molecular

MIT AI learns molecular language for rapid materials development and drug discovery

by
AI molecular properties concept illustration

MIT-Watson AI Labs’ new AI system dramatically simplifies drug and material discovery by accurately predicting molecular properties with minimal data. The system takes advantage of a molecular grammar learned through reinforcement learning to efficiently generate new molecules. This method has demonstrated remarkable effectiveness even with datasets of less than 100 samples. This AI system needs … Read more

Microsoft AI Research introduces a new deep learning framework called Distributional Graphormer (DiG) to predict the equilibrium distribution of molecular systems.

by
Microsoft AI Research introduces a new deep learning framework called Distributional Graphormer (DiG) to predict the equilibrium distribution of molecular systems.

https://arxiv.org/abs/2306.05445 The structure of a molecule determines its properties and functions. This is why structure prediction is an important problem in molecular science. Molecular scientists are appreciating the groundbreaking accuracy of deep learning approaches like AlphaFold and RoseTTAFold in identifying the most likely structures for proteins from their amino acid sequences. However, structural prediction can … Read more