Molecular latent representations, derived from autoencoders (AEs), have been widely used for drug or material discovery over the past couple of years. In particular, a variety of machine learning methods based on latent representations have shown excellent performance on quantitative structure–activity relationship (QSAR) modeling. However, the sequence feature of them has not been considered in most cases. In addition, data scarcity is still the main obstacle for deep learning strategies, especially for bioactivity datasets. In this study, we propose the convolutional recurrent neural network and transfer learning (CRNNTL) method inspired by the applications of polyphonic sound detection and electrocardiogram classification. Our model takes advantage of both convolutional and recurrent neural networks for feature extraction, as well as the data augmentation method. According to QSAR modeling on 27 datasets, CRNNTL can outperform or compete with state-of-art methods in both drug and material properties. In addition, the performances on one isomers-based dataset indicate that its excellent performance results from the improved ability in global feature extraction when the ability of the local one is maintained. Then, the transfer learning results show that CRNNTL can overcome data scarcity when choosing relative source datasets. Finally, the high versatility of our model is shown by using different latent representations as inputs from other types of AEs.

A 7.07-kW Yb:YAG composite crystal zigzag slab amplifier at room temperature is developed and demonstrated, the optical-to-optical efficiency is about 29.5% and the slope efficiency is 39.5%, respectively, with the pumping power of 19.98 kW and the incident seed laser power of 1.18 kW. The influence of the pump coupling efficiency, the intensity of the seed laser, and the temperature to the extracted power from the Yb:YAG slab are analyzed by simulation. The fluorescence distribution at the end face and the transmission wavefront of the Yb:YAG slab are measured, the nonuniformity of the fluorescence distribution is less than 6%, and the P-V value of the transmission wave front distortion is less than 1.6 μm.

De novo molecular design has facilitated the exploration of large chemical space to accelerate drug discovery. Structure-based de novo method can overcome the data scarcity of active ligands by incorporating drug-target interaction into deep generative architectures. However, these strategies are bottlenecked by the small fraction of experimentally determined protein or complex structures. In addition, the cost of the molecular generation is computationally expensive due to 3D representations of both molecule and protein. Here, we demonstrate a widely used and fast protein sequence-based reinforcement learning (RL) model for drug discovery. In the generative model, one of the reward components, a binding affinity predictor, is based on 1D protein sequence and molecular SMILES. As a proof of concept, the RL model was utilized to design molecules for four targets. The generated compounds showed bioactivities by the validation of both QSAR and molecular docking with experimental 3D binding pockets. We also found that the performance of generated molecules depends on the selection of data source training for the binding predictor. Furthermore, drug design for a kinase without any experimental structure, CDK20, was studied. By only 1D protein sequence as input, the generated novel compounds showed favorable binding affinity based on the AlphaFold predicted structure.

Transformer models have been developed in molecular science with excellent performance in applications including quantitative structure-activity relationship (QSAR) and virtual screening (VS). Compared with other types of models, however, they are large and need voluminous data for training, which results in a high hardware requirement to abridge time for both training and inference processes. In this work, cross-layer parameter sharing (CLPS), and knowledge distillation (KD) are used to reduce the sizes of transformers in molecular science. Both methods not only have competitive QSAR predictive performance as compared to the original BERT model, but also are more parameter efficient. Furthermore, by integrating CLPS and KD into a two-state chemical network, we introduce a new deep lite chemical transformer model, DeLiCaTe. DeLiCaTe accomplishes 4× faster rate for training and inference, due to a 10- and 3-times reduction of the number of parameters and layers, respectively. Meanwhile, the integrated model achieves comparable performance in QSAR and VS, because of capturing general-domain (basic structure) and task-specific knowledge (specific property prediction). Moreover, we anticipate that the model compression strategy provides a pathway to the creation of effective generative transformer models for organic drugs and material design.

In recent years, the development of large language models (LLMs) has revolutionized various fields of natural science. However, their application in dealing with various molecular data remains constrained due to the reliance on single‐modality inputs and outputs. ChatMolData, a novel LLM‐based multimodal agent designed to handle diverse molecular data forms, including molecular databases, images, structure‐specific files, and unstructured and structured documents, is introduced. ChatMolData integrates the capabilities of LLMs (e.g., GPT‐4 and GPT‐3.5) with the robust toolset that supports data retrieval, structuring, prediction, visualization, and search tasks. The agent employs a systematic cycle of reasoning and action to efficiently process complex tasks in molecular science. The evaluation demonstrates that ChatMolData achieves over 90% accuracy for 128 diverse tasks, effectively bridging the gap between experimenters and computational tools. Moreover, it is anticipated that the multimodal‐agent strategy provides a pathway to expand data size and improve data accessibility, ultimately promoting molecular research and innovation.