Adobe et al. Present Variational r-Adaptive Cloth Simulation Method
ssh4net · x · 2026-08-19
Researchers from Adobe and USC published a paper on cloth simulation, introducing the first variational r-adaptive method for modern cloth pipelines capable of handling frictional contact dynamics and statics.
Core Problem:
Thin cloth requires high effective spatial resolution to reproduce wrinkles, folds, buckling, and sharp contact features. However, applying existing variational r-adaptivity to piecewise-linear shells reveals two coupled failure modes:
- Discretized incremental-potential (IP) optimization can become trapped in poor local minima, yielding suboptimal physical configurations.
- It can artificially lower IP by collapsing elements, invalidating the finite-element approximation.
Solution:
The paper addresses both issues with "degeneracy-activated quality regularization." This regularizer remains inactive when mesh quality is good and activates only when elements degenerate, preventing optimization traps and numerical failures.
More from Research
- Anthropic Sets Up Bay Area Wet Lab, Plans to Have Claude Run Experiment Robots — FinanceYF5 · 2026-09-21
- KLPO: a critic-free, single-rollout RL method for agentic LLMs, framed as 'Q* solved' — inductionheads · 2026-09-21
- FutureHouse publishes 'Millennium Problems for Biology' as wetlab-verifiable evals for AI — burny_tech · 2026-09-21
- No scientific test for consciousness exists even in humans — how can we declare AI unconscious? — burny_tech · 2026-09-21
- Formal Reasoning in AI: Song guest-lectures Virginia Tech course, slides public — wellecks · 2026-09-21
- Biologist: FutureHouse's Bio Millennium List Is Too Narrow — Really Just Synbio — burny_tech · 2026-09-21