BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Australia/Melbourne
X-LIC-LOCATION:Australia/Melbourne
BEGIN:DAYLIGHT
TZOFFSETFROM:+1000
TZOFFSETTO:+1100
TZNAME:AEDT
DTSTART:19721003T020000
RRULE:FREQ=YEARLY;BYMONTH=4;BYDAY=1SU
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:19721003T020000
TZOFFSETFROM:+1100
TZOFFSETTO:+1000
TZNAME:AEST
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260114T163751Z
LOCATION:Meeting Room C4.9+C4.10\, Level 4 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231213T152000
DTEND;TZID=Australia/Melbourne:20231213T162500
UID:siggraphasia_SIGGRAPH Asia 2023_sess168@linklings.com
SUMMARY:Applications & Innovations
DESCRIPTION:Extended Path Space Manifolds for Physically Based Differentia
 ble Rendering\n\nPhysically based differentiable rendering has become an i
 ncreasingly important topic in recent years. A common pipeline computes lo
 cal color derivatives of light paths or pixels with respect to arbitrary s
 cene parameters, and enables optimizing or recovering the scene parameters
  through iterative gr...\n\n\nJiankai Xing and Xuejun Hu (Tsinghua Univers
 ity), Fujun Luan (Adobe Research), Ling-Qi Yan (University of California S
 anta Barbara), and Kun Xu (Tsinghua University)\n---------------------\nWa
 rped-Area Reparameterization of Differential Path Integrals\n\nPhysics-bas
 ed differentiable rendering is becoming increasingly crucial for tasks in 
 inverse rendering and machine learning pipelines. To address discontinuiti
 es caused by geometric boundaries and occlusion, two classes of methods ha
 ve been proposed: 1) the edge sampling methods that directly sample...\n\n
 \nPeiyu Xu (University of California Irvine), Sai Bangaru (MIT CSAIL), Tzu
 -Mao Li (University of California San Diego), and Shuang Zhao (University 
 of California Irvine)\n---------------------\nProjective Sampling for Diff
 erentiable Rendering of Geometry\n\nDiscontinuous visibility changes at ob
 ject boundaries remain a persistent source of difficulty in the area of di
 fferentiable rendering. Left untreated, they bias computed gradients so se
 verely that even basic optimization tasks fail.\n\nPrior path-space method
 s addressed this bias by decoupling bounda...\n\n\nZiyi Zhang, Nicolas Rou
 ssel, and Wenzel Jakob (Ecole Polytechnique Fédérale de Lausanne)\n-------
 --------------\nAmortizing Samples in Physics-Based Inverse Rendering usin
 g ReSTIR\n\nRecently, great progress has been made in physics-based differ
 entiable rendering. Existing differentiable rendering techniques typically
  focus on static scenes, but during inverse rendering—a key application fo
 r differentiable rendering—the scene is updated dynamically by each gradie
 nt s...\n\n\nYu-Chen Wang (University of California Irvine), Chris Wyman a
 nd Lifan Wu (NVIDIA), and Shuang Zhao (University of California Irvine)\n-
 --------------------\nJoint Sampling and Optimisation for Inverse Renderin
 g\n\nWhen dealing with difficult inverse problems such as inverse renderin
 g, using Monte Carlo estimated gradients to optimise parameters can slow d
 own convergence due to variance. Averaging many gradient samples in each i
 teration reduces this variance trivially. However, for problems that requi
 re thousa...\n\n\nMartin Balint, Karol Myszkowski, Hans-Peter Seidel, and 
 Gurprit Singh (Max Planck Institute for Informatics)\n\nRegistration Categ
 ory: Full Access\n\nSession Chair: Soo-Mi Choi (Sejong University)
END:VEVENT
END:VCALENDAR