BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:Europe/Zagreb
BEGIN:DAYLIGHT
DTSTART:20250330T030000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20251026T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20251002T085600Z
UID:A8D2EEB0-BD6B-49B5-9C33-F61EB0FB5D4A
DTSTART;TZID=Europe/Zagreb:20250929T140000
DTEND;TZID=Europe/Zagreb:20250929T150000
DESCRIPTION:With the rise of online social media platforms\, written conver
 sations among multiple participants have become an essential part of daily
  communication\, enabling users to exchange ideas\, share opinions\, and e
 ngage in discussions. Beyond the textual content of individual messages\, 
 the dynamics of participant interactions influence the meaning and the pra
 gmatic aspects of the conversations. In this talk\, we will explore the ch
 allenges faced by models in processing Multi-Party Conversations (MPCs)\, 
 the key downstream tasks in this domain\, and the importance of incorporat
 ing interaction dynamics. Specifically\, we will discuss the impact of tra
 ining data size\, highlight critical evaluation challenges\, and propose a
  potential solution to address these issues based on the generation of syn
 thetic MPCs. We will mainly focus on the latter\, exploring different mode
 ls and strategies for generating MPCs under precise constraints. We also i
 ntroduce an evaluation framework to assess both constraint compliance and 
 generation quality\, in order to identify which model-strategy combination
 s are more effective.\n\nSpeaker(s): Nicolò \, \n\nUnska 3\, Zagreb\, Gra
 d Zagreb\, Croatia\, 10000
LOCATION:Unska 3\, Zagreb\, Grad Zagreb\, Croatia\, 10000
ORGANIZER:ana.milas@fer.hr
SEQUENCE:6
SUMMARY:Interacting in Natural Language in the era of LLMs: NLP for Online 
 Discussions and Multi-Party Conversations
URL;VALUE=URI:https://events.vtools.ieee.org/m/501500
X-ALT-DESC:Description: <br /><p><span data-olk-copy-source="MessageBody">W
 ith the rise of online social media platforms\, written conversations amon
 g multiple participants have become an essential part of daily communicati
 on\, enabling users to exchange ideas\, share opinions\, and engage in dis
 cussions. Beyond the textual content of individual messages\, the dynamics
  of participant interactions influence the meaning and the pragmatic aspec
 ts of the conversations. In this talk\, we will explore the challenges fac
 ed by models in processing Multi-Party Conversations (MPCs)\, the key down
 stream tasks in this domain\, and the importance of incorporating interact
 ion dynamics. Specifically\, we will discuss the impact of training data s
 ize\, highlight critical evaluation challenges\, and propose a potential s
 olution to address these issues based on the generation of synthetic MPCs.
  We will mainly focus on the latter\, exploring different models and strat
 egies for generating MPCs under precise constraints. We also introduce an 
 evaluation framework to assess both constraint compliance and generation q
 uality\, in order to identify which model-strategy combinations are more e
 ffective.</span></p>
END:VEVENT
END:VCALENDAR