Image by Matthis Lehmkühler.

Frontiers of Statistical Mechanics and Theoretical Computer Science 2021

Online and in-person, December 13-15, 2021

Organizers: Tyler Helmuth (Durham) and Will Perkins (UIC)

Description

A central question in theoretical computer science concerns computational phase transitions that separate parameter regimes for which efficient algorithms do/do not exist. It is not a coincidence that this terminology has been borrowed from the physical sciences: there are, in some instances, rigorous links between computational and physical phase transitions. This confluence has led to an exciting research area exploring these connections and transferring ideas between fields.

This mini-workshop series is aimed at promoting further interaction between the theoretical computer science and mathematical physics communities. We hope this contributes to the development of new techniques for algorithms, statistical mechanics, and combinatorics, as well as the discovery of further common research interests.


The workshop will be held in a satellite format: talks will be given online via Zoom on Monday December 13 - Wednesday December 15, 2021 from 3:00pm-6:00pm UTC time (10am-1pm Eastern Time; 4pm-7pm Central European Time). Participants are welcome to attend talks remotely from anywhere in the world; there will also be satellite workshops in Chicago and Durham (and other locations TBD) at which particpants can watch the talks together and then partipcate in discussions and working groups in-person. If you're interested in organizing a satellite meeting, please contact us and we will assist in any way possible.


2020 Workshop: Uniqueness Methods in Statistical Mechanics

Registration

Registration is now closed; to participate, please email the organizers for the zoom link.

Program

The workshop will feature three introductory survey talks and eight research talks, with the remaining time reserved for discussions and work in the satellite groups.
Schedule with abstracts (pdf)

All times below are UTC time (same as London time)

Monday December 13:

  • 14:55-15:00: Introductory remarks
  • 15:00-15:40: Jian Ding (University of Pennsylvania): Recent progress on random field Ising model
  • 15:50-16:30: Wojciech Samotij (Tel Aviv): The upper tail for triangles in random graphs
  • 16:40-17:20: Tselil Schramm (Stanford): In hopes of systematizing information-computation gaps

Tuesday December 14:

  • 15:00-15:30: Christophe Garban (Lyon): Continuous symmetry breaking along the Nishimori line
  • 15:40-16:10: Jonathan Leake (Weierstrass Institute): Sampling matrices from HCIZ densities
  • 16:20-16:50: Holden Lee (Duke): Changing the temperature for algorithm design
  • 17:00-17:30: Reza Gheissari (Berkeley): Low-temperature Ising dynamics from random initializations

Wednesday December 15:

  • 15:00-15:30: Vishesh Jain (Stanford): Approximate counting and sampling via local central limit theorems
  • 15:40-16:10: Gal Kronenberg (Oxford): Independent sets in random subgraphs of the hypercube
  • 16:20-16:50: Mark Jerrum (Queen Mary): Mixing in frustrated systems at low temperatures
  • 17:00-17:30: Nike Sun (MIT): Generalized Ising perceptron models

Acknowledgements

The workshop is supported in part by the NSF through the UIC TRIPODS Foundations of Data Science Institute (CCF-1934915) and NSF award DMS-1847451.