Deep Generative Methods and Tire Architecture Design

2025-07-15

AI Summary

Industrial tire design involves complex multi-component architectures, necessitating generative models capable of unconditional generation, part-conditioned synthesis, and dimension-constrained output. Method: This work systematically evaluates five deep generative models—VAE, GAN, MMVAE, DDPM, and MDM—across these three scenarios. We introduce a categorical imputation mechanism enabling discrete diffusion models to preserve known categorical labels in conditional generation without retraining. Additionally, we propose a geometry-aware evaluation framework quantifying structural coherence, inter-part compatibility, and perceptual quality. Results: Diffusion models achieve overall superior performance: MDM excels in in-distribution generation, while DDPM demonstrates stronger out-of-distribution generalization to unseen dimensions. Notably, masked-training VAE surpasses MMVAE+ on part-conditioned tasks. The study establishes a reproducible model selection paradigm and delivers key technical enablers—including label-preserving conditioning and geometric evaluation—for industrial generative design.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Computer Vision: Diffusion Models for Vision
• Natural Language Processing: Generation

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Web Mining and Content Analysis: Web data generation and simulation

Abstract

As deep generative models proliferate across the AI landscape, industrial practitioners still face critical yet unanswered questions about which deep generative models best suit complex manufacturing design tasks. This work addresses this question through a complete study of five representative models (Variational Autoencoder, Generative Adversarial Network, multimodal Variational Autoencoder, Denoising Diffusion Probabilistic Model, and Multinomial Diffusion Model) on industrial tire architecture generation. Our evaluation spans three key industrial scenarios: (i) unconditional generation of complete multi-component designs, (ii) component-conditioned generation (reconstructing architectures from partial observations), and (iii) dimension-constrained generation (creating designs that satisfy specific dimensional requirements). To enable discrete diffusion models to handle conditional scenarios, we introduce categorical inpainting, a mask-aware reverse diffusion process that preserves known labels without requiring additional training. Our evaluation employs geometry-aware metrics specifically calibrated for industrial requirements, quantifying spatial coherence, component interaction, structural connectivity, and perceptual fidelity. Our findings reveal that diffusion models achieve the strongest overall performance; a masking-trained VAE nonetheless outperforms the multimodal variant MMVAE extsuperscript{+} on nearly all component-conditioned metrics, and within the diffusion family MDM leads in-distribution whereas DDPM generalises better to out-of-distribution dimensional constraints.

Problem

• Evaluating deep generative models for industrial tire design tasks
• Addressing conditional generation in tire architecture scenarios
• Comparing model performance using geometry-aware industrial metrics

Innovation

• Evaluates five deep generative models for tire design
• Introduces categorical inpainting for conditional generation
• Uses geometry-aware metrics for industrial performance assessment