BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:"Machine learning in chemistry and beyond" (ChE-650) seminar by Bh
 arath Ramsundar: "Language based Pre-training for Drug Discovery"
DTSTART:20211207T170000
DTEND:20211207T180000
DTSTAMP:20260510T164839Z
UID:f3c3a01619bdc471828795c1bed0a4c848b0d7a912f45c99f691de96
CATEGORIES:Conferences - Seminars
DESCRIPTION:Bharath received a BA and BS from UC Berkeley in EECS and Math
 ematics and was valedictorian of his graduating class in mathematics. He d
 id his PhD in computer science at Stanford University where he studied the
  application of deep-learning to problems in drug-discovery. At Stanford\,
  Bharath created the deepchem.io open-source project to grow the deep dr
 ug discovery open source community\, co-created the moleculenet.ai bench
 mark suite to facilitate development of molecular algorithms\, and more. 
 Bharath’s graduate education was supported by a Hertz Fellowship\, the 
 most selective graduate fellowship in the sciences. After his PhD\, Bharat
 h co-founded Computable a startup that built better tools for collaborat
 ive dataset management. Bharath is currently working actively on growing t
 he DeepChem community and on exploring a few early projects still in steal
 th. Bharath is the lead author of “TensorFlow for Deep Learning: From L
 inear Regression to Reinforcement Learning”\, a developer’s introducti
 on to modern machine learning\, with O’Reilly Media\, and the lead auth
 or of “Deep Learning for the Life Sciences”\nLanguage based Pre-train
 ing for Drug Discovery\n\nPretraining has taken the NLP world by storm as 
 ever larger language models have broken successive benchmarks. In this ta
 lk\, I'll review some recent work applying pretraining to scientific chall
 enges\, and in particular\, will discuss the challenges of pretraining for
  molecular machine learning. I'll introduce our new architecture\, Che
 mBERTa\, which explores the use of BERT-style pretraining for machine le
 arning problems inspired by drug discovery applications.
LOCATION:https://epfl.zoom.us/j/64473017589?pwd=Vmpnd1pleGhEb1hFb3kxUlNIUW
 JyQT09
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR