The optimized multi-head graph attention network (OPNet) uses the multi-head graph attention network to predict both the thermal decomposition temperature with a 5% mass loss (Td5%) and the glass transition temperature (Tg) of polymers as a machine learning model. The OPNet model provides reliable performance predictions for the Td5% and Tg datasets. The OPNet regression evaluation metrics R2 of Td5% is 0.76, which is the best, and the regression evaluation metrics R2 of Tg is 0.91, which is better than the current existing method models (the best model R2 ≈ 0.90). The OPNet model is an end-to-end network model, eliminating the need for manual data filtering or feature extraction. By analyzing the feature weights of the OPNet model, it is found that structures such as benzene rings play a more important role. At the same time, we have confirmed through other literature that such structures do have better stability and a higher thermal decomposition temperature and glass transition temperature. Therefore, the OPNet model exhibits interpretability and holds significant reference value for the field of materials science.
Deep generative models are becoming a tool of choice for exploring the molecular space. One important application area of deep generative models is the reverse design of drug compounds for given attributes (solubility, ease of synthesis, etc.). Although there are many generative models, these models cannot generate specific intervals of attributes. This paper proposes a AC-ModNet model that effectively combines VAE with AC-GAN to generate molecular structures in specific attribute intervals. The AC-ModNet is trained and evaluated using the open 250K ZINC dataset. In comparison with related models, our method performs best in the FCD and Frag model evaluation indicators. Moreover, we prove the AC-ModNet created molecules have potential application value in drug design by comparing and analyzing them with medical records in the PubChem database. The results of this paper will provide a new method for machine learning drug reverse design.