Constraining Generative Models for Engineering Design with Negative Data
This addresses safety-critical constraints in engineering design, such as ship hull synthesis and vehicle safety, representing a novel method for a known bottleneck rather than a foundational advancement.
The paper tackles the problem of generative models producing unrealistic or unsafe outputs in engineering design by introducing a training method that uses negative data to guide models toward constraint-satisfying results, achieving 1/6 as many constraint-violating samples with 1/8 as much data in some cases and outperforming baselines in 12 out of 14 tested problems.
Generative models have recently achieved remarkable success and widespread adoption in society, yet they often struggle to generate realistic and accurate outputs. This challenge extends beyond language and vision into fields like engineering design, where safety-critical engineering standards and non-negotiable physical laws tightly constrain what outputs are considered acceptable. In this work, we introduce a novel training method to guide a generative model toward constraint-satisfying outputs using `negative data' -- examples of what to avoid. Our negative-data generative model (NDGM) formulation easily outperforms classic models, generating 1/6 as many constraint-violating samples using 1/8 as much data in certain problems. It also consistently outperforms other baselines, achieving a balance between constraint satisfaction and distributional similarity that is unsurpassed by any other model in 12 of the 14 problems tested. This widespread superiority is rigorously demonstrated across numerous synthetic tests and real engineering problems, such as ship hull synthesis with hydrodynamic constraints and vehicle design with impact safety constraints. Our benchmarks showcase both the best-in-class performance of our new NDGM formulation and the overall dominance of NDGMs versus classic generative models. We publicly release the code and benchmarks at https://github.com/Lyleregenwetter/NDGMs.