BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Tokyo
X-LIC-LOCATION:Asia/Tokyo
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:JST
DTSTART:18871231T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260817T171531Z
LOCATION:Hall B7 (1)\, B Block\, Level 7
DTSTART;TZID=Asia/Tokyo:20241204T104500
DTEND;TZID=Asia/Tokyo:20241204T105600
UID:siggraphasia_SIGGRAPH Asia 2024_sess114_papers_464@linklings.com
SUMMARY:LLM-enhanced Scene Graph Learning for Household Rearrangement
DESCRIPTION:Wenhao Li, Zhiyuan Yu, Qijin She, Zhinan Yu, Yuqing Lan, and C
 henyang Zhu (National University of Defense Technology (NUDT)); Ruizhen Hu
  (Shenzhen University (SZU)); and Kai Xu (National University of Defense T
 echnology (NUDT))\n\nThe household rearrangement task involves spotting mi
 splaced objects in a scene and accommodate  them with proper places. It de
 pends both on common-sense knowledge on the objective side and human user 
 preference on the subjective side. In achieving such task, we propose to m
 ine object functionality with user preference alignment directly from the 
 scene itself, without relying on human intervention. To do so, we work wit
 h scene graph representation and propose LLM-enhanced scene graph learning
  which transforms the input scene graph into an affordance-enhanced graph 
 (AEG) with information-enhanced nodes and newly discovered edges (relation
 s). In AEG, the nodes corresponding to the receptacle objects are augmente
 d with context-induced affordance which encodes what kind of carriable obj
 ects can be placed on it. New edges are discovered with newly discovered n
 on-local relations. With AEG, we perform task planning for scene rearrange
 ment by detecting misplaced carriables and determining a proper placement 
 for each of them. We test our method by implementing a tiding robot in sim
 ulator and perform evaluation on a new benchmark we build. Extensive evalu
 ations demonstrate that our method achieves state-of-the-art performance o
 n misplacement detection and the following rearrangement planning.\n\nRegi
 stration Category: Full Access, Full Access Supporter\n\nLanguage Format: 
 English Language\n\nSession Chair: Kai Wang (Simon Fraser University, Shan
 ghaiTech University)\n\n
URL:https://asia.siggraph.org/2024/program/?id=papers_464&sess=sess114
END:VEVENT
END:VCALENDAR