Structured Coupling for Flow Matching (SCFM) accepted at NeurIPS 2026
liyzhen2 · x · 2026-09-25
The authors announced that Structured Coupling for Flow Matching (SCFM, arXiv:2605.07676) is accepted at NeurIPS 2026. Standard flow matching scales well but uses unstructured source distributions, limiting interpretable latent structure; latent-variable models capture structure but sacrifice generative quality. SCFM bridges the gap with a cooperative framework: VAE-based structured coupling, shared networks between VAE encoder and flow, and VAE-decoder-initialized sampling refined by flow. A shared time-dependent recognition network handles both variational inference and flow velocity estimation. SCFM enables unsupervised latent representation learning for clustering, disentanglement and downstream tasks while matching standard flow matching in sample quality.
More from Research
- EMNLP paper analyzes 79,286 posts: MT users and AI devs sharply disagree on what matters — EhudReiter · 2026-09-25
- NVIDIA Open-Sources Nemotron-Terminal: Boosts Qwen3-32B from 3.4% to 27.4% on Terminal-Bench 2.0 — _weiping · 2026-09-25
- Cusp AI co-founder Max Welling on AI-powered materials design — NandoDF · 2026-09-25
- Sharpa's World Synesthesia Model for dexterous hands accepted at CoRL 2026 — jiqizhixin · 2026-09-25
- DeltaWAM cuts video model cost for bimanual manipulation, lifting RoboTwin success to 85.4% — Han Yan · 2026-09-25
- Amazon open-sources Rufus-Air: full 8-stage post-training recipe on GLM-4.5-Air-Base — amazon · 2026-09-25