BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Interactive and Grounded Learning - Prof. Prithviraj Ammanabrolu (
 University of California\, San Diego)
DTSTART:20250605T160000Z
DTEND:20250605T170000Z
UID:TALK231043@talks.cam.ac.uk
CONTACT:Shun Shao
DESCRIPTION:Abstract: Natural language communication has long been conside
 red a defining characteristic of human intelligence\; as such\, it acts as
  a north star in the pursuit of artificially intelligent agents. This talk
  thus focuses on the research question of how to imbue learning agents wit
 h the ability to understand and generate contextually relevant natural lan
 guage and actions in service of achieving a goal. It will focus on creatin
 g reinforcement learning agents inspired by communication strategies in hu
 mans that can: (1) interactively gather and align to human preferences via
  feedback to shape “how” a given task is executed\, and (2) learn worl
 d models that tell an agent “what” action to take in the current conte
 xt and “why” to take that action given the dynamics of a grounded envi
 ronment.\n\n\n\nBio:\n\nPrithviraj or Prithvi or Raj is an Assistant Profe
 ssor at UCSD leading the PEARLS Lab and a Research Scientist at Databricks
  (via MosaicML). He was previously a researcher at Ai2 and before that rec
 eived his PhD at Georgia Tech. You can find out more about his work at htt
 ps://prithvirajva.com\n
LOCATION:https://cam-ac-uk.zoom.us/j/97599459216?pwd=QTRsOWZCOXRTREVnbTJBd
 XVpOXFvdz09
END:VEVENT
END:VCALENDAR