Generative pretrained autoregressive transformer graph neural network applied to the analysis and discovery of novel proteins

RT @ProfBuehlerMIT: Predicting the properties of materials based on a flexible description of their structure, environment, or process, is…

Predicting the properties of materials based on a flexible description of their structure, environment, or process, is a long-standing challenge in multiscale modeling. Our MaterioFormer language model, trained to solve forward and inverse tasks, incorpora

The MaterioFormer language-model solves multimodal forward and inverse tasks through attention and graph strategies, demonstrated via design of mechanically stable, soluble, antimicrobial proteins. @ProfBuehlerMIT https://t.co/SYNWf9u5zb

Generative pretrained autoregressive transformer #graphneuralnetwork applied to the analysis and discovery of novel proteins | Journal of Applied Physics #ArtificialIntelligence #AI #ML #MachineLearning #DeepLearning #DL #Tech #Technology https://t.co/