I suppose I'm thinking about transformer architecture rather than strictly GenAI, but the computer science aspects of protein folding and GenAI seem like they overlap significantly.
People are using these models to generate candidate molecules for specific purposes. According to some estimates I've seen, their hit rate is about 50% instead of 25-33%, and doesn't take two years.
I suppose I'm thinking about transformer architecture rather than strictly GenAI, but the computer science aspects of protein folding and GenAI seem like they overlap significantly.
People are using these models to generate candidate molecules for specific purposes. According to some estimates I've seen, their hit rate is about 50% instead of 25-33%, and doesn't take two years.