Tue, 25 Oct 2022
14:00
C3

Nonbacktracking spectral clustering of nonuniform hypergraphs

Dr. Phil Chodrow
(Department of Computer Science, Middlebury College)

Note: we would recommend to join the meeting using the Zoom client for best user experience.

Abstract

Spectral methods offer a tractable, global framework for clustering in graphs via eigenvector computations on graph matrices. Hypergraph data, in which entities interact on edges of arbitrary size, poses challenges for matrix representations and therefore for spectral clustering. We study spectral clustering for arbitrary hypergraphs based on the hypergraph nonbacktracking operator. After reviewing the definition of this operator and its basic properties, we prove a theorem of Ihara-Bass type which allows eigenpair computations to take place on a smaller matrix, often enabling faster computation. We then propose an alternating algorithm for inference in a hypergraph stochastic blockmodel via linearized belief-propagation which involves a spectral clustering step, again using nonbacktracking operators. We provide proofs related to this algorithm that both formalize and extend several previous results. We pose several conjectures about the limits of spectral methods and detectability in hypergraph stochastic blockmodels in general, supporting these with in-expectation analysis of the eigeinpairs of our studied operators. We perform experiments with real and synthetic data that demonstrate the benefits of hypergraph methods over graph-based ones when interactions of different sizes carry different information about cluster structure.

Joint work with Nicole Eikmeier (Grinnell) and Jamie Haddock (Harvey Mudd).

Tue, 01 Nov 2022
14:00
C3

Large network community detection by fast label propagation

Dr. Vincent Traag
(Leiden University)
Abstract

Many networks exhibit some community structure. There exists a wide variety of approaches to detect communities in networks, each offering different interpretations and associated algorithms. For large networks, there is the additional requirement of speed. In this context, the so-called label propagation algorithm (LPA) was proposed, which runs in near linear time. In partitions uncovered by LPA, each node is ensured to have most links to its assigned community. We here propose a fast variant of LPA (FLPA) that is based on processing a queue of nodes whose neighbourhood recently changed. We test FLPA exhaustively on benchmark networks and empirical networks, finding that it runs up to 700 times faster than LPA. In partitions found by FLPA, we prove that each node is again guaranteed to have most links to its assigned community. Our results show that FLPA is generally preferable to LPA.

Lecture banner

Wednesday 2 November, 5 pm, Mathematical Institute, Oxford

A calculator processes numbers without caring that these numbers refer to items in our shopping, or the calculations involved in designing an airplane. Number without context is a remarkable abstraction that we learn as infants and which has profoundly affected our world.

Rectangular eigenvalue problems
Hashemi, B Nakatsukasa, Y Trefethen, L Advances in Computational Mathematics volume 48 issue 6 (16 Nov 2022)
Reply to Comment on ‘Quantum principle of relativity’
Dragan, A Ekert, A New Journal of Physics volume 24 issue 9 098002 (01 Sep 2022)
Fri, 02 Dec 2022
10:00
L6

Closest Point of Approach problem

Dr. Nikhil Banda MIOA and Dan Pollard
(Drumgrange)
Abstract

Consider an environment with two vehicles/platforms moving at a relative velocity (v). The objective is to predict the Closest Point of Approach (CPA) between the two platforms as defined by the parameters: CPA time (t0), CPA bearing (θ0), CPA distance (r0)[†].The challenge is to identify mathematical operations - either using geometric methods, or by use of tracking algorithms such as Kalman Filters (EKF, UKF), or a combination of both - to estimate the CPA parameters. The statistical errors in estimation of CPA parameters also need to be quantified with each observations at time ti. The signals to be employed are acoustic in nature and the receiver platform has one sensor. The parameters that can extracted from acoustic signals are current relative bearing (θ) and current doppler or range rate (S) 


[†]Defined currently using polar coordinate system.

Correction to: The space of barcode bases for persistence modules
Jacquard, E Nanda, V Tillmann, U Journal of Applied and Computational Topology volume 7 issue 1 31-31 (26 Mar 2023)
AdS Virasoro-Shapiro from dispersive sum rules
Alday, L Hansen, T Silva, J JOURNAL OF HIGH ENERGY PHYSICS volume 2022 issue 10 (05 Oct 2022)
Estimation of heterogeneous instantaneous reproduction numbers with application to characterize SARS-CoV-2 transmission in Massachusetts counties
Zhou, Z Kolaczyk, E Thompson, R White, L PLoS Computational Biology volume 18 issue 9 (01 Sep 2022)
Snap-induced morphing: from a single bistable shell to the origin of shape bifurcation in interacting shells
Liu, M Domino, L Dupont de Dinechin, I Taffetani, M Vella, D Journal of the Mechanics and Physics of Solids volume 170 (25 Oct 2022)
Subscribe to