Noetik Uses Autoregressive Transformers to Address 95% Failure Rate in Cancer Clinical Trials

Summary

Noetik, a biotech AI company co-founded by Ron Alfa and Daniel Bear, is applying autoregressive transformer architectures to the problem of cancer clinical trial failures. Their model, TARIO-2, aims to reframe the 95% failure rate of cancer treatments in clinical trials as a patient-treatment matching problem rather than a drug efficacy problem. The approach uses transformer-based models trained on biological and clinical data to better predict which patients will respond to specific treatments.