Designing novel materials is greatly dependent on understanding the design principles, physical mechanisms, and modeling methods of material microstructures, requiring experienced designers with expertise and several rounds of trial and error. Although recent advances in deep generative networks have enabled the inverse design of material microstructures, most studies involve property-conditional generation and focus on a specific type of structure, resulting in limited generation diversity and poor human–computer interaction. In this study, a pioneering text-to-microstructure deep generative network (Txt2Microstruct-Net) is proposed that enables the generation of 3D material microstructures directly from text prompts without additional optimization procedures. The Txt2Microstruct-Net model is trained on a large microstructure-caption paired dataset that is extensible using the algorithms provided. Moreover, the model is sufficiently flexible to generate different geometric representations, such as voxels and point clouds. The model's performance is also demonstrated in the inverse design of material microstructures and metamaterials. It has promising potential for interactive microstructure design when associated with large language models and could be a user-friendly tool for material design and discovery.

PUBLICATION

Zheng, X., Watanabe, I., Paik, J., Li, J., Guo, X., & Naito, M. (2024). Text‐to‐Microstructure Generation Using Generative Deep Learning. Small, 2402685.

https://doi.org/10.1002/smll.202402685

Museum exhibitions, both temporary and permanent, form an essential link between a society and its cultural, historical, and artistic heritage sites. Curating artifacts and thematic displays in museum exhibitions can promote dialogue, foster cultural appreciation, and contribute to heritage preservation. The traditional way of holding museum exhibitions, heavily reliant on the expertise of designers and curatorial staff, makes them a labor-intensive process, from conceptualization to visitor engagement analysis. This review systematically compiles and examines how the application of digital transformation technologies (DTTs) has revolutionized museum exhibitions and augmented their future potential. DTTs such as artificial intelligence, immersive technologies, additive manufacturing, the Internet of Things, and cloud computing can help create engaging designs, improve accessibility and inclusivity, enhance educational potential, and allow for sophisticated visitor experience data collection and analyses, improving exhibit management. However, despite multiple specialized studies on DTTs and their roles in museum exhibitions, the connections between technology and application scenarios remain underexplored. By addressing this gap, this study is expected to inform and inspire practitioners in the museum and heritage sectors and present new research avenues for scholars.

PUBLICATION

Li, J., Zheng, X., Watanabe, I., & Ochiai, Y. (2024). A systematic review of digital transformation technologies in museum exhibition. Computers in Human Behavior, 108407.

https://doi.org/10.1016/j.chb.2024.108407

Mechanical metamaterials are meticulously designed structures with exceptional mechanical properties determined by their microstructures and constituent materials. Tailoring their material and geometric distribution unlocks the potential to achieve unprecedented bulk properties and functions. However, current mechanical metamaterial design considerably relies on experienced designers' inspiration through trial and error, while investigating their mechanical properties and responses entails time-consuming mechanical testing or computationally expensive simulations. Nevertheless, recent advancements in deep learning have revolutionized the design process of mechanical metamaterials, enabling property prediction and geometry generation without prior knowledge. Furthermore, deep generative models can transform conventional forward design into inverse design. Many recent studies on the implementation of deep learning in mechanical metamaterials are highly specialized, and their pros and cons may not be immediately evident. This critical review provides a comprehensive overview of the capabilities of deep learning in property prediction, geometry generation, and inverse design of mechanical metamaterials. Additionally, this review highlights the potential of leveraging deep learning to create universally applicable datasets, intelligently designed metamaterials, and material intelligence. This article is expected to be valuable not only to researchers working on mechanical metamaterials but also those in the field of materials informatics. 

PUBLICATION

Zheng, X., Zhang, X., Chen, T. T., & Watanabe, I. (2023). Deep learning in mechanical metamaterials: from prediction and generation to inverse design. Advanced Materials, 35(45), 2302530.

https://doi.org/10.1002/adma.202302530 

Mechanical metamaterials are artificial structures with structure-dependent properties. They often harness zero-energy deformation modes, e.g., a single shape change that limits their applications, resulting in the need for changeable mechanical responses. We address this limitation by using a flexible material, called light-responsive shape-memory polydimethylsiloxane (SM-PDMS), to introduce reprogrammability into flexible mechanical metamaterials. The SM-PDMS is a rubber-like functional material with shape-memory and photothermal effects. Specfically, we propose three different reprogrammable SM-PDMS metamaterials with different mechanical responses, namely, an auxetic SM-PDMS, a chiral SM-PDMS, and a buckling-induced SM-PDMS. Finally, a buckling-induced SM-PDMS was harnessed to make a soft actuator with a reprogrammable preferred locomotion direction. Despite focusing on reprogramming flexible metamaterials using the light-induced SM effect, our strategy can be easily extended to other structures and smart materials. More importantly, our strategy paves the way to change the mechanical responses for similar architectures. Furthermore, our designed flexible metamaterials have the potential for different applications, such as soft robots, actuation, adaptive safety, and sports equipment.

PUBLICATION

Xiaoyang Zheng, Koichiro Uto, Wei-Hsun Hua, Ta-Te Chen, Masanobu Naito, Ikumu Watanabe. Reprogrammable flexible mechanical metamaterials. Applied Materials Today 29(11):101662. https://doi.org/10.1016/j.apmt.2022.101662

The application of sodium metal batteries (SMBs) is hindered by the notable challenges of side reactions between electrolyte and Na as well as the growth of Na dendrites. Herein, we report that a unique design of dual electrolytes within double separators, namely lean diglyme-based electrolyte hosted by the polypropylene separator together with sulfolane-based electrolyte hosted by the glass fiber separator (denoted as D-e@PP/S-e@GF), can effectively prevent Na dendrite growth and suppress side reactions. D-e@PP/S-e@GF ingeniously combines the advantages of D-e and S-e by using their wettability difference on separators. Finite element method (FEM) simulations and in-situ optical microscopy reveal that the D-e@PP layer promotes uniform deposition and significantly reduces side reactions. The S-e@GF layer can be paired with high voltage cathodes. D-e@PP/S-e@GF enables a high CE of Na plating/stripping as high as 97.22% over 560 cycles at 0.5 mAh cm−2 and ultra-long cycle life of more than 1900 h at 1 mAh cm−2 in Na||Na symmetric cell. Furthermore, this design also demonstrates excellent cycling stability of full cells using Na3V2(PO4) (NVP) and Prussian Blue (PB) as cathodes. The unique design of dual electrolytes within double separators provides a new and promising avenue to develop high-performance SMBs.

PUBLICATION

Ming Zhu, Xiaoyang Zheng, Lulu Li, Xiaolong Zhu, Zhongyi Huang, Guanyao Wang, Yuanjun Zhang et al. "Towards stable sodium metal battery with high voltage output through dual electrolyte design." Energy Storage Materials 48 (2022): 466-474. https://doi.org/10.1016/j.ensm.2022.03.040

Museums and galleries are places that stimulate the imagination and creativity of people. However, against the backdrop of the COVID-19 pandemic, many museums have been forced to close. As a result, more visitors have had to start visiting online exhibitions. Under this trend, the quality of the online exhibition experience is facing unprecedented challenges, and visitors need to be provided with a high-quality online experience. With the rapid development of emerging technologies, museums and galleries have started to use technology to create “digitalization” experiences, such as 3D virtual exhibitions and VR interactive experiences. The current online exhibition format focuses more on improving the visitor’s experience of viewing and browsing works and related information and less on how visitors create their artworks. In order to remedy the experience of visitors actively creating artworks, this study uses Yayoi Kusama’s works as a case study to help form an utterly interactive process between visitors and artworks. We propose a Polka Dot Arts Transfer Network (PDAT-net), mixing realistic-looking and Polka Dot Art-style reference images. The output image looks like an actual natural image “drawn” in the polka dot art style. The PDAT-net is trained to compose an image in the style of Polka Dot Arts, based on neural style transfer. Convolutional networks implemented the training. Plant images performed the final test. The results showed that the generated polka dot images have similar styles to the style images, proving that our PDAT-network has good transfer performance of Polka Dot Arts.

PUBLICATION

Jingjing Li, Xiaoyang Zheng, Jun-Li Lu, Vargas Meza Xanat, Yoichi Ochiai . (2022). Transformation of Plants into Polka Dot Arts: Kusama Yayoi as an Inspiration for Deep Learning. In: Antona, M., Stephanidis, C. (eds) Universal Access in Human-Computer Interaction. Novel Design Approaches and Technologies. HCII 2022. Lecture Notes in Computer Science, vol 13308. Springer, Cham. https://doi.org/10.1007/978-3-031-05028-2_18