavel Turek)

Add to your calendar or Include in your list

Thu 06 Feb 14:30: Decomposing Latin squares into transversals.

School of Physical Sciences - Mon, 27/01/2025 - 09:33
Decomposing Latin squares into transversals.

A Latin square of order n is an n x n grid filled with n symbols such that each symbol appears exactly once in each row and column. A transversal in a Latin square of order n is a collection of n cells such that each row, column and symbol appears exactly once in the collection.

Latin squares were introduced by Euler in the 1700s and he was interested in the question of when a Latin square decomposes fully into transversals.

We’ll discuss some of the history of this question, including some recent joint work with Richard Montgomery.

Add to your calendar or Include in your list

Wed 05 Mar 13:30: Title tbc

School of Physical Sciences - Mon, 27/01/2025 - 07:38
Title tbc

Abstract not available

Add to your calendar or Include in your list

Thu 27 Mar 14:00: Title to be confirmed

School of Physical Sciences - Sun, 26/01/2025 - 17:20
Title to be confirmed

Abstract not available

Add to your calendar or Include in your list

Tue 28 Jan 13:00: Debiased regression adjustment in completely randomized experiments with moderately high-dimensional covariates

School of Physical Sciences - Sun, 26/01/2025 - 12:17
Debiased regression adjustment in completely randomized experiments with moderately high-dimensional covariates

Completely randomized experiment is the gold standard for causal inference. When the covariate information for each experimental candidate is available, one typical way is to include them in covariate adjustments for more accurate treatment effect estimation. In this paper, we investigate this problem under the randomization-based framework, i.e., that the covariates and potential outcomes of all experimental candidates are assumed as deterministic quantities and the randomness comes solely from the treatment assignment mechanism. Under this framework, to achieve asymptotically valid inference, existing estimators usually require either (i) that the dimension of covariates p grows at a rate no faster than O(n3/4) as sample size n→∞; or (ii) certain sparsity constraints on the linear representations of potential outcomes constructed via possibly high-dimensional covariates. In this paper, we consider the moderately high-dimensional regime where p is allowed to be in the same order of magnitude as n. We develop a novel debiased estimator with a corresponding inference procedure and establish its asymptotic normality under mild assumptions. Our estimator is model-free and does not require any sparsity constraint on potential outcome’s linear representations. We also discuss its asymptotic efficiency improvements over the unadjusted treatment effect estimator under different dimensionality constraints. Numerical analysis confirms that compared to other regression adjustment based treatment effect estimators, our debiased estimator performs well in moderately high dimensions.

Add to your calendar or Include in your list

Fri 07 Feb 14:00: Identification and Estimation of Graphical Continuous Lyapunov Models Note unusual location

School of Physical Sciences - Sun, 26/01/2025 - 12:13
Identification and Estimation of Graphical Continuous Lyapunov Models

Graphical continuous Lyapunov models offer a new perspective on modeling causally interpretable dependence structure in multivariate data by treating each independent observation as a one-time cross-sectional snapshot of a temporal process. Specifically, the models consider multivariate Ornstein-Uhlenbeck processes in equilibrium. This leads to Gaussian models in which the covariance matrix is determined by the continuous Lyapunov equation. In this setting, each graphical model assumes a sparse drift matrix with support defined by a directed graph. The talk will discuss the identifiability of such sparse drift matrices and their regularized estimation.

Note unusual location

Add to your calendar or Include in your list

Tue 25 Feb 14:00: Branching random walk with non-local competition

School of Physical Sciences - Sat, 25/01/2025 - 16:33
Branching random walk with non-local competition

Abstract not available

  • Speaker: Sarah Penington (Bath)
  • Tuesday 25 February 2025, 14:00-15:00
  • Venue: MR12.
  • Series: Probability; organiser: ww295.

Add to your calendar or Include in your list