iiinova — where your digital self keeps evolving

Reposted by Mike Dereviannykh

Valentin Deschaintre @vdeschaintre.bsky.social · 13/08/2025

I will be presenting in the 10.30 session in room 208-209. I will share thoughts about implicit/generative and explicit appearance representations. See you there if that sounds interesting!

021

Mike Dereviannykh @mishok43.bsky.social · 13/08/2025

In just less than an hour we'll present "Neural Two-Level Monte Carlo Real-Time Rendering" on SIGGRAPH 2025 at "Best of Eurographics" session Please join us, it'll be fun! 😊 🏠Room 208-209 🕜10:30-11:30

030

Reposted by Mike Dereviannykh

Christoph Peters @momentsingraphics.bsky.social · 03/08/2025

Physically-based differentiable rendering enables inverse rendering, but handling visibility is hard. Our SIGGRAPH 2025 paper uses quadrics to importance sample silhouette edges--outperforming all existing unidirectional differentiable path tracers. momentsingraphics.de/Siggraph2025...

1417

Mike Dereviannykh @mishok43.bsky.social · 28/07/2025

Done, thank you so much for the info!

010

Mike Dereviannykh @mishok43.bsky.social · 28/07/2025

And it's a fantastic coincidence that I'm already next to Vancouver and don't need to suffer from jet lag the 2nd time... a horrible thing

010

Mike Dereviannykh @mishok43.bsky.social · 28/07/2025

Session's time and location for your schedule: s2025.conference-schedule.org/presentation...

s2025.conference-schedule.org

Presentation - SIGGRAPH 2025 Conference Schedule

000

Mike Dereviannykh @mishok43.bsky.social · 28/07/2025

Big news! The Eurographics Association invited us to present our work "Neural Two-Level Monte Carlo Real-Time Rendering" at #SIGGRAPH2025 🎉 Super honored - my first SIGGRAPH! Let’s discuss neural & real-time rendering, grab a coffee, or just hang out - feel free to leave a DM

280

Mike Dereviannykh @mishok43.bsky.social · 26/07/2025

Just joined RealityLabs at Meta to do some real-time neural rendering research. Unfortunately without a 9- digits compensation package, but WIP 😁 I'm in the Redmond office, but already visited Seattle. If you wanna grab ☕ - DMs are opened

020

Mike Dereviannykh @mishok43.bsky.social · 25/06/2025

Currently, the issue with efficient spatial encoding is more or less resolved (iNGP or the new GATE by @boksajak.bsky.social ) But it's not the case for the directional domain at all. Despite its importance for the unbiased rendering, Cache-Based Resampling and 2-Level Monte-Carlo Estimator

020

Mike Dereviannykh @mishok43.bsky.social · 25/06/2025

To be completely honest, it's a equal-time-comparison: 3SPP vs 1SPP+25 Neural Resamples I believe we should invest more resources in more rapid adaptivity of neural caches and more aggressive quantizations, so we could deliver it to production real-time rendering

120

Mike Dereviannykh @mishok43.bsky.social · 25/06/2025

It's a cool work which deserve a lot attentions from real-time rendering community! Previously I've got a little bit of time to conduct similar experiments on top of our NIRC, as each additional neural sample costs just pure tensor FLOPs ~ 1.5ms on 4080 1 spp vs 1 spp + 25 cache resamples

120

Mike Dereviannykh @mishok43.bsky.social · 25/06/2025

I found it cool to hear the motivation for High-Frequency Learnable Encoding for NRC\NIRC\Neural Ambient Occlusion from the perspective of Kernel Machines! Classics 😊

051

Mike Dereviannykh @mishok43.bsky.social · 06/06/2025

Hmmm, 1 min sounds amazing, I didn't expect such speed!

000

Mike Dereviannykh @mishok43.bsky.social · 04/06/2025

Congrats to you and to the whole team! I've got a question: basically the volumetric blocks don't have any transparency, do they? And you rely on pruning to make them look right because of it?

100

Reposted by Mike Dereviannykh

Valentin Deschaintre @vdeschaintre.bsky.social · 13/05/2025

I was very honoured to receive one of the two Eurographics Young Researcher Award 2025 yesterday! This is the combination of the work of many people, mentors, collaborators, students, friends who trusted me and taught me so much along the way!

9374

Mike Dereviannykh @mishok43.bsky.social · 17/05/2025

Huge congrats, Valentin! 100% Deserved

110

Mike Dereviannykh @mishok43.bsky.social · 16/05/2025

We've received an "Honorable Mention" at the Eurographics 2025 for our work on "Neural Two-Level Monte Carlo Real-Time Rendering" in London! 🥳 Huge thanks to everyone who supported me along the way, and to the EG chairs, committee, and organizers for this recognition

140

Mike Dereviannykh @mishok43.bsky.social · 15/05/2025

That's what I mean by the lack of compute

000

Mike Dereviannykh @mishok43.bsky.social · 15/05/2025

If you're interested in something like AlphaEvolve but focused on CG, GameDev, or offline rendering, feel free to reach out. I’ve been leading research in this space with strong results so far. But we need support with compute, expertise and even maybe with engineering

110

Mike Dereviannykh @mishok43.bsky.social · 11/05/2025

And of course, huge thanks to my amazing co-authors — Dmitrii Klepikov, Johannes Hanika, Carsten Dachsbacher — and industry friends @kaplanyan.bsky.social , Sebastian Herholz, @momentsingraphics.bsky.social, who helped me on the way!

