HERS: Hidden-Pattern Expert Learning for Risk-Specific Vehicle Damage Adaptation in Diffusion Models

ICLR 2026 Conference Submission18319 Authors

Keywords: Text-to-Image (T2I) Diffusion, Vehicle Damage Synthesis, Domain-Specific Generative Modeling, Self-Supervised Expert Adaptation, Fraud Detection / Safety-Critical AI, Controllable Image Generation

TL;DR: HERS enables domain-specific, self-supervised text-to-image diffusion for realistic vehicle damage generation, improving fidelity, controllability, and safety in high-stakes insurance applications.

Abstract: Recent advances in text-to-image (T2I) diffusion models have enabled increasingly realistic synthesis of vehicle damage, raising concerns about their reliability in automated insurance workflows. The ability to generate crash-like imagery challenges the boundary between authentic and synthetic data, introducing new risks of misuse in fraud or claim manipulation. To address these issues, we propose HERS (Hidden-Pattern Expert Learning for Risk-Specific Damage Adaptation), a framework designed to improve fidelity, controllability, and domain alignment of diffusion-generated damage images. HERS fine-tunes a base diffusion model via domain-specific expert adaptation without requiring manual annotation. Using self-supervised image–text pairs automatically generated by a large language model and T2I pipeline, HERS models each damage category—such as dents, scratches, broken lights, or cracked paint—as a separate expert. These experts are later integrated into a unified multi-damage model that balances specialization with generalization. We evaluate HERS across four diffusion backbones and observe consistent improvements: +5.5\% in text faithfulness and +2.3\% in human preference ratings compared to baselines. Beyond image fidelity, we discuss implications for fraud detection, auditability, and safe deployment of generative models in high-stakes domains. Our findings highlight both the opportunities and risks of domain-specific diffusion, underscoring the importance of trustworthy generation in safety-critical applications such as auto insurance.

Primary Area: alignment, fairness, safety, privacy, and societal considerations

Submission Number: 18319