Facebook Employs AI Tech To Combat Complex Illnesses

Facebook said it is rolling out artificial intelligence (AI) tools aimed at predicting the best combination of pharmaceuticals to combat complex illnesses, VentureBeat reported.

The social media behemoth said it is launching the first AI solution that is able to predict the effects of drug combinations and the timing of doses for the best outcome. The tool will also suggest additional interventions that might help a person’s condition, such as gene deletion.

The tools are being launched in collaboration with Helmholtz Zentrum München, and Facebook said it thinks the solutions will be a boon to the identification of drug interactions and will be able to offer up the best combination of therapies and medications to combat complex diseases.

Repurposing existing pharmaceuticals has already changed the way doctors and clinicians prescribe drugs and the different combinations have been found to fight deadly conditions that were not indicated by the suggested use of a particular drug. Further, a personalized therapeutic approach could produce more effective modalities for patients, Facebook said, per VentureBeat.

Facebook is using an open-source model — Compositional Perturbation Autoencoder (CPA) — that is thought to address the issue while also employing a self-supervision technique to predict the most favorable outcomes. 

The new approach differs from supervised models that collect information from labeled datasets. Facebook said its approach will reveal the relationships between the varying parts of the equation, to the point of being on par with human-level intelligence.

The telehealth market is anticipated to escalate more than 20 percent from now until 2028, and globally, the market is expected to top 2020 estimates of $56 billion. Verizon introduced a pilot of BlueJeans video conferencing for a range of medical functions, including telehealth.

Verizon’s BlueJeans Telehealth initiative is expected to enable patients to use time before a medical appointment to fill out forms and questionnaires. The model will also extend library resources that a provider can offer patients.