Artificial Intelligence in Drug Discovery
Using artificial intelligence for drug discovery is the practice of using computational methods to research new pharmaceuticals, and repurpose existing compounds for new use cases.
Computational Drug Discovery Using AI
Artificial intelligence (AI), and more accurately Machine Learning (ML), is accelerating drug discovery and development timelines, and simultaneously reducing costs associated with pharmaceutical research.
BioSymetrics uses a patent pending Contingent-AI™ platform, that trains machine learning models to reduce bias by making decisions “contingent” on early stage data collection and feature engineering variables. The result provides a higher level of confidence in the output of the model, leading to greater precision.
How Contingent-AI™ Works
Artificial Intelligence (AI) or Machine Learning (ML) models are commonly run against “feature sets”. The data scientists’ decision on how to create feature sets can dramatically affect the outcome of the machine learning by unintentionally inferring bias. Contingent-AI™ seeks to train and run machine learning models against various permutations of feature sets in an effort to reduce bias and de-noise the machine learning output.
For example, imaging data can vary greatly by the contrast settings set by the technician when an MRI was captured. Contingent-AI will evaluate different variations of the imaging feature set with image contrast at different settings, reporting back if results are “contingent” on a decision made in early stages of data collection or feature engineering. BioSymetrics output reports will showcase the contingency of a machine learning model’s results based on variances in the feature engineering stage and data de-bias stages.