@_TheodoreKim One year I finished up and went to sleep at 9pm. I woke up and the clock said 7 so I thought, great, I’ll get a coffee and a bagel! The coffee shop was empty, and as I sat down they started mopping the floors. I thought huh, that’s weird, and realized it was 7pm! 22 hours asleep!
Check out @DonglaiXiang's cool new @SIGGRAPHAsia 2022 paper on combing learned appearance models with physically-based clothing simulations. Awesome looking results! https://t.co/tAH17usJc7
@topher_batty @ana_dodik@JustinMSolomon@_AlecJacobson TikTok sponsorships to promote publications! *A little dance, points at text bubble* “Adaptive octree simulations” *Makes a funny face at new text bubble* “Not practical, right?” *Smiles and does new dance as exciting music drops and new text bubble appears* “Not anymore!”
@milesmacklin@_AlecJacobson I’ve come across references in mechanics from at least the 1970s! (https://t.co/L3LwShdQXq) but have not yet dived backwards in depth through that literature
@amirvaxman_dgp@_AlecJacobson R0 is slightly different than local-global. Without dynamics, I agree local-global is just a different optimization strategy (equivalent to replacing ARAP’s Hessian with a Laplacian, Kovalsky’s AQP paper I think?). For a given dt R0 vs R(F) is a different objective at each step.
@amirvaxman_dgp@_AlecJacobson The compute a fixed R0 at the start of each timestep energy is *not* rotaionally variant, I believe. In the context of dynamics this would break angular momentum conservation.
Announcing WiGRAPH's Rising Stars in Computer Graphics! Ph.D. students and postdocs of underrepresented genders: apply for a two-year program of mentorship and workshops co-located with SIGGRAPH 2022&2023. Travel support provided. https://t.co/M8KO569Kay
#WiGRAPHRisingStars [1/6]
@_AlecJacobson Synthesis? Image, motion, design architechture… But then again understanding is inextricably linked to synthesis, and papers on purely understanding are presented in graphics venues. Practically/cynically: Whatever the ‘community’ feels like reviewing at any given moment.
@_AlecJacobson @_TheodoreKim I bet even a less aggressive line search would give a gentle deflation, here. IIRC the inversion avoiding LS *wants* a singularity at the origin or it can lock, this was in the SLIM paper for ARAP. ‘Log-variant’ NH should deflate nicely here with the inversion aware LS, I bet.
@_AlecJacobson @_TheodoreKim Yep, pages 8/9 of https://t.co/EwWe1wjLwZ. Would be curious to see if Stable Neo-Hookean also locks with the inversion avoiding line search!