Adaptive Linear Path Model-Based Diffusion

2026-02-02

AI Summary

This work addresses the high sensitivity of diffusion models in robotic control to scheduling parameters, which complicates tuning and undermines performance stability. The authors propose a diffusion method based on a linear path model that leverages flow matching to construct geometrically interpretable linear probability paths, replacing conventional variance-preserving schedules and thereby decoupling and simplifying scheduling parameters. Furthermore, they integrate reinforcement learning to enable online adaptive adjustment of the number of diffusion steps and noise levels, dynamically accommodating task and environmental variations. Evaluated on numerical experiments, Brax benchmarks, and mobile robot trajectory tracking tasks, the approach significantly simplifies scheduler design while maintaining high performance, enhancing both robustness and real-time efficiency.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Computer Vision: Diffusion Models for Vision
• Intelligent Robots: State Estimation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

The interest in combining model-based control approaches with diffusion models has been growing. Although we have seen many impressive robotic control results in difficult tasks, the performance of diffusion models is highly sensitive to the choice of scheduling parameters, making parameter tuning one of the most critical challenges. We introduce Linear Path Model-Based Diffusion (LP-MBD), which replaces the variance-preserving schedule with a flow-matching-inspired linear probability path. This yields a geometrically interpretable and decoupled parameterization that reduces tuning complexity and provides a stable foundation for adaptation. Building on this, we propose Adaptive LP-MBD (ALP-MBD), which leverages reinforcement learning to adjust diffusion steps and noise levels according to task complexity and environmental conditions. Across numerical studies, Brax benchmarks, and mobile-robot trajectory tracking, LP-MBD simplifies scheduling while maintaining strong performance, and ALP-MBD further improves robustness, adaptability, and real-time efficiency.

Problem

• diffusion models
• scheduling parameters
• parameter tuning
• model-based control
• robotic control

Innovation

• Linear Path Diffusion
• Model-Based Control
• Flow Matching
• Adaptive Scheduling
• Reinforcement Learning