Gurprit Singh

Researcher working at the intersection of Monte Carlo sampling, rendering, optimization, and generative AI.

For more than a decade, I have developed mathematical models to understand the role of randomness. I am fascinated by noise: currently, by the pivotal role it plays in generative AI, and previously by how it affects the convergence of physically based light transport.

At the core of my research, I develop Monte Carlo sampling strategies for high-dimensional numerical integration. I am equally interested in Monte Carlo, Quasi-Monte Carlo, and MCMC methods for generative modeling.

The art of noise should not limit itself to an imitative reproduction
— Luigi Russolo

Research

My research focuses on Monte Carlo, Quasi-Monte Carlo, and MCMC methods for high-dimensional numerical integration, physically based light transport, inverse rendering, optimization, and generative modeling.

Publications

2026

Rao-Blackwellized Markov chain Monte Carlo Light Transport

Sascha Holl, Gurprit Singh, Hans-Peter Seidel

SIGGRAPH North America 2026

Q: How can Rao-Blackwellization substantially reduce variance and accelerate convergence in MCMC light transport?

arXiv

Diffusion Restore: Real-Time Markov Chain Monte Carlo Light Transport

Sascha Holl, Gurprit Singh, Hans-Peter Seidel

SIGGRAPH Asia 2026

Q: Can nonreversible diffusion dynamics make MCMC light transport both state of the art and real-time?

arXiv

Score-Based Generative Modeling through Anisotropic Stochastic Partial Differential Equations

Sascha Holl, Jente Vandersanden, Gurprit Singh, Hans-Peter Seidel

arXiv, 2026

Q: Can anisotropic diffusion preserve geometric structure longer and improve score-based image generation?

arXiv

2025

Jump restore light transport

Sascha Holl, Gurprit Singh, Hans-Peter Seidel

SIGGRAPH Asia 2025

project page / arXiv

Gaussian integral linear operators for precomputed graphics

Haolin Lu, Yash Belhe, Gurprit Singh, Tzu-Mao Li, Toshiya Hachisuka

SIGGRAPH Asia 2025

project page

Histogram Stratification for Spatio-Temporal Reservoir Sampling

Corentin Salaün, Martin Bálint, Laurent Belcour, Eric Heitz, Gurprit Singh, Karol Myszkowski

SIGGRAPH North America 2025

project page

Demystifying noise: Role of randomness in generative AI

Gurprit Singh, Xingchang Huang, Jente Vandersanden, Cengiz Öztireli, Niloy Mitra

SIGGRAPH North America Courses 2025 / Eurographics Tutorial 2025

project page

Edge-preserving noise for diffusion models

Jente Vandersanden, Sascha Holl, Xingchang Huang, Gurprit Singh

ICLR Workshop 2025

Q: What would be the impact of content-aware anisotropic noise on diffusion models?

project page / arXiv

Online importance sampling for stochastic gradient optimization

Corentin Salaun, Xingchang Huang, Iliyan Georgiev, Niloy Mitra, Gurprit Singh

ICPRAM 2025 — Best Student Paper Award

Q: Is there an efficient way to assign importance weights to mini-batch samples in gradient estimation?

project page / arXiv

Multiple importance sampling for stochastic gradient estimation

Corentin Salaun, Xingchang Huang, Iliyan Georgiev, Niloy Mitra, Gurprit Singh

ICPRAM 2025

Q: What if we have multiple importance strategies for gradient estimation?

project page / arXiv

2024

MCMC: Bridging Rendering, Optimization and Generative AI

Gurprit Singh, Wenzel Jakob

SIGGRAPH Asia Courses 2024

X: These notes are an effort to understand the role of MCMC sampling methods in rendering, optimization and generative AI.

project page / arXiv

Blue noise for diffusion models

Xingchang Huang, Corentin Salaun, Cristina Vasconcelos, Christian Theobalt, Cengiz Öztireli, Gurprit Singh

SIGGRAPH North America 2024

Q: How can we enhance generated samples simply from noise manipulation?

project page / arXiv

2023

Joint sampling and optimisation for inverse rendering

Martin Bálint, Karol Myszkowski, Hans-Peter Seidel, Gurprit Singh

SIGGRAPH Asia 2023

Q: How to reduce variance in gradient estimation during inverse rendering?

project page / arXiv

Perceptual error optimization for Monte Carlo animation rendering

Misa Korac*, Corentin Salaun*, Iliyan Georgiev, Pascal Grittmann, Philipp Slusallek, Karol Myszkowski, Gurprit Singh

SIGGRAPH Asia 2023 — conference track

Q: How to design perceptually motivated spatio-temporal masks for Monte Carlo animation rendering?

project page / arXiv

Joint first authors.

Patternshop: Editing point patterns with image manipulations

Xingchang Huang, Tobias Ritschel, Hans-Peter Seidel, Pooran Memari, Gurprit Singh

SIGGRAPH North America 2023

Q: How can we design a 2D color-space that allows editing point patterns with Photoshop?

project page / arXiv

2022

Informatik Spektrum: Scalable multi-class sampling via filtered sliced optimal transport

Corentin Salaun, Iliyan Georgiev, Hans-Peter Seidel, Gurprit Singh

Cover image for Informatik Spektrum, October 2022

journal

Scalable multi-class sampling via filtered sliced optimal transport

Corentin Salaun, Iliyan Georgiev, Hans-Peter Seidel, Gurprit Singh

SIGGRAPH Asia 2022 / ACM Transactions on Graphics, Volume 41, Issue 6, December 2022

Q: How can we build a unified framework for stippling, object placement and perceptually pleasing rendering?

project page / arXiv

Point-pattern synthesis using Gabor and random filters

Xingchang Huang, Pooran Memari, Hans-Peter Seidel, Gurprit Singh

EGSR 2022 / Computer Graphics Forum, Volume 41, Issue 6, July 2022

Q: How can we perform point pattern texture synthesis without training a network?

project page

Regression-based Monte Carlo integration

Corentin Salaun, Adrien Gruson, Binh-Son Hua, Toshiya Hachisuka, Gurprit Singh

SIGGRAPH North America 2022 / ACM Transactions on Graphics, Volume 41, Issue 4, July 2022

Q: What happens if we use a polynomial function to average Monte Carlo estimates?

project page / arXiv

Perceptual error optimization for Monte Carlo rendering

Vassillen Chizhov, Iliyan Georgiev, Karol Myszkowski, Gurprit Singh

ACM Transactions on Graphics, Volume 41, Issue 3, June 2022 — presented at SIGGRAPH North America 2022

Q: How can we use a perception-based human visual system model to control the error distribution in rendering?

project page / arXiv

2021

Informatik Spektrum: Neural Light Field 3D Printing

Quan Zheng, Vahid Babaei, Gordon Wetzstein, Hans-Peter Seidel, Matthias Zwicker, Gurprit Singh

Cover image for Informatik Spektrum, October 2021

magazine / journal

Neural Relightable Participating Media Rendering

Quan Zheng, Gurprit Singh, Hans-Peter Seidel

NeurIPS 2021

arXiv

Blue Noise Plots

Christian van Onzenoodt, Gurprit Singh, Timo Ropinski, Tobias Ritschel

Eurographics 2021 / Computer Graphics Forum, Volume 40, Issue 2, May 2021

arXiv / source code

2020

Neural Light Field 3D Printing

Quan Zheng, Vahid Babaei, Gordon Wetzstein, Hans-Peter Seidel, Matthias Zwicker, Gurprit Singh

SIGGRAPH Asia 2020 / ACM Transactions on Graphics, Volume 39, Issue 6, December 2020

project page

LadyBird: Quasi-Monte Carlo Sampling for Deep Implicit Field Based 3D Reconstruction with Symmetry

Yifan Xu*, Tianqi Fan*, Yi Yuan, Gurprit Singh

ECCV 2020 — Oral

project page

Contributed equally.

Real-time Monte Carlo Denoising with the Neural Bilateral Grid

Xiaoxu Meng, Quan Zheng, Amitabh Varshney, Gurprit Singh, Matthias Zwicker

Eurographics Symposium on Rendering 2020

project page

2019

Deep Point Correlation Design

Thomas Leimkühler, Gurprit Singh, Karol Myszkowski, Hans-Peter Seidel, Tobias Ritschel

SIGGRAPH Asia 2019 / ACM Transactions on Graphics, Volume 38, Issue 6, October 2019

project page

Analysis of Sample Correlations for Monte Carlo Rendering

Gurprit Singh, Cengiz Öztireli, Abdalla G. M. Ahmed, David Coeurjolly, Kartic Subr, Oliver Deussen, Victor Ostromoukhov, Ravi Ramamoorthi, Wojciech Jarosz

Computer Graphics Forum — Proceedings of Eurographics State of the Art Reports 2019

project page

Fourier Analysis of Correlated Monte Carlo Importance Sampling

Gurprit Singh, Kartic Subr, David Coeurjolly, Victor Ostromoukhov, Wojciech Jarosz

Computer Graphics Forum, Volume 38, Issue 1, 2019

project page

A Perception-driven Hybrid Decomposition for Multi-layer Accommodative Displays

Hyeonseung Yu, Mojtaba Bemana, Marek Wernikowski, Michał Chwesiuk, Okan Tarhan Tursun, Gurprit Singh, Karol Myszkowski, Radosław Mantiuk, Hans-Peter Seidel, Piotr Didyk

IEEE VR 2019

project page

2018

Spectral Measures of Distortion for Change Detection in Dynamic Graphs

Luca Castelli Aleardi, Semih Salihoglu, Gurprit Singh, Maks Ovsjanikov

Complex Networks 2018 — Oral

project page

Sampling Analysis using Correlations for Monte Carlo Rendering

Cengiz Öztireli, Gurprit Singh

SIGGRAPH Asia Courses 2018

project page

End-to-end Sampling Patterns

Thomas Leimkühler, Gurprit Singh, Karol Myszkowski, Hans-Peter Seidel, Tobias Ritschel

Technical Report

arXiv

2017

Convergence Analysis for Anisotropic Monte Carlo Sampling Spectra

Gurprit Singh, Wojciech Jarosz

SIGGRAPH 2017 / ACM Transactions on Graphics, Volume 36, Issue 4, July 2017

project page

Variance and Convergence Analysis of Monte Carlo Line and Segment Samples

Gurprit Singh, Bailey Miller, Wojciech Jarosz

Computer Graphics Forum — Proceedings of EGSR, Volume 36, Issue 4, June 2017

project page / source code

2016

Monte Carlo Convergence Analysis for Anisotropic Sampling Power Spectra

Gurprit Singh, Wojciech Jarosz

Technical Report

technical report

Fourier Analysis of Numerical Integration in Monte Carlo Rendering: Theory and Practice

Kartic Subr, Gurprit Singh, Wojciech Jarosz

SIGGRAPH Courses 2016

project page / ACM / source code

2015

Variance and Sampling Analysis for Monte Carlo Integration in the Spherical Domain

Gurprit Singh

Ph.D. Dissertation, Université Lyon 1, France, September 2015

HAL

Variance Analysis for Monte Carlo Integration

Adrien Pilleboue*, Gurprit Singh*, David Coeurjolly, Michael Kazhdan, Victor Ostromoukhov

SIGGRAPH 2015 / ACM Transactions on Graphics, Volume 34, Issue 4, 2015

project page / ACM / source code

Joint first authors.

Variance Analysis for Monte Carlo Integration: A Representation-Theoretic Perspective

Michael Kazhdan, Gurprit Singh, Adrien Pilleboue, David Coeurjolly, Victor Ostromoukhov

Technical Report

arXiv

2014

Fast Tile-Based Adaptive Sampling with User-Specified Fourier Spectra

Florent Wachtel, Adrien Pilleboue, David Coeurjolly, Katherine Breeden, Gurprit Singh, Gaël Cathelin, Fernando de Goes, Mathieu Desbrun, Victor Ostromoukhov

SIGGRAPH 2014 / ACM Transactions on Graphics, Volume 33, Issue 4, 2014

project page / ACM

Contact

The best way to reach me is by email. You can also find me on GitHub and LinkedIn.