Watson in healthcare