At Micron’s Insight 2018 event in San Francisco last October, I spoke about memory’s role in enabling a future where our lives will be enriched through artificial intelligence (AI). I have a strong interest in the role of memory and machine learning in enabling precision healthcare as both the parent of a cancer survivor and Micron’s Vice President of Technology Strategy. From everything I’ve read, I believe that memory innovations will help revolutionize healthcare, especially when it comes to healthcare data analytics at the edge.
The paradigm shift that comes with applying AI in healthcare will enable clinicians to have real-time diagnostic capabilities to give patients immediate feedback — and with one visit to the doctor, understand the diagnosis, prognosis and treatment plan. Consider that millions of routine abdominal scans are taken every year. What if deep learning can be used effectively on those routine scans to provide early, highly accurate computer-aided detection of abnormal tissue? For some terrible diseases, like pancreatic cancer, early detection is the difference between life and death.
Now, take this capability and put it in the hands of rural doctors who are far from big city hospitals, and the opportunity to save lives increases exponentially. In the not-too-distant future, it is easy to imagine a proliferation of more readily accessible diagnostic devices that generate enormous data sets being analyzed and used for better predictive diagnosis and treatment. Even today, personal devices have the potential to monitor heart rates, blood pressure and act like a personal electrocardiogram. Possibly even more intriguing are the start-up companies working to develop affordable, automated ultrasound imaging platforms for self-monitoring. These platforms will use deep learning to differentiate benign and malignant lesions for early cancer detection. Imagine that for individuals with family histories of cancer, these devices could become as prevalent as having a blood pressure monitor in your home. This is precision healthcare at the edge.
The key to using this technology to solve our most difficult challenges is to get from data to insight faster. Data analytics at the edge require high-performance memory systems that deliver speed of data transfer and high-density storage. If we imagine a future where massive amounts of data can be collected more easily from local sources, we can imagine how the data can be used to tailor treatments for specific diseases. But, if data can’t be analyzed quickly enough, its value, particularly in healthcare, is reduced. In fact, storage at the edge will become more and more important because local data will help enhance the speed of data processing and analysis. Today, much of the analysis is done in the cloud, on servers far removed from the data source. In a future where we gather data locally in clinical or even non-clinical environments, our ability to process, transfer, integrate and store this data with other data sets will be a key enabler to delivering insights quickly and affordably to clinicians and patients.
Micron’s ability to innovate and create memory and storage solutions that meet these needs for fast, powerful and economical data processing will pave the way for the future. We have teams working on these problems at Micron today: business units working with customers to deliver fast data for AI applications, researchers working with leading universities and government agencies to uncover new methods and uses for our memory as well as engineers looking at future compute architectures. This pathfinding is very important to understanding how Micron will add value to the global AI ecosystem of the future, and how we can help usher in innovations in personalized medicine and other AI applications. I’m excited to be part of this exciting new future where we can truly enrich life on the planet.
