BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Model Merging — A Tale of Two Settings - Donato Crisostomi
DTSTART:20251121T110000Z
DTEND:20251121T120000Z
UID:TALK239923@talks.cam.ac.uk
CONTACT:Lucas Resck
DESCRIPTION:Abstract: In this talk\, I will introduce the emerging field o
 f model merging: the process of combining multiple neural networks into a 
 single model without retraining. We'll begin with foundational concepts su
 ch as linear mode connectivity and task vectors\, and explore two main set
 tings: (1) merging models trained from scratch on the same task but with d
 ifferent initializations\, and (2) merging models finetuned on different t
 asks from a shared pretrained base. I will then present a series of recent
  works that expand the model merging toolkit. These include the use of cyc
 le consistency in permutation-based merging\, insights into how task vecto
 rs relate to gradients\, SVD-based approaches for low-rank model fusion\, 
 and the application of evolutionary algorithms to discover optimal merging
  coefficients. Throughout\, we'll see how these techniques can be applied 
 in real-world scenarios\, from model compression in computer vision to the
  synthesis of state-of-the-art LLMs for low-resource languages.\n\nBio: Do
 nato Crisostomi is an ELLIS PhD student at the Sapienza University of Rome
  & University of Cambridge\, currently interning at Cohere. His research f
 ocuses on model merging and representational alignment. He currently leads
  the “Model Reuse” work package for the 1.5M€ project “NEXUS”. H
 e previously held roles as a visiting researcher at the University of Camb
 ridge\, a Research Scientist at Amazon Alexa\, and an Applied Scientist at
  Amazon Search. His research has been featured in top-tier AI conferences 
 and journals\, including CVPR\, NeurIPS\, ACM\, ACL\, and LoG. In addition
  to his scientific contributions\, he has played an active role in the res
 earch community as the organizer of the UniReps workshop at NeurIPS\, ment
 or at LOGML\, and as a program committee member for leading conferences su
 ch as CVPR\, NeurIPS\, ICLR\, etc.
LOCATION:GR06/07\, English Faculty Building\, 9 West Road\, Sidgwick Site 
 and online https://cam-ac-uk.zoom.us/j/97599459216?pwd=QTRsOWZCOXRTREVnbTJ
 BdXVpOXFvdz09
END:VEVENT
END:VCALENDAR