Disordered systems insights on computational hardness

D Gamarnik, C Moore… - Journal of Statistical …, 2022 - iopscience.iop.org
In this review article we discuss connections between the physics of disordered systems,
phase transitions in inference problems, and computational hardness. We introduce two …

Sampling with flows, diffusion, and autoregressive neural networks from a spin-glass perspective

D Ghio, Y Dandi, F Krzakala, L Zdeborová - Proceedings of the National …, 2024 - pnas.org
Recent years witnessed the development of powerful generative models based on flows,
diffusion, or autoregressive neural networks, achieving remarkable success in generating …

Reducibility and statistical-computational gaps from secret leakage

M Brennan, G Bresler - Conference on Learning Theory, 2020 - proceedings.mlr.press
Inference problems with conjectured statistical-computational gaps are ubiquitous
throughout modern statistics, computer science, statistical physics and discrete probability …

Inference of dynamic hypergraph representations in temporal interaction data

A Kirkley - Physical Review E, 2024 - APS
A range of systems across the social and natural sciences generate data sets consisting of
interactions between two distinct categories of items at various instances in time. Online …

Fundamental limits to learning closed-form mathematical models from data

O Fajardo-Fontiveros, I Reichardt… - Nature …, 2023 - nature.com
Given a finite and noisy dataset generated with a closed-form mathematical model, when is
it possible to learn the true generating model from the data alone? This is the question we …

Marvels and pitfalls of the langevin algorithm in noisy high-dimensional inference

S Sarao Mannelli, G Biroli, C Cammarota, F Krzakala… - Physical Review X, 2020 - APS
Gradient-descent-based algorithms and their stochastic versions have widespread
applications in machine learning and statistical inference. In this work, we carry out an …

Community detection in bipartite networks with stochastic block models

TC Yen, DB Larremore - Physical Review E, 2020 - APS
In bipartite networks, community structures are restricted to being disassortative, in that
nodes of one type are grouped according to common patterns of connection with nodes of …

Passed & spurious: Descent algorithms and local minima in spiked matrix-tensor models

SS Mannelli, F Krzakala, P Urbani… - … on machine learning, 2019 - proceedings.mlr.press
In this work we analyse quantitatively the interplay between the loss landscape and
performance of descent algorithms in a prototypical inference problem, the spiked matrix …

Exact phase transitions for stochastic block models and reconstruction on trees

E Mossel, A Sly, Y Sohn - Proceedings of the 55th Annual ACM …, 2023 - dl.acm.org
In this paper, we rigorously establish the predictions in ground breaking work in statistical
physics by Decelle, Krzakala, Moore, Zdeborová (2011) regarding the block model, in …

Limits and performances of algorithms based on simulated annealing in solving Sparse hard inference problems

MC Angelini, F Ricci-Tersenghi - Physical Review X, 2023 - APS
The planted-coloring problem is a prototypical inference problem for which thresholds for
Bayes-optimal algorithms, like belief propagation (BP), can be computed analytically. In this …