BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Towards Knowledge-Robust and Multimodally-Grounded NLP - Mohit Ban
 sal\, UNC Chapel Hill
DTSTART:20210610T160000Z
DTEND:20210610T170000Z
UID:TALK160972@talks.cam.ac.uk
CONTACT:Marinela Parovic
DESCRIPTION:In this talk\, I will present our group's work on NLP models t
 hat are knowledge-robust and multimodally-grounded. First\, we will descri
 be multi-task and reinforcement learning methods to incorporate novel auxi
 liary-skill tasks such as saliency\, entailment\, and back-translation val
 idity (including bandit-based methods for automatic auxiliary task selecti
 on+mixing and multi-reward mixing). Next\, we will discuss developing adve
 rsarial robustness against reasoning shortcuts\, missing commonsense gaps\
 , and cross-domain/lingual generalization in QA and dialogue models (inclu
 ding auto-adversary generation). Lastly\, we will discuss multimodally-gro
 unded models which condition and reason on dynamic spatio-temporal informa
 tion in images and videos\, and action-based robotic navigation and assemb
 ling tasks (including commonsense reasoning for ambiguous robotic instruct
 ions).
LOCATION:https://cam-ac-uk.zoom.us/j/97599459216?pwd=QTRsOWZCOXRTREVnbTJBd
 XVpOXFvdz09
END:VEVENT
END:VCALENDAR