Enabling AI-Driven Healthcare Advances Without Sacrificing Patient Privacy | MIT News


There is a lot of excitement at the intersection of artificial intelligence and healthcare. AI has already been used to improve treatment and detection of disease, discover promising new drugs, identify links between genes and disease, and more.

By analyzing large datasets and finding patterns, virtually any new algorithm has the potential to help patients – AI researchers just need access to the right data to train and test these algorithms. Hospitals, naturally, are reluctant to share sensitive patient information with research teams. When sharing data, it’s hard to verify that researchers are only using the data they need and deleting it when they’re done.

Secure AI Labs (SAIL) solves these problems with technology that allows AI algorithms to run on encrypted datasets that never leave the data owner’s system. Healthcare organizations can control how their datasets are used, while researchers can protect the privacy of their models and research queries. Neither party needs to see the data or the model to collaborate.

SAIL’s platform can also combine data from multiple sources, creating rich information that powers more efficient algorithms.

“You shouldn’t have to chat with hospital executives for five years before you can run your machine learning algorithm,” says SAIL co-founder and MIT professor Manolis Kellis, who co-founded the company with CEO Anne Kim ’16, SM ’17. “Our goal is to help patients, help scientists with machine learning and create new therapies. We want new algorithms – the best algorithms – to be applied to the largest dataset possible. “

SAIL has already partnered with hospitals and life science companies to unlock anonymized data for researchers. Next year, the company hopes to work with around half of the top 50 academic medical centers in the country.

Unlocking the full potential of AI

As an undergraduate student at MIT studying computer science and molecular biology, Kim worked with researchers at the Laboratory for Computing and Artificial Intelligence (CSAIL) to analyze data from clinical trials, studies combination of genes, hospital intensive care units, etc.

“I realized that there was something seriously wrong with data sharing, whether it was hospitals using hard drives, an old file transfer protocol, or even sending messages. things in the mail, ”Kim explains. “Everything was just not well followed. “

Kellis, who is also a fellow of the Broad Institute at MIT and Harvard, has spent years partnering with hospitals and consortia for a variety of illnesses, including cancer, heart disease, schizophrenia, and obesity. He knew that small research teams would struggle to access the same data his lab was working with.

In 2017, Kellis and Kim decided to commercialize the technology they were developing to allow AI algorithms to run on encrypted data.

In the summer of 2018, Kim participated in the delta v startup accelerator managed by the Martin Trust Center for MIT Entrepreneurship. The founders also received support from the Sandbox Innovation Fund and the Venture Mentoring Service, and established various early connections through their MIT network.

To participate in the SAIL program, hospitals and other healthcare organizations make some of their data available to researchers by installing a node behind their firewalls. SAIL then sends encrypted algorithms to the servers where the datasets reside in a process called federated learning. The algorithms analyze the data locally in each server and feed the results to a central model, which updates itself. No one, not the researchers, not the data owners, not even SAIL, has access to the models or the datasets.

The approach allows a much larger set of researchers to apply their models to large datasets. To further engage the research community, Kellis’ lab at MIT has started hosting competitions in which he provides access to datasets in areas such as protein function and gene expression, and puts researchers the challenge of predicting outcomes.

“We invite machine learning researchers to come and train on last year’s data and predict this year’s data,” says Kellis. “If we find that there is a new kind of algorithm that works best in these community-level assessments, people can adopt it locally in many different institutions and level the playing field. So the only thing that is the quality of your algorithm rather than the strength of your connections matters.

By enabling large numbers of data sets to be anonymized into aggregated information, SAIL’s technology also enables researchers to study rare diseases, in which small pools of relevant patient data are often spread across many institutions. . This has historically made the data difficult to apply to AI models.

“We hope that all of these datasets will eventually be opened,” says Kellis. “We can break through all silos and usher in a new era where every patient with every rare disease around the world can come together with one keystroke to analyze data. “

Enabling the medicine of the future

To work with large amounts of data on specific diseases, SAIL has increasingly sought to partner with patient associations and consortia of healthcare groups, including an international healthcare consulting firm and the Kidney Cancer Association. The partnerships also align SAIL with patients, the group they try to help the most.

Overall, the founders are happy to see SAIL solve the problems they face in their labs for researchers around the world.

“The right place to solve this problem is not an academic project. The right place to solve this problem is in industry, where we can provide a platform not just for my lab, but for any researcher, ”says Kellis. “It’s about creating an ecosystem of universities, researchers, pharmaceutical, biotechnology and hospital partners. I think it is the mixture of all these different fields that will make this vision of the medicine of the future a reality.


About Author

Comments are closed.