Earlier this year, I wrote a number of posts here about IBM’s Watson project, to build a computer system that could compete on the popular TV game show, Jeopardy!. Now, according to an article at Technology Review, a system using some of the same technology in Watson is being deployed to assist in caring for infants in a neonatal ICU at a hospital in Toronto. Dr. Carolyn McGregor, an associate professor at the University of Ontario Institute of Technology, says that the system attempts to interpret and analyze a constant stream of information produced by monitoring equipment in the ICU.
McGregor leads a project that has developed software to ensure that no scrap of that data goes to waste. At the neonatal ICU of the Hospital for Sick Children in Toronto, that software, dubbed Artemis, collects data from eight infant beds. The system can monitor the baby’s electrocardiogram, heart rate, breathing rate, blood oxygen level, temperature, and blood pressure. It can also access data from medical records, such as the baby’s birth weight. McGregor and colleagues are developing algorithms that use those signals to spot signs of hospital-borne infection before doctors and nurses do.
The conventional method of monitoring the infants’ condition is to record data from the monitors periodically. This, according to the article, tends to a high false-positive rate in infection diagnosis; one might surmise that this stems from an excess of caution, because the medical staff knows that a short-term symptom might be missed.
“The processing paradigms we had before just didn’t fit with the kind of streaming data we are dealing with,” says McGregor. Software has traditionally performed analysis by systematically scouring a fixed, well-organized store of data, like a person navigating the stacks of a library, she explains.
The system uses an IBM technology called InfoSphere Streams, that allows parallel processing of several concurrent sources or “streams” of data. Just as the Watson system was able to launch many concurrent analytical algorithms to seek the answer for a single Jeopardy! clue, the Artemis system can deal with the parallel data streams produced by the range of medical monitors. Although the system is being tested on a small scale now, the hope is that the technology can be developed into a remote diagnostic resource for ICUs in many places.
IBM is also continuing work on other projects to deploy the lessons learned from Watson to medical diagnosis.
Although there will undoubtedly be many kinks to be worked out along the way, I have a feeling that systems like Artemis, and Watson, are going to become much more common in the near future. As I’ve noted before, there have been significant improvements in machine translation, stemming from the adoption of more empirically and statistically based approaches, and less emphasis on trying to specify a complete set of formal language rules in advance. Likewise, I suspect that we will begin to build what we might regard as truly intelligent systems, when they start working more like real brains seem to work.