BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260805T205504Z
LOCATION:Room 403 A
DTSTART;TZID=America/Los_Angeles:20260721T090000
DTEND;TZID=America/Los_Angeles:20260721T103000
UID:siggraph_SIGGRAPH 2026_sess106@linklings.com
SUMMARY:Layouts &  Procedural
DESCRIPTION:LLM-enhanced Scene Graph Learning for Household Rearrangement\
 n\nWe present an LLM-enhanced scene graph framework for autonomous househo
 ld rearrangement. By transforming scene graphs into affordance-enhanced gr
 aphs that encode scene-grounded object functionality and user preferences,
  our method detects misplaced objects and plans correct placements. We dem
 onstrat...\n\n\nWenhao Li, Shilong Zou, Zhinan Yu, Zheng Zhou, Wenxuan Li,
  and Chenyang Zhu (National University of Defense Technology); Ruizhen Hu 
 (Shenzhen University); Kai Xu (Institute of AI for Industries Chinese Acad
 emy of Science); and Wenhao Li\n---------------------\nCasLayout: Cascaded
  3D Layout Diffusion for Indoor Scene Synthesis with Implicit Relation Mod
 eling\n\nCasLayout is a cascaded diffusion framework that decomposes 3D in
 door scene generation into four interpretable stages. This design reduces 
 data requirements while enabling flexible, controllable synthesis and seam
 less integration with LLMs and VLMs for zero-shot tasks.\n\n\nYingrui Wu (
 Institute of Automation, Chinese Academy of Sciences; School of Artificial
  Intelligence, University of Chinese Academy of Sciences); You-Kang Kong (
 Tsinghua University, Microsoft Research Asia); Mingyang Zhao (State Key La
 boratory of Mathematical Sciences, Academy of Mathematics and Systems Scie
 nce, Chinese Academy of Sciences); Weize Quan and Dongmin Yan (Institute o
 f Automation, Chinese Academy of Sciences); and Yang Liu (Microsoft Resear
 ch Asia)\n---------------------\nRaster2Seq: Polygon Sequence Generation f
 or Floorplan Reconstruction\n\nRaster2Seq is an anchor-based autoregressiv
 e formulation that frames floorplan reconstruction as scalable, sequential
  polygon prediction, transforming rasterized floorplan images into vectori
 zed format. Through using our proposed labeled polygon sequence representa
 tion, the method jointly encodes sp...\n\n\nHao Phung and Hadar Averbuch-E
 lor (Cornell University)\n---------------------\nToward Richer Material Ge
 neration via Procedural Data Enhancement\n\nGenerative material models are
  limited by simple PBR data. We augment single-lobe GGX materials into lay
 ered, multi-lobe models capturing richer effects (e.g., dust, clearcoat). 
 These are encoded as neural materials in a shared 6D latent space. The res
 ulting dataset enables generative models to prod...\n\n\nYunchen Yu (Corne
 ll University, NVIDIA); Jacob Munkberg, Jon Hasselgren, and Chris Cummings
  (NVIDIA); Steve Marschner (Cornell University, NVIDIA); and Andrea Weidli
 ch (NVIDIA)\n---------------------\nInfiniteDiffusion: Bridging Learned Fi
 delity and Procedural Utility for Open-World Terrain Generation\n\nWe intr
 oduce InfiniteDiffusion, a training-free algorithm enabling diffusion mode
 ls to generate spatially infinite outputs with the core properties of proc
 edural noise: seed-consistency (determinism) and constant-time random acce
 ss. We incorporate a novel architecture for terrain generation to produ...
 \n\n\nAlexander Goslin (Independent)\n---------------------\nRCGP: Resourc
 e Contracts for Graphics Programming\n\nRCGP introduces a type system for 
 graphics programming that statically enforces resource contracts, the agre
 ements on type, layout, binding, and synchronization between shaders, pipe
 lines, and host code. Through contracts, modules, combinators, and witness
 es, RCGP catches descriptor mismatches, layo...\n\n\nVenkataram Sivaram (C
 omputer Science and Artificial Intelligence Laboratory (CSAIL), Massachuse
 tts Institute of Technology (MIT)); Sai Praveen Bangaru (NVIDIA); Ravi Ram
 amoorthi and Tzu-Mao Li (University of California San Diego); and Jonathan
  Ragan-Kelley and Fredo Durand (Computer Science and Artificial Intelligen
 ce Laboratory (CSAIL), Massachusetts Institute of Technology (MIT))\n-----
 ----------------\nLayouts & Procedural - Interactive Discussion\n\nAfter t
 he summary presentations, attendees will participate in an interactive dis
 cussion. Outside the room will be a series of poster boards for authors to
  gather around with the audience. Authors are invited to bring any materia
 l related to their paper that could instigate further conversation such...
 \n\n\nInterest Area: Research & Education\n\nKeyword: Artificial Intellige
 nce/Machine Learning, Generative AI, Geometry, Modeling, Rendering\n\nRegi
 stration Category: Full Conference Supporter, Full Conference\n\nSession C
 hair: I-Chao Shen (The University of Tokyo)
END:VEVENT
END:VCALENDAR