010

Mike Dereviannykh @mishok43.bsky.social · 11/05/2025

But there are so many other cool questions that we tried to cover So please check out our webpage, demo videos, and paper itself: mishok43.github.io/nirc/ 📍I’ll be at Eurographics next week in London — if you're around and want to talk rendering, AI, or just grab coffee, DM me!

mishok43.github.io

Neural Two-Level Monte Carlo Real-Time Rendering

Efficient real-time global illumination rendering using Neural Incident Radiance Cache combined with Two-Level Monte Carlo. By Mikhail Dereviannykh, Dmitrii Klepikov, Johannes Hanika, and Carsten Dach...

110

Mike Dereviannykh @mishok43.bsky.social · 11/05/2025

In equal-time comparisons, NIRC achieves surprisingly cool results both in the biased and unbiased cases But yeah... variance may increase next to foliage, brush, trees🌿 — still the eternal pain in CG 😅

100

Mike Dereviannykh @mishok43.bsky.social · 11/05/2025

Using Two-Level Monte Carlo, we can debias NIRC while still cutting variance — thanks to fast cache sampling - dozens of times for the cost of 1 real path It works like (N)CV, but doesn't introduce any architectural constraints! No need to train Normalizing Flows on-the-fly

110

Mike Dereviannykh @mishok43.bsky.social · 11/05/2025

Another positive scalability property: deeper MLPs do improve quality here Downside: not all scenes benefit from it (esp. with high-variance MC estimator. must be further researched)

100

Mike Dereviannykh @mishok43.bsky.social · 11/05/2025

And we got basically classical Monte-Carlo integration, but over the neural domain! But it scales pretty well with the number of neural samples!

100

Mike Dereviannykh @mishok43.bsky.social · 11/05/2025

NIRC amortizes iNGP costs via task-reformulation: from outgoing to incident radiance 1. Use hash-grid on surface point → get latent light rep 2. Sample incoming dirs via BSDF 3. Decode radiance using MLPs (per-dir) The more directions, the more we leverage GPU tensor FLOPS 💥

100

Mike Dereviannykh @mishok43.bsky.social · 11/05/2025

Inspired by NRC + iNGP’s adaptivity from amazing Tomas Müller, Christoph Schied, Jan Novák, Alex Evans and et al, but found key limits: – Up to 70% time spent on iNGP → memory-bound – MLP depth ≠ sign. better quality → poor FLOPs scaling – Biased for specular & detailed BSDFs with normals

110

Mike Dereviannykh @mishok43.bsky.social · 11/05/2025

🚨 CG Paper, EG 2025 As scenes & lighting in games grow in complexity, we introduce Neural Incident Radiance Cache (NIRC) – a real-time, online-trainable cache that: 🚄 Costs just ~1ms/neural-sample for 1080p ☘️ Decreases MC variance 🥳 Saves on bounces www.youtube.com/watch?v=Y791...

youtube.com

Neural Two-Level Monte Carlo Real-Time Rendering

YouTube video by Mike Derevyannykh

2309

Reposted by Mike Dereviannykh

Shannon Woods 🦡🖥️👾 @shannonin3d.bsky.social · 04/04/2025

I've finally gotten a chance to put together an in-depth introduction to neural graphics for beginners in Slang using a simplified version of our 2D Gaussian splatting example! shader-slang.org/blog/2025/04...

shader-slang.org

Neural Graphics in an Afternoon

In this article, I’ll show you how to write your first neural graphics program with Slang and SlangPy by walking through our 2D Gaussian Splatting example.

13410

Mike Dereviannykh @mishok43.bsky.social · 14/04/2025

I believe you significantly underestimate the employees' benefits...

010

Reposted by Mike Dereviannykh

Anton @antonhand.bsky.social · 12/04/2025

Reality labs had 17k employees in 2022. 17k employees times say... $130k a year (avg fb salary) is 2.2 billion a year. So 10+ bil of the 45 in salary. Then the office / Benefit costs. Then whatever it costs to make crazy AR glass prototypes with 10k/unit screens. All the AI build-out/compute.

4353

Reposted by Mike Dereviannykh

Nathan Hoobler @hoobler.net · 16/03/2025

To anyone at GDC interested in neural rendering/texture compression: Alexey Panteleev is speaking tomorrow at 9:30am on how the new Cooperative Vector functionality in HLSL was used to implement the RTXNTC library. Start your week off strong! schedule.gdconf.com/session/adva...

schedule.gdconf.com

0135

Mike Dereviannykh @mishok43.bsky.social · 20/02/2025

Stay tuned if you're interested in something like this but in a more applied domain We've been working on it already for a while 😊

010

Reposted by Mike Dereviannykh

Bart Wronski 🇺🇦🇵🇸 @bartwr.bsky.social · 19/02/2025

I'm thrilled to share a new JCGT paper, "GPU Friendly Laplacian Texture Blending." It uses simple image processing and computational photography techniques to smoothly blend/layer textures and materials while preserving local contrast and texture details without ghosting or over-blurring. 1/3

813235

Mike Dereviannykh @mishok43.bsky.social · 09/02/2025

I still don't know frankly speaking... Sometimes I feel that if you're 100% into AI, there's almost no reason to use bsky but sometimes I can find here pretty useful posts about Computer Graphics and Rendering overall

010

Reposted by Mike Dereviannykh

Jonathan Stone @sheriffstone.bsky.social · 20/11/2024

It would be amazing to see the computer graphics community engage on Bluesky, so here's a starter pack for those who might be interested in joining in: go.bsky.app/K8ZL1kq

3305