SCALLOP paper at NeurIPS 2026 trains few-step flow maps with tractable likelihood for Boltzmann generation
gottapatchemall · x · 2026-10-17
The team introduces SCALLOP, a new paper accepted at NeurIPS 2026. Flow maps are popular, but scientific applications like Boltzmann generation still require the likelihood of the flow. SCALLOP provides a scalable, efficient way to train few-step flow maps with a tractable likelihood, enabling effective Boltzmann generation — sampling from Boltzmann distributions in few steps. Paper and project page are available.
Related event: SCALLOP: Few-Step Flow Maps with Likelihoods Accepted to NeurIPS(2 posts)→
More from Research
- APS Global Physics Summit opens Neural Systems session, abstracts due Oct 2026 — chklovskii · 2026-10-17
- RAX open-sourced: frozen coding agents learn robot tasks via online exploration, zero demos — du_yilun · 2026-10-17
- Open textbook "Elements of Robotics" by Ben-Ari & Mondada now freely available — burkov · 2026-10-17
- U. Tokyo's TwinX robot: one wire links two wheeled-legged bots for morphing and teamwork — LexiLove · 2026-10-17
- NVIDIA open-sources midtraining recipe: 500B tokens lift Nemotron agentic benchmark scores — denisyarats · 2026-10-17
- PySR v2 ships after two years of work: plugins, custom types, much faster — MilesCranmer · 2026-10-17