AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Latent Barrier Steering: Hierarchical Safety for Generative Planning

Renhao Zhang, Mingzhe Li, Bruno Silva

NeurIPS 2026 · Image generation › Fast diffusion sampling · approach: Guidance and alignment for flow matching · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes Latent Barrier Steering (LBS), a hierarchical semantic-to-path framework that applies safety barriers to flow-matching generative planners to handle constraints that block a whole selected mode.

Problem

Control-barrier-function safety filters for diffusion and flow-matching planners act locally on generated samples, which can fail when a new constraint blocks the entire selected behavior mode.

Approach

Moves safety handling into a hierarchical latent space: a semantic level selects behavior modes and barrier steering operates on latent variables before path generation, rather than repairing individual generated samples locally during or after sampling.

Experimental setup

Baselines: CBF safety filters for diffusion planners CBF safety filters for flow-matching planners

Builds on: control barrier function flow matching

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy