BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:IC Colloquium: Unsupervised Discovery of Interpretable Structure i
 n Complex Systems
DTSTART:20250303T101500
DTEND:20250303T111500
DTSTAMP:20260413T103444Z
UID:a87315e081c691beb1167284aa9132b1a9fced7ec416fdfbb2bac451
CATEGORIES:Conferences - Seminars
DESCRIPTION:By : Mark Hamilton - MIT/Microsoft\nIC Faculty candidate\n\nAb
 stract\nHow does the human mind make sense of raw information without bein
 g taught how to see or hear? In this talk we will explore how to build alg
 orithms that can uncover interpretable structure from large collections of
  unsupervised data like images and video. First\, I will describe how to c
 lassify every pixel of a collection of images without any human annotation
 s (Unsupervised semantic segmentation) by distilling self-supervised visio
 n models. Second\, we’ll see how this basic idea leads us to a new unify
 ing theory of representation learning\, and I will show how 20 different c
 ommon machine learning methods such as dimensionality reduction\, clusteri
 ng\, contrastive learning\, and spectral methods emerge from a single unif
 ied equation. Finally\, we’ll use this unified theory to create algorith
 ms that can decode natural language just by watching unlabeled videos of p
 eople talking\, without any knowledge of text. This work is the first step
  in our broader effort to translate animals using large scale\, unsupervis
 ed\, and interpretable learners\, and the talk will conclude with some of 
 our most recent efforts to analyze the complex vocalizations of Atlantic s
 potted dolphins.\n\nBio\nMark Hamilton is a PhD student in William T Freem
 an's lab at the MIT Computer Science & Artificial Intelligence Laboratory.
  He is also a Senior Engineering Manager at Microsoft where he leads a tea
 m building a large-scale distributed ML products for Microsoft’s largest
  databases.  Mark is interested in how we can use unsupervised machine le
 arning to discover scientific "structure" in complex systems. Mark values 
 working on projects for social\, cultural\, and environmental good and aim
 s to use his algorithms to help humans solve challenges they cannot solve 
 alone.\n\nMore information\n 
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420 https://epfl.zoom.us/
 j/68600056278
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR