Pinned Tweet
A new blog post! bartwronski.com/2020/03/15/usi… Using JAX, numpy, and optimization techniques to improve the separable image filters - a follow-up to my previous post with some new hopefully useful programming and mathematical tools for graphics and image processing problems.
6
60
272
Bart Wronski 🇺🇦🇵🇸 retweeted
We are looking forward to seeing you at the Light Transport session at #SIGGRAPH 2026 on Tuesday!
Check out our work with my mentor @marcosalvi, @tzumaoli , @BartWronsk @winmad4869 our labmates @sneezygiraffe, Cheng and Ravi! : A generalizable light transport 3D embedding for ere we learn light transport in a generalizable, scalable manner. lnkd.in/gsBpmCmy ⚽
#siggraph2026
Check out our work with my mentor @marcosalvi, @tzumaoli , @BartWronsk @winmad4869 our labmates @sneezygiraffe, Cheng and Ravi! : A generalizable light transport 3D embedding for ere we learn light transport in a generalizable, scalable manner. lnkd.in/gsBpmCmy ⚽
#siggraph2026
0
3
12
1,923
Bart Wronski 🇺🇦🇵🇸 retweeted
What a journey this was! 🚀
Couldn't be prouder of my intern @_bingxu for leading this work. It took immense persistence to turn the initial hunch, that an attention mechanism could model long-range interactions in light transport, into a robust, working system. Huge congrats to the entire team! 🎉
You can read the paper here:
arxiv.org/abs/2510.18189…
Couldn't be prouder of my intern @_bingxu for leading this work. It took immense persistence to turn the initial hunch, that an attention mechanism could model long-range interactions in light transport, into a robust, working system. Huge congrats to the entire team! 🎉
You can read the paper here:
arxiv.org/abs/2510.18189…
🌲Introducing our paper A generalizable light transport 3D embedding for Global Illumination arxiv.org/pdf/2510.18189.
Just as Transformers learn long-range relationships between words or pixels, our new paper shows they can also learn how light interacts and bounces around a 3D analogy of light transport operator and Attention, our fully scalable model learns a generalizable light transport 3D embedding that captures global illumination.
It takes as input the 3D scene assets—geometry, materials, and lighting represented as a point cloud—and encodes the complex light interaction into a 3D latent space, with codes anchored at the sampled scene points.
Remarkably, the model is independent of both viewpoint and resolution, enabling view-consistent rendering across diverse scenes. (See teaser and attention figure.)
(Our team work: with my mentor Marco @marcosalvi, labmates Mukund @mukundvermar, Cheng, NV folks Bart @BartWronsk, Lifan @winmad4869, and Tzu-Mao @tzumaoli & Ravi.)
🧵
2
6
45
7,696
Bart Wronski 🇺🇦🇵🇸 retweeted
🌲Introducing our paper A generalizable light transport 3D embedding for Global Illumination arxiv.org/pdf/2510.18189.
Just as Transformers learn long-range relationships between words or pixels, our new paper shows they can also learn how light interacts and bounces around a 3D analogy of light transport operator and Attention, our fully scalable model learns a generalizable light transport 3D embedding that captures global illumination.
It takes as input the 3D scene assets—geometry, materials, and lighting represented as a point cloud—and encodes the complex light interaction into a 3D latent space, with codes anchored at the sampled scene points.
Remarkably, the model is independent of both viewpoint and resolution, enabling view-consistent rendering across diverse scenes. (See teaser and attention figure.)
(Our team work: with my mentor Marco @marcosalvi, labmates Mukund @mukundvermar, Cheng, NV folks Bart @BartWronsk, Lifan @winmad4869, and Tzu-Mao @tzumaoli & Ravi.)
🧵
Just as Transformers learn long-range relationships between words or pixels, our new paper shows they can also learn how light interacts and bounces around a 3D analogy of light transport operator and Attention, our fully scalable model learns a generalizable light transport 3D embedding that captures global illumination.
It takes as input the 3D scene assets—geometry, materials, and lighting represented as a point cloud—and encodes the complex light interaction into a 3D latent space, with codes anchored at the sampled scene points.
Remarkably, the model is independent of both viewpoint and resolution, enabling view-consistent rendering across diverse scenes. (See teaser and attention figure.)
(Our team work: with my mentor Marco @marcosalvi, labmates Mukund @mukundvermar, Cheng, NV folks Bart @BartWronsk, Lifan @winmad4869, and Tzu-Mao @tzumaoli & Ravi.)
🧵
4
33
128
20,116
Bart Wronski 🇺🇦🇵🇸 retweeted
Our upcoming HPG 2025 paper ”Collaborative Texture Filtering” is up!
Research done with
@_pandersson, Matt Pharr, and @BartWronsk.
research.nvidia.com/labs/rtr/publi…
Research done with
@_pandersson, Matt Pharr, and @BartWronsk.
research.nvidia.com/labs/rtr/publi…
1
30
81
8,661
I'm excited to share our latest paper (work with Matt Pharr and @inversepixel ), "Improved Stochastic Texture Filtering Through Sample Reuse" accepted to I3D'2025.
arxiv.org/abs/2504.05562
research.nvidia.com/labs/rtr/publi… 1/
arxiv.org/abs/2504.05562
research.nvidia.com/labs/rtr/publi… 1/
4
51
209
14,005
3. Finally, we put those contributions and the idea of work sharing together and show it can allow for high quality higher order filtering, such as a B-spline bicubic filter, still at 1spp and with a negligible performance overhead. 6/
1
0
7
1,663
Check out our paper arxiv.org/abs/2504.05562 and the video on the project website! research.nvidia.com/labs/rtr/publi…
I'll be presenting this work on behalf of the team at I3D'25 (together with JCGT Laplacian blending paper) - if you're there, say "hi"! :) 7/7
I'll be presenting this work on behalf of the team at I3D'25 (together with JCGT Laplacian blending paper) - if you're there, say "hi"! :) 7/7
0
0
9
1,542
I'm thrilled to share a JCGT paper, "GPU Friendly Laplacian Texture Blending." It uses simple image processing and computational photography techniques to smoothly blend textures and materials while preserving local contrast and texture details without ghosting or blurring. 1/3
GPU Friendly Laplacian Texture Blending
jcgt.org/published/0014…
8
42
217
12,187
This family of techniques can aid the material authoring process (artist-guided or procedural) and does not require any pre-processing other than the presence of regular texture mip chains. The additional cost is simply sampling additional coarser mip levels. 2/3
1
0
11
1,647
You can find a live WebGL demo here: jcgt.org/published/0014…
Applications can go beyond textures and images and could be useful on other signals or even in ML applications, such as inpainting or regularization.
Hope you find it inspiring or useful! 3/3
Applications can go beyond textures and images and could be useful on other signals or even in ML applications, such as inpainting or regularization.
Hope you find it inspiring or useful! 3/3
0
0
14
1,498
Bart Wronski 🇺🇦🇵🇸 retweeted
it's time again !
A majority of Poles, 53%, say life has got harder in 2024 under Tusk's government than it was in 2023 under the former PiS government. Only 33% say it has got easier.
Even among the ruling coalition's voters, 29% say things are worse, finds the poll wiadomosci.wp.pl/nowy-sondaz-ja…
117
2,066
19,515
828,009
Bart Wronski 🇺🇦🇵🇸 retweeted
It's important to distinguish between supporting legal immigration generally and advocating specifically for tech immigration to fill high-skilled jobs. While both positions have merit, they're driven by different motivations and have distinct implications.
Often, those in the ss concerns about unfilled positions and visa challenges primarily focus on the need for skilled workers. This perspective, while understandable, overshadows the broader context of immigration and the diverse needs and contributions of immigrants from all walks of life.
We need to remember that contributions to American society come in many forms, and a narrow focus on high-skilled immigration can obscure the immense potential of individuals who may arrive with very limited resources and experience.
Examples abound, but consider the remarkable cases of just two refugees who had immense impact on tech: Sergey Brin fled the Soviet Union with his family as a child; and Andy Grove, who escaped Hungary and revolutionized tech with the microprocessor. Neither were “skilled workers” at the time of admission to the US.
A truly balanced immigration system should consider not only the demands of the tech sector but also the value of immediate family reunification, humanitarian concerns, and the overall enrichment that diverse forms of immigration bring to our society. To fulfill the long-term promise of America as a nation of opportunity, we must recognize and embrace the potential of all legal immigrants, not just those who would serve the short term needs of the tech sector.
Often, those in the ss concerns about unfilled positions and visa challenges primarily focus on the need for skilled workers. This perspective, while understandable, overshadows the broader context of immigration and the diverse needs and contributions of immigrants from all walks of life.
We need to remember that contributions to American society come in many forms, and a narrow focus on high-skilled immigration can obscure the immense potential of individuals who may arrive with very limited resources and experience.
Examples abound, but consider the remarkable cases of just two refugees who had immense impact on tech: Sergey Brin fled the Soviet Union with his family as a child; and Andy Grove, who escaped Hungary and revolutionized tech with the microprocessor. Neither were “skilled workers” at the time of admission to the US.
A truly balanced immigration system should consider not only the demands of the tech sector but also the value of immediate family reunification, humanitarian concerns, and the overall enrichment that diverse forms of immigration bring to our society. To fulfill the long-term promise of America as a nation of opportunity, we must recognize and embrace the potential of all legal immigrants, not just those who would serve the short term needs of the tech sector.
3
7
106
12,547
Bart Wronski 🇺🇦🇵🇸 retweeted
No, the people singing “there are no schools in Gaza because we burned all the children” are not the fucking victims here, and no matter how many paid-for politicians send tweets pretending otherwise they never will be
113
5,480
21,181
285,741
Bart Wronski 🇺🇦🇵🇸 retweeted
This is the genocidal trash @SenSchumer is painting as innocent pogrom victims
It bugged me that I couldn't find a video of the entire genocidal song, but then came Maccabi Tel Aviv fans: they arrived at Ben Gurion airport, fleeing from the fake pogrom, and started singing the entire song...
Ole ole
Ole ole ole
Let the IDF win and fuck the Arabs
Ole ole
ol out in Gaza? There are no children left there
70
907
2,643
64,050
