AI Talk: Lung cancer, ear infection, clinical trials and fake news

May 24, 2019 / By V. “Juggy” Jagannathan, PhD

This week’s AI Talk…

Lung cancer detection

I saw this article in one of my usual haunts—MIT Technology Review! Using a corpus of CT Scans licensed from Northwestern, Google researcher Danial Tse trained a deep learning model to detect malignant lung nodules. The models did better than radiologists in identifying the nodules, with 11 percent fewer false positives and 5 percent fewer false negatives. This technology is one of a string of models that are able to read all kinds of images. It will be interesting to see when these types of models make it into the practical world.

Smartphone App to diagnose ear infection in children

It’s been a while since I had to deal with the pernicious ear infection which afflicts children. Eighty percent of children will experience an ear infection during their childhood and many parents have agonized over it! Well, parents have a new tool to determine if their kids are suffering from an ear infection. It’s a smart phone app that emits a bird-chirping sound. You simply use a folded piece of paper in the form of a funnel to pipe this sound into the child’s ear and based on the sound emitted back, the machine learning algorithm will predict if there is fluid in the ear—a sign of infection. The algorithm is correct 80 percent of the time, apparently!

Verily and Big Pharma

Verily, an Alphabet company, was created to focus on life-sciences. They are involved in a number of projects, ranging from collecting evidence directly from patients for studies to supporting the precision medicine initiative from NIH—the All of Us project. Now, they have just announced partnership with four big pharma companies: Novartis, Sanofi, Otsuka and Pfizer. The partnership aims to identify patients who are eligible to participate in clinical studies conducted by the drug manufacturers. One way to do this is to use Google search capability. If someone does a search for say, “asthma” and there is a trial drug for that condition, an advertisement pops up for them to consider joining the research trial. Targeted advertisement of research trials could mean more participation and lead to the development of new and better pharmaceuticals.

Cracking down on fake news

Whatever you may think of Singapore, they are getting ahead of the fake news world (albeit using draconian tactics). If you are caught posting fake news, you can be charged with ten years of jail time or fined one million Singapore dollars (US $730K). The law dubbed “The Protection from Online Falsehoods and Manipulation Bill” has stirred up a bit of controversy, but Singapore is known for its strict laws and unlikely to back down from its stance.

I am always looking for feedback and if you would like me to cover a story, please let me know. “See something, say something”!  Leave me a comment below or ask a question on my blogger profile page.

V. “Juggy” Jagannathan, PhD, is Vice President of Research for M*Modal, with four decades of experience in AI and Computer Science research.