BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Hong_Kong
X-LIC-LOCATION:Asia/Hong_Kong
BEGIN:STANDARD
TZOFFSETFROM:+0800
TZOFFSETTO:+0800
TZNAME:HKT
DTSTART:19911015T033000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20251218T030656Z
LOCATION:Meeting Room S426+S427\, Level 4
DTSTART;TZID=Asia/Hong_Kong:20251217T135300
DTEND;TZID=Asia/Hong_Kong:20251217T140400
UID:siggraphasia_SIGGRAPH Asia 2025_sess139_papers_1258@linklings.com
SUMMARY:Off-Centered WoS-Type Solvers with Statistical Weighting
DESCRIPTION:Anchang Bao (School of Software, Tsinghua University); Jie Xu 
 (University of Electronic Science and Technology of China); Enya Shen (Sch
 ool of Software, Tsinghua University; Haihe Lab of ITAI); and Jianmin Wang
  (School of Software, Tsinghua University)\n\nStochastic solvers have emer
 ged as a powerful alternative to traditional discretization-based methods 
 for solving partial differential equations (PDEs), especially in geometry 
 processing and graphics. While off-centered estimators enhance sample reus
 e in Monte Carlo solvers, they introduce correlation artifacts and bias wh
 en Green’s functions are approximated. In this paper, we propose a statist
 ically weighted off-centered Monte Carlo estimator that leverages local si
 milarity filtering to selectively combine samples across neighboring evalu
 ation points. Our method balances bias and variance through a principled w
 eighting strategy that suppresses unreliable estimators. We demonstrate ou
 r approach's effectiveness on various PDEs—including screened Poisson equa
 tions—and boundary conditions, achieving consistent improvements over exis
 ting solvers such as vanilla Walk on Spheres, mean value caching, and boun
 dary value caching. Our method also naturally extends to gradient field es
 timation and mixed boundary problems.\n\nRegistration Category: Full Acces
 s, Full Access Supporter\n\nSession Chair: Sheng Li (Peking University)\n\
 n
END:VEVENT
END:VCALENDAR