BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Narrative Summarization from Multiple Views - Pinelopi Papalampidi
  (DeepMind)
DTSTART:20230217T120000Z
DTEND:20230217T130000Z
UID:TALK196516@talks.cam.ac.uk
CONTACT:Rami Aly
DESCRIPTION:Abstract: \n\nAlthough summarizing movies and TV shows comes n
 aturally to humans\, it is very challenging for machines. They have to com
 bine different input sources (i.e.\, video\, audio\, subtitles)\, process 
 long videos of 1-2 hours\, and their transcripts\, and learn from a handfu
 l of examples\, since collecting and processing such videos is hard. Given
  the challenges of multimodal summarization\, most prior work does not con
 sider all facets of the computational problem at once but instead focuses 
 on either processing multiple but short input sources or long text-only na
 rratives.\n\nIn contrast\, we aim at summarizing full-length movies and TV
  episodes while considering all input sources for creating video trailers 
 and textual summaries. For trailer creation\, we propose an algorithm for 
 selecting trailer moments in movies based on interpretable criteria such a
 s the narrative importance and sentiment intensity of events. We further d
 emonstrate how we can convert our algorithm into an interactive tool for t
 railer creation with a human in the loop. Next\, for producing textual sum
 maries from full-length TV episodes\, we move to a video-to-text setting a
 nd hypothesize that multimodal information from the full-length video and 
 audio can directly facilitate abstractive dialogue summarization. We propo
 se a parameter-efficient way for incorporating such information into a pre
 -trained textual summarizer and demonstrate improvements in the generated 
 summaries.\n\n\nBio:\n\nPinelopi (Nelly) Papalampidi is a Research Scienti
 st at DeepMind working at the intersection of language and vision. She rec
 ently completed her PhD at the University of Edinburgh under the supervisi
 on of Mirella Lapata and Frank Keller and interned as a Research Scientist
  at DeepMind and Meta AI. Her PhD thesis focuses on structure-aware movie 
 understanding and summarization via multimodal and graph-based methods.
LOCATION:Computer Lab\, SS03
END:VEVENT
END:VCALENDAR