analiza aiqbench
This study applies cooperative neural networks to solve the partial inverse design problem in high-performance concrete, enabling the generation of valid mix designs that meet specific strength targets without retraining. By integrating imputation models with surrogate predictors, the framework achieves strong predictive consistency and significantly reduces error compared to baseline autoencoders and Gaussian process methods. The research highlights a practical application of machine learning techniques to optimize complex engineering constraints in material science.
Partial Inverse Design of High-Performance Concrete Using Cooperative Neural Networks for Constraint-Aware Mix Generation
arXiv:2512.06813v3 Announce Type: replace-cross Abstract: High-performance concrete (HPC) requires complex mix design decisions involving interdependent variables and practical constraints. While data-driven methods have improved predictive modeling for forward design in concrete engineering, inverse design remains limited, especially when some variables are fixed and only the remaining ones must be inferred. This study proposes a cooperative neural network framework for the partial inverse design of HPC. The framework integrates an imputation model with a surrogate strength predictor and learns through cooperative training. Once trained, it generates valid and performance-consistent mix designs in a single forward pass without retraining for different constraint scenarios. Compared with baseline models, including autoencoder models and Bayesian inference with Gaussian process surrogates, the proposed method achieves strength consistency between the surrogate-predicted strength of the generated mixes and the target strength with R-squared values of 0.84 to 0.89 and substantially reduces the mean squared error of this strength consistency by approximately 42% and 60%, respectively. The results demonstrate a novel, accurate, and computationally efficient application of artificial intelligence in concrete science by applying a cooperative neural network for constraint-aware partial inverse design of HPC mix generation.
Abstract:High-performance concrete (HPC) requires complex mix design decisions involving interdependent variables and practical constraints. While data-driven methods have improved predictive modeling for forward design in concrete engineering, inverse design remains limited, especially when some variables are fixed and only the remaining ones must be inferred. This study proposes a cooperative neural network framework for the partial inverse design of HPC. The framework integrates an imputation model with a surrogate strength predictor and learns through cooperative training. Once trained, it generates valid and performance-consistent mix designs in a single forward pass without retraining for different constraint scenarios. Compared with baseline models, including autoencoder models and Bayesian inference with Gaussian process surrogates, the proposed method achieves strength consistency between the surrogate-predicted strength of the generated mixes and the target strength with R-squared values of 0.84 to 0.89 and substantially reduces the mean squared error of this strength consistency by approximately 42% and 60%, respectively. The results demonstrate a novel, accurate, and computationally efficient application of artificial intelligence in concrete science by applying a cooperative neural network for constraint-aware partial inverse design of HPC mix generation.
22 pages, 12 figures. All experiments were rerun under a revised training protocol with an improved Bayesian-GP baseline. Significance tests and a clipping ablation appendix were added, and the abstract, figures, and tables are updated accordingly
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2512.06813 [cs.LG]
(or arXiv:2512.06813v3 [cs.LG] for this version)
Text integral sub licența deschisă a publicației · 233 cuvinte
Subiecte conexe
AI Model Co-Design: Hardware-Friendly LLM Design
NVIDIA · 10 iul. 2026
[AINews] not much happened today
Latent Space · 3 oct. 2026
[AINews] Reality Checks on AI News (Yegge shuts down Gas Town, Databricks’ +60% Astra cost)
Latent Space · 17 sept. 2026
The basic argument for making AI better at moral philosophy
LessWrong · 14 oct. 2026