Artificial intelligence (AI) is fast on its way to being part of routine practice in health care thanks to companies like IBM. Now, General Electric is throwing their hat in the ring with a recently announced partnership with the University of California San Francisco.
We’ve already talked about the benefits of IBM’s Watson, Google’s DeepMind and medical apps that could help with diagnosis, treatment and population monitoring. The collaboration with UCSF positions GE as another viable competitor in the growing health AI market. As with those other ventures, the goal is to mine the massive amounts of health data being centralized in electronic health records, health trackers, and other repositories. Not only can AI help generate insights retrospectively, it can also help move us to a learning healthcare system where new data is constantly being integrated into, say, predictive models for drug benefits in a particular disease.
In this case, GE will utilize its analytical algorithms to give providers new insight into their patients through deep learning — a way a computer can process lots of data, including imaging and text — in “ways that traditional machine learning methods cannot.”
The initial focus on the collaboration will be on “high impact, high volume imaging.” That makes a lot of sense given that GE makes X-ray systems, MRI scanners, and everything in between. As an example, they describe development of an algorithm focused on chest imaging, like identifying a patient with a collapsed lung.
Using these algorithms, machines could be “taught” to notice abnormalities, helping radiologists focus their attention on the patients who may be in need for rapid intervention. And over time, using deep learning and other forms of AI, these algorithms would likely get more accurate and faster over time.
As Dr. Michael Blum, professor at UCSF, put it:
Next generation data science techniques have already transformed the industrial and consumer world. With this collaboration, these technologies will be applied to our clinical data and images to provide clinicians with actionable information in near realtime [sic]. Together, we will develop tools and algorithms that will allow clinicians and researchers to identify problems and ask questions that are only achievable with vast computing power and datasets.
Source: Press Release