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:20260805T205513Z
LOCATION:Room 403  B
DTSTART;TZID=America/Los_Angeles:20260720T104500
DTEND;TZID=America/Los_Angeles:20260720T121500
UID:siggraph_SIGGRAPH 2026_sess142@linklings.com
SUMMARY:Learning for CAD
DESCRIPTION:DualBrep: A Dual-Field Continuous Representation for B-rep Mod
 elling\n\nDualBrep presents a novel continuous representation for CAD mode
 ls, resolving modeling challenges caused by heterogeneous data structures.
  It uses dual scalar fields: a Signed Distance Function for geometry and a
 n Unsigned Distance Field for topology. Compressing these into a single la
 tent space ena...\n\n\nYilin Liu, Pradeep Jayaraman, Chinthala Reddy, Xian
 g Xu, and Hooman Shayani (Autodesk Research)\n---------------------\nAutor
 egressive B-Rep Shape Generation with Parametric Surfaces\n\nParaCAD is a 
 point-cloud-conditioned autoregressive framework for B-Rep generation that
  directly predicts native parametric surface types and continuous paramete
 rs via a surface-centric tokenization, preserving CAD semantics and downst
 ream editability beyond grid/point-based representations.\n\n\nDafei Qin a
 nd Rui Xu (University of Hong Kong); Zeyu Shen (Institute of Software, Chi
 nese Academy of Sciences; University of Chinese Academy of Sciences); Kaic
 hun Qiao, Hongyang Lin, and Qixuan Zhang (ShanghaiTech University); Huaiji
 n Pi (University of Hong Kong); Lan Xu and Jingyi Yu (ShanghaiTech Univers
 ity); Wenping Wang (Texas A&M University); and Taku Komura (University of 
 Hong Kong)\n---------------------\nB-repLer: Language-guided Editing of CA
 D Models\n\nCurrent language-guided CAD editing struggles with dataset sho
 rtages and relies on construction history. We introduce B-repLer, a novel 
 framework enabling text-based CAD edits in a learned latent space without 
 requiring construction history. We also present BrepEDIT-240K, the first l
 arge-scale editi...\n\n\nYilin Liu and Niladri Shekhar Dutt (University Co
 llege London (UCL)); Changjian Li (University of Edinburgh); and Niloy J. 
 Mitra (University College London (UCL), Adobe Research)\n-----------------
 ----\nFlatten the Complex: Joint B-Rep Generation via Compositional k-Cell
  Particles\n\nGenerating CAD B-Reps is challenging due to complex topology
 -geometry entanglement. We solve this by reformulating B-Reps into composi
 tional k-cell particles. Our flow matching framework unifies topological e
 ntities, enabling the joint generation, precise 3D reconstruction, and loc
 al in-painting of ...\n\n\nJunran Lu, Yuanqi Li, Hengji Li, Jie Guo, and Y
 anwen Guo (Nanjing University, State Key Laboratory for Novel Software Tec
 hnology)\n---------------------\nBrepForge: Factorized B-rep Synthesis via
  Wireframe Composition and Boundary-Conditioned Surface Instantiation\n\nB
 repForge is a generative framework for CAD B-rep modeling that factorizes 
 synthesis into wireframe construction and boundary-conditioned surface gen
 eration. This two-stage design improves topological validity and geometric
  precision, enabling the creation of complex and realistic CAD models.\n\n
 \nJing Li, Yihang Fu, and Falai Chen (University of Science and Technology
  of China)\n---------------------\nImg2CADSeq: Image-to-CAD Generation via
  Sequence-Based Diffusion\n\nImg2CADSeq is a multi-stage pipeline converti
 ng single-view images into parametrically editable 3D CAD models. Using a 
 coarse-to-fine hierarchical codebook and VQ-Diffusion , it bridges 2D visu
 al features with 3D semantics , generating standard STEP files ready for d
 irect commercial software use.\n\n\nShiyu Tan, Zixuan Zhao, Hao Gao, Zhihe
 ng Chen, and Xiaolong Yin (School of Software and BNRist, Tsinghua Univers
 ity) and Enya Shen (School of Software and BNRist, Tsinghua University; Ha
 ihe Lab of ITAI)\n---------------------\nLearning for CAD - Interactive Di
 scussion\n\nAfter the summary presentations, attendees will participate in
  an interactive discussion. Outside the room will be a series of poster bo
 ards for authors to gather around with the audience. Authors are invited t
 o bring any material related to their paper that could instigate further c
 onversation such...\n\n\nInterest Area: Research & Education\n\nKeyword: A
 rtificial Intelligence/Machine Learning, Geometry, Modeling\n\nRegistratio
 n Category: Full Conference Supporter, Full Conference\n\nSession Chair: C
 hangjian Li (University of Edinburgh)
END:VEVENT
END:VCALENDAR
