通过不断进化，生物体形成了适合其生存环境的复杂材料结构。基于这些久经考验的天然设计，人工仿生结构呈现出优越的材料配置。尽管如此，通过调控结构特征来获得理想材料性能的方式依然是重要的设计策略。本研究工作将针对人体最坚硬的生物组织——牙釉质，阐明其结构与性能之间的关系。材料硬度的实验测量方法耗时且具有破坏性，而本文提出的人工智能模型能够直接预测材料性能，实现高通量的无损表征。在模型中，将深度图像回归神经网络作为代理模型进行训练，通过梯度上升和显著图的可视化来识别硬度最高的材料结构特征。与实验硬度图相比，该模型能够提高空间分辨率和灵敏度。利用这一快速硬度测试模型，通过耦合对比生成模型和具有潜在空间的遗传算法来实现材料设计，从而精准构筑硬度可控的仿生结构。
Over evolution, organisms develop complex material structures fit to their environments. Based on these time-tested designs, human-engineered bioinspired structures offer exciting possible materials configurations. However, navigating diverse structure spaces for attaining desired properties remains nontrivial. We focus on the hardest biological tissue in humans, tooth enamel, to examine the structure-property relationship. While typical hardness measurements are time consuming and destructive, we propose that artificial intelligence models can predict properties directly and enable high-throughput nondestructive characterization. We train a deep image regression neural network as a surrogate model, and visualize with gradient ascent and saliency maps to identify structural features contributing most to hardness. This model demonstrates improved spatial resolution and sensitivity compared to experimental hardness maps. Using this rapid hardness testing model, a generative adversarial model, and a genetic algorithm that operates in latent space allows for guided materials design, yielding proposed designs for bioinspired structures with precisely controlled hardness.
