Séminaires Quantiques IVADO

IVADO propose une nouvelle série de séminaires qui débutera au printemps 2025.

Ces séminaires quantiques IVADO ont pour objectif de stimuler la collaboration entre les communautés de recherche en physique quantique et en intelligence artificielle. Ils s’inscrivent dans le cadre de l’Alliance en Algorithmique Quantique dont IVADO est membre.

Ces séminaires auront généralement lieu les vendredis peu après midi.

Ils se tiennent en présence et un lunch est offert.

On demande aux participants de s’inscrire et d’indiquer leur choix de repas.

Bienvenue!

Dr. Aram Harrow (Center for Theoretical Physics, MIT)

Résumé : The idea of quantum computers is that « More is different » when it comes to qubits. One qubit is not so interesting but many of them together can create exotic and computationally powerful forms of many-body entanglement. Given this, we might expect that many-body physics and quantum information would often be related. I will describe two recent examples.
1. The Ising model is a simple model of a magnet. But it turns out to also describe the competition between quantum interactions creating entanglement and measurements destroying entanglement. This can tell us how about the power of near-term quantum computers.
2. How can a closed system reach thermal equilibrium? There have been many answers to this question dating back to the 19th century. I will explain how entanglement is a plausible source of thermalization.

Alexander Schmidhuber (Center for Theoretical Physics, MIT)

Résumé : One of the central challenges in quantum computing is to identify new computational problems for which quantum algorithms offer exponential speedups over any classical method. In this talk, I’ll argue that algebraic topology provides a surprisingly rich source of such problems. I will focus on two recent developments in this directions.

The first concerns Khovanov homology [1], a categorification of the Jones polynomial and a powerful knot invariant that also appears as a physical observable in 4D supersymmetric Yang-Mills theory. I’ll present a quantum algorithm that combines techniques from quantum algorithms for the Jones polynomial and recent advances in quantum homology computation.

The second is the estimation of persistent Betti numbers [2,3] — a core subroutine of Topological Data Analysis (TDA) that captures the shape of data across scales. I’ll discuss quantum algorithms for estimating these invariants, as well as recent complexity-theoretic results that clarify the limitations and potential of quantum approaches in TDA.

No prior background in algebraic topology is required.

Based on:

• [1] A quantum algorithm for Khovanov homology, Schmidhuber et al., arXiv:2501.12378

• [2] Complexity-theoretic limitations on quantum algorithms for topological data analysis, Schmidhuber & Lloyd, arXiv:2209.14286

• [3] Quantum computing and persistence in topological data analysis, Gyurik, Schmidhuber, et al., arXiv:2410.21258

Risi Kondor (Department of Computer Science and Department of Statistics, University of Chicago)

Résumé : TBA

Dr. Martin Larroca (Postdoctoral Research Associate at Los Alamos National Laboratory)

Résumé : Kostka, Littlewood-Richardson, Plethysm and Kronecker coefficients are the multiplicities of irreducible representations in the decomposition of representations of the symmetric group that play an important role in representation theory, geometric complexity and algebraic combinatorics. We give quantum algorithms for computing these coefficients whenever the ratio of dimensions of the representations is polynomial and study the computational complexity of this problem. We show that there is an efficient classical algorithm for computing the Kostka numbers under this restriction and conjecture the existence of an analogous algorithm for the Littlewood-Richardson coefficients. We argue why such a classical algorithm does not straightforwardly work for the Plethysm and Kronecker coefficients and conjecture that our quantum algorithms lead to superpolynomial speedups for these problems. The conjecture about Kronecker coefficients was disproved by Greta Panova in [arXiv:2502.20253] with a classical solution which, if optimal, points to a O(n^{4+2k}) vs Ω(n^{4k^2+1}) polynomial gap in quantum vs classical computational complexity for an integer parameter k.

Reference: https://arxiv.org/abs/2407.17649

Mark Wilde (School of Electrical and Computer Engineering, Cornell University)

Résumé : TBA