Want to efficiently compute smooth, conservative distance fields on co-dimensional geometry ? Then @_abhishekmadan's #SIGGRAPH2022 paper has you covered. Super useful for collision detection and resolution on things like lidar data! Great work !.

Any “real” content we place here effectively content-washes that and I just don’t want to be part of that anymore.

There is still a lot of great graphics and science-based content here but it’s also abundantly clear that Xwitter is effectively a propaganda/recruiting arm for a very dangerous brand of alt-right ideology.

Like many people I’ve reached the conclusion that it’s time to leave Xwitter . I’ll be here now: @diwlevin.bsky.social . Reasons:.

RT @rms80: if you would like to talk about something else today - like unreal engine's interactive tools framework, geometry script, modeli….

30 years of graphics and we’re back to “nah blocks are fine” ;).

It's rec letter time again 🙃. Time to answer the tough questions

RT @nmwsharp: My team at nvidia is recruiting PhD research interns for next year. Do you like ML + graphics/vision/geometry/simulation? App….

The HiFi Physics Team:

The NVIDIA HiFi Physics team is looking for interns with interest/skills in. 1⃣ Physics simulation (solids/fluids/differentiable etc).2⃣ Digital Humans.3⃣ Neural Rendering/Geometry, Video Models.4⃣ Monte Carlo Methods.5⃣ Digital Fabrication . Apply:

Growing up in the cynical and sarcastic 90s has been great protection against the age of omnipresent hype.

7b. Caveat -- if the citation is extremely obscure that does not constitute "obvious" -- consider the value to the community of republishing useful but less known ideas with proper attribution.

7. Everything is obvious in hindsight (ie any elegant algorithm can seem obvious once you read the paper). It's only obvious if it was previously published and you can provide the citation.

Be polite . there are real humans behind these submissions who worked hard. Make sure you critiques are directed to the submission not the authors. Bad: "Authors didn't do this".Good: "The submission does not discuss Y".

6.

6. It's ok to change your opinion based on other reviews or the rebuttal (READ THE REBUTTAL SERIOUSLY). You will make mistakes or misunderstand things. That's fine, just be open to being corrected (either +ve or -ve).

5. Number 4 again and again and again.Bad: "This paper is not novel".Good: "Unsure if method is significantly better than sota. Smith et al. propose an identical algorithm with the only difference being step 7. L2 accuracy is only 1% better in the proposed method on same data".

4. Be evidence-driven. Every critique in your review must be backed up with evidence. Bad: "paper doesn't have enough evaluation". Better: "previous works on topic over the last 5 yrs (see paper 1,2,3,4) include following ablation study to test blah. Why is it omitted here ?".

3. Know your role: A tertiary reviewer (non-pc) is meant to give an in-depth technical vetting of the paper. Your reviews should be long and detailed. Don't accept more reviews if you won't have time to do this for each one.

2. Read the reviewer guidelines. Familiarize yourself with the SIGGRAPH ethos (preference to new exciting ideas etc. ) and the rules (ie cannot hold missing comparisons against a submission unless you can provide a link to code/data and comparison is reasonable amount of work.

1. Read the desiderata for each track on the SIGGRAPH website. Authors will write their submissions with these in mind. To be fair to authors any decision you make must be justified using these desiderata.