AI approaches like neutral networks work well when they have a lot of information to train on as a foundation for starting to learn patterns and see how things fit together. For novel diseases, targets or chemistry, however, there is often no training data available. For early stage drug discovery, this means that artificial intelligence approaches in isolation—and without the benefit of iterative cycles of wet lab assays—may not work well.
At TandemAI, our physics-based approach enables us to apply artificial intelligence to ingest data, perform computations, and create models which lead to insights. These insights produce increasingly accurate predictions our clients use to develop novel therapies—with the significant advantage of not needing to rely on pre-existing training data.