# Diffusion-Based Failure Sampling for Evaluating Safety-Critical Autonomous Systems

arXiv: https://arxiv.org/abs/2406.14761 · HTML: https://mlsanity.com/paper/2406.14761/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Placed in Diffusion and flow policies for robotics, one of 14 directions in Diffusion models.
- Builds on 19 of its 35 references that are in the map.
- First posted on arXiv 20 Jun 2024.

## Placed in

- Diffusion models / [Diffusion and flow policies for robotics](https://mlsanity.com/diffusion/diffusion-and-flow-policies-for-robotics/)

## Builds on (most cited first, 19 of 19)

- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [Score-Based Generative Modeling through Stochastic Differential Equations](https://mlsanity.com/paper/2011.13456/) (2020)
- [Improved Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2102.09672/) (2021)
- [UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction](https://mlsanity.com/paper/1802.03426/) (2018)
- [Planning with Diffusion for Flexible Behavior Synthesis](https://mlsanity.com/paper/2205.09991/) (2022)
- [Reliable Fidelity and Diversity Metrics for Generative Models](https://mlsanity.com/paper/2002.09797/) (2020)
- [Motion Planning Diffusion: Learning and Planning of Robot Motions with Diffusion Models](https://mlsanity.com/paper/2308.01557/) (2023)
- [Self-Driving Cars: A Survey](https://mlsanity.com/paper/1901.04407/) (2019)
- [Scalable End-to-End Autonomous Vehicle Testing via Rare-event Simulation](https://mlsanity.com/paper/1811.00145/) (2018)
- [A Survey of Algorithms for Black-Box Safety Validation of Cyber-Physical Systems](https://mlsanity.com/paper/2005.02979/) (2020)
- [Neural Bridge Sampling for Evaluating Safety-Critical Autonomous Systems](https://mlsanity.com/paper/2008.10581/) (2020)
- [Efficient Black-box Assessment of Autonomous Vehicle Safety](https://mlsanity.com/paper/1912.03618/) (2019)
- [DiffAIL: Diffusion Adversarial Imitation Learning](https://mlsanity.com/paper/2312.06348/) (2023)
- [Falsification of Cyber-Physical Systems Using Deep Reinforcement Learning](https://mlsanity.com/paper/1805.00200/) (2018)
- [ShiftDDPMs: Exploring Conditional Diffusion Models by Shifting Diffusion Trajectories](https://mlsanity.com/paper/2302.02373/) (2023)
- [A Flow-Based Generative Model for Rare-Event Simulation](https://mlsanity.com/paper/2305.07863/) (2023)
- [Model-based Validation as Probabilistic Inference](https://mlsanity.com/paper/2305.09930/) (2023)
- [Requirement falsification for cyber-physical systems using generative models](https://mlsanity.com/paper/2310.20493/) (2023)
- [Transfer Learning for Efficient Iterative Safety Validation](https://mlsanity.com/paper/2012.05336/) (2020)

## Built on by (most cited first, 6 of 6)

- [Predictive Uncertainty for Runtime Assurance of a Real-Time Computer Vision-Based Landing System](https://mlsanity.com/paper/2508.09732/) (2025)
- [Diffusion Models for Safety Validation of Autonomous Driving Systems](https://mlsanity.com/paper/2506.08459/) (2025)
- [The FABRIC Strategy for Verifying Neural Feedback Systems](https://mlsanity.com/paper/2603.08964/) (2026)
- [Robust Planning for Autonomous Vehicles with Diffusion-Based Failure Samplers](https://mlsanity.com/paper/2507.11991/) (2025)
- [A New Strategy for Verifying Reach-Avoid Specifications in Neural Feedback Systems](https://mlsanity.com/paper/2601.08065/) (2026)
- [Importance Sampling and PCA for Finding Failures in Commercial Autonomous Vehicles](https://mlsanity.com/paper/2607.18106/) (2026)

## Abstract (arXiv)

Validating safety-critical autonomous systems in high-dimensional domains such as robotics presents a significant challenge. Existing black-box approaches based on Markov chain Monte Carlo may require an enormous number of samples, while methods based on importance sampling often rely on simple parametric families that may struggle to represent the distribution over failures. We propose to sample the distribution over failures using a conditional denoising diffusion model, which has shown success in complex high-dimensional problems such as robotic task planning. We iteratively train a diffusion model to produce state trajectories closer to failure. We demonstrate the effectiveness of our approach on high-dimensional robotic validation tasks, improving sample efficiency and mode coverage compared to existing black-box techniques.