BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Random Matrices and Tensors for Large-Dimensional Statistical Lear
 ning
DTSTART:20260505T091500Z
DTEND:20260505T101500Z
DTSTAMP:20260607T053100Z
UID:indico-event-14480@indico.math.cnrs.fr
DESCRIPTION:Speakers: Hugo Lebeau (Inria ENS Lyon)\n\nThis presentation wi
 ll have two parts. The first part introduces an extension of spectral clus
 tering on data streams. Assuming observations are made in an online fashio
 n\, we show how spectral clustering can be performed on the fly with a fix
 ed amount of available memory. Studying this problem amounts to the spectr
 al analysis of a particular Gram matrix in the high-dimensional regime\, w
 hich we perform using tools from Random Matrix Theory. Based on our result
 s\, we describe the optimal memory policy and the corresponding clustering
  performance.The second part of the presentation tackles the estimation of
  a planted low-rank signal in a large-dimensional tensor. This problem rev
 eals a statistical-to-computational gap: a regime in which the maximum lik
 elihood estimator is efficient\, yet no polynomial-time algorithm can comp
 ute it. We study the performance of a procedure based on unfoldings\, whic
 h is known to achieve the best algorithmic threshold\, thereby revealing i
 nsights into the computational barrier.\n\nhttps://indico.math.cnrs.fr/eve
 nt/14480/
LOCATION:Salle K. Johnson (1R3\, 1er étage)
URL:https://indico.math.cnrs.fr/event/14480/
END:VEVENT
END:VCALENDAR