# From Attacks to Curricula: Learnability-Guided Adversarial Training for Safe Autonomous Driving

arXiv: https://arxiv.org/abs/2606.14032 · HTML: https://mlsanity.com/paper/2606.14032/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Adversarial scenarios and safety testing, one of 9 directions in End-to-end autonomous driving.
- Builds on 23 of its 28 references that are in the map.
- First posted on arXiv 12 Jun 2026.

## Placed in

- End-to-end autonomous driving / [Adversarial scenarios and safety testing](https://mlsanity.com/driving/adversarial-scenarios-and-safety-testing/)

## Builds on (most cited first, 23 of 23)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Large Scale Interactive Motion Forecasting for Autonomous Driving : The Waymo Open Motion Dataset](https://mlsanity.com/paper/2104.10133/) (2021)
- [MetaDrive: Composing Diverse Driving Scenarios for Generalizable Reinforcement Learning](https://mlsanity.com/paper/2109.12674/) (2021)
- [Generating Useful Accident-Prone Driving Scenarios via a Learned Traffic Prior](https://mlsanity.com/paper/2112.05077/) (2021)
- [KING: Generating Safety-Critical Driving Scenarios for Robust Imitation via Kinematics Gradients](https://mlsanity.com/paper/2204.13683/) (2022)
- [CAT: Closed-loop Adversarial Training for Safe End-to-End Driving](https://mlsanity.com/paper/2310.12432/) (2023)
- [Autonomous Overtaking in Gran Turismo Sport Using Curriculum Reinforcement Learning](https://mlsanity.com/paper/2103.14666/) (2021)
- [SAFE-SIM: Safety-Critical Closed-Loop Traffic Simulation with Diffusion-Controllable Adversaries](https://mlsanity.com/paper/2401.00391/) (2023)
- [CausalAF: Causal Autoregressive Flow for Safety-Critical Driving Scenario Generation](https://mlsanity.com/paper/2110.13939/) (2021)
- [AdvDiffuser: Generating Adversarial Safety-Critical Driving Scenarios via Guided Diffusion](https://mlsanity.com/paper/2410.08453/) (2024)
- [Embedding Synthetic Off-Policy Experience for Autonomous Driving via Zero-Shot Curricula](https://mlsanity.com/paper/2212.01375/) (2022)
- [Safety-Critical Scenario Generation Via Reinforcement Learning Based Editing](https://mlsanity.com/paper/2306.14131/) (2023)
- [FREA: Feasibility-Guided Generation of Safety-Critical Scenarios with Reasonable Adversariality](https://mlsanity.com/paper/2406.02983/) (2024)
- [Adv-BMT: Bidirectional Motion Transformer for Safety-Critical Traffic Scenario Generation](https://mlsanity.com/paper/2506.09485/) (2025)
- [Reward-Driven Automated Curriculum Learning for Interaction-Aware Self-Driving at Unsignalized Intersections](https://mlsanity.com/paper/2403.13674/) (2024)
- [GOOSE: Goal-Conditioned Reinforcement Learning for Safety-Critical Scenario Generation](https://mlsanity.com/paper/2406.03870/) (2024)
- [Steerable Adversarial Scenario Generation through Test-Time Preference Alignment](https://mlsanity.com/paper/2509.20102/) (2025)
- [Seeking to Collide: Online Safety-Critical Scenario Generation for Autonomous Driving with Retrieval Augmented Large Language Models](https://mlsanity.com/paper/2505.00972/) (2025)
- [Automatic Curriculum Learning for Driving Scenarios: Towards Robust and Efficient Reinforcement Learning](https://mlsanity.com/paper/2505.08264/) (2025)
- [Heterogeneous Adversarial Play in Interactive Environments](https://mlsanity.com/paper/2510.18407/) (2025)
- [AED: Automatic Discovery of Effective and Diverse Vulnerabilities for Autonomous Driving Policy with Large Language Models](https://mlsanity.com/paper/2503.20804/) (2025)
- [VCAT: Vulnerability-aware and Curiosity-driven Adversarial Training for Enhancing Autonomous Vehicle Robustness](https://mlsanity.com/paper/2409.12997/) (2024)
- [RCG: Safety-Critical Scenario Generation for Robust Autonomous Driving via Real-World Crash Grounding](https://mlsanity.com/paper/2507.10749/) (2025)

## Abstract (arXiv)

Closed-loop adversarial training improves autonomous driving safety by exposing policies to rare safety-critical scenarios. Standard pipelines first generate adversarial scenarios and then sample them for policy optimization. However, most existing frameworks remain attack-oriented: collision-driven generators often synthesize unsolvable extreme situations, which can degrade learning, while heuristic samplers ignore the evolving capability of the driving policy, causing sample inefficiency and delayed convergence. We propose AlignADV, a learnability-guided closed-loop adversarial training framework that converts adversarial scenarios into resolvable and capability-aligned curricula. First, we reformulate adversarial scenario generation as a preference alignment problem and employ direct preference optimization to guide the generator toward critical yet resolvable scenarios. Second, we introduce behavioral fingerprints to capture the intrinsic characteristics of the evolving policy and construct a multi-modal capability prediction model that estimates policy performance without expensive closed-loop simulations. By combining resolvability-aligned scenarios with capability predictions, AlignADV develops a dynamic curriculum sampling mechanism that prioritizes scenarios targeting the current policy's vulnerabilities. Experiments on the Waymo Open Motion Dataset demonstrate that AlignADV improves convergence efficiency and final performance, reducing training steps by up to 40.6 percent compared with baseline methods while lowering collision rate and improving route completion under both normal and adversarial traffic conditions. These results highlight a shift from attack-oriented scenario generation to learnability-guided policy improvement, offering a principled direction for safer and more efficient autonomous driving training. Project page: https://meiyuewen.github.io/AlignADV/.