The ScholarNet and Artificial Intelligence (AI) Supervisor in Material Science Research.

The past decade has witnessed extensive applications of artificial intelligence (AI) and robotics in chemistry and material science. However, the current focus mainly revolves around idea execution, neglecting the significance of idea generation, which plays a pivotal role in determining research novelty and potential breakthroughs. Concurrently, the exponential growth of scientific publications has resulted in overpublishing, making it challenging for researchers to grasp multiple fields effectively. As most opportunities for innovation lie in interdisciplinary realms, there is a risk of missing out on the development of new ideas. To address these challenges, we present a deep learning-based AI supervisor trained on correlation-based ScholarNet data of publications. Primarily tailored for material science, this AI supervisor excels in recommending research ideas, analyzing their novelty, and providing comprehensive guidance to researchers. By offering invaluable support in idea generation and novelty assessment, our AI supervisor has emerged as a promising digital infrastructure for future material science research.

Tensor Network-Encrypted Physical Anti-counterfeiting Passport for Digital Twin Authentication.

The trend of digitalization has produced rapidly increasing data interaction and authentication demand in today's internet of things ecosystem. To face the challenge, we demonstrated a micro-scale label by direct laser writing to perform as a passport between the physical and digital worlds. On this label, the user information is encrypted into three-dimensional geometric structures by a tensor network and then authenticated through the decryption system based on computer vision. A two-step printing methodology is applied to code the randomly distributed fluorescence from doped quantum dots, which achieved physical unclonable functions (PUFs) of the passport. The 105 bits/mm2 data storage density enables abundant encrypted information from physical worlds, for example, the biometric data of human users. This passport guarantees the strong correlation between the user's privacy data and the PUF-assisted codes, successfully overcoming the illegal transfer of authentication information. Due to its ultra-high security level and convenience, the printed passport has enormous potential in future digital twin authentication anytime anywhere, including personal identity, valuable certificates, and car networking.

Catalyst deep neural networks (Cat-DNNs) in singlet fission property prediction.

Many current deep neural network (DNN) models only focus on straightforward optimization over the given database. However, most numerical fitting procedures depart from physical laws. By introducing the concept of "catalysis" from physical chemistry, we propose that the physical correlations among molecular properties could spontaneously act as a catalyst in the DNNs, which increases the accuracy, and more importantly, guides the DNNs in the right way. These Catalysis-DNNs (Cat-DNNs) could precisely predict both the ground and excited-state properties, especially the molecules' screening with singlet fission character. We show that traditional machine learning metrics are not suitable for evaluating model accuracy in physical-chemical tasks and issue new physical errors. We believe that the agile transfer of fundamental physics or chemistry domain knowledge, like the catalyst, could significantly benefit both the architecture and application of artificial intelligence technology in the future.

QM-symex, update of the QM-sym database with excited state information for 173 kilo molecules.

In the research field of material science, quantum chemistry database plays an indispensable role in determining the structure and properties of new material molecules and in deep learning in this field. A new quantum chemistry database, the QM-sym, has been set up in our previous work. The QM-sym is an open-access database focusing on transition states, energy, and orbital symmetry. In this work, we put forward the QM-symex with 173-kilo molecules. Each organic molecular in the QM-symex combines with the Cnh symmetry composite and contains the information of the first ten singlet and triplet transitions, including energy, wavelength, orbital symmetry, oscillator strength, and other quasi-molecular properties. QM-symex serves as a benchmark for quantum chemical machine learning models that can be effectively used to train new models of excited states in the quantum chemistry region as well as contribute to further development of the green energy revolution and materials discovery.

Symmetrical Graph Neural Network for Quantum Chemistry with Dual Real and Momenta Space.

Most of the current neural network models in quantum chemistry (QC) exclude the molecular symmetry and separate the well-correlated real space (R space) and momenta space (K space) into two individuals, which lack the essential physics in molecular chemistry. In this work, by endorsing the molecular symmetry and elementals of group theory, we propose a comprehendible method to apply symmetry in the graph neural network (SY-GNN), which extends the property-predicting coverage to orbital symmetry for both ground and excited states. SY-GNN is an end-to-end model that can predict multiple properties in both K and R space within a single model, and it shows excellent performance in predicting both the absolute and relative R and K space quantities. Besides the numerical properties, SY-GNN can also predict orbital properties, providing the active regions of chemical reactions. We believe the symmetry-endorsed deep learning scheme covers the significant physics inside and is essential for the application of neural networks in QC and many other research fields in the future.

AIR-Chem: Authentic Intelligent Robotics for Chemistry.

The new era with prosperous artificial intelligence (AI) and robotics technology is reshaping the materials discovery process in a more radical fashion. Here we present authentic intelligent robotics for chemistry (AIR-Chem), integrated with technological innovations in the AI and robotics fields, functionalized with modules including gradient descent-based optimization frameworks, multiple external field modulations, a real-time computer vision (CV) system, and automated guided vehicle (AGV) parts. AIR-Chem is portable and remotely controllable by cloud computing. AIR-Chem can learn the parametric procedures for given targets and carry on laboratory operations in standalone mode, with high reproducibility, precision, and availability for knowledge regeneration. Moreover, an improved nucleation theory of size focusing on inorganic perovskite quantum dots (IPQDs) is theoretically proposed and experimentally testified to by AIR-Chem. This work aims to boost the process of an unmanned chemistry laboratory from the synthesis of chemical materials to the analysis of physical chemical properties, and it provides a vivid demonstration for future chemistry reshaped by AI and robotics technology.