Skip to content. | Skip to navigation

Personal tools

Sections

UMR 5672

logo de l'ENS de Lyon
logo du CNRS
You are here: Home / Seminars / Machine Learning and Signal Processing / Learning of narrow neural networks in high dimensions

Learning of narrow neural networks in high dimensions

Hugo Chao Cui (Assistant-doctorant, Laboratoire de physique statistique des systèmes computationnels)
When Jul 23, 2024
from 01:00 to 02:00
Attendees Hugo Chao Cui
Add event to calendar vCal
iCal

 

Hugo Chao Cui

Title: Learning of narrow neural networks in high dimensions

Abstract: This talk explores the interplay between neural network architectures and data structure through the lens of high-dimensional asymptotics. We introduce the sequence multi-index model, which encompasses as special instances several previously studied models of feed-forward fully-connected networks, but also auto-encoder and attention architectures. In the limit of large data dimension and comparably large number of samples, but finite number of hidden units, we derive a tight asymptotic characterization of the learning of this generic model. As an illustration, we discuss how this characterization enables the analysis of the learning of a dot-product attention layer. We show how the latter can learn to implement either a positional attention mechanism (with tokens attending to each other based on their respective positions), or a semantic attention mechanism (with tokens attending to each other based on their meaning), and evidence a phase transition with sample complexity from positional to semantic learning.

Website: https://people.epfl.ch/hugo.cui

In Room M7 101, 1st floor, Monod campus, ENSL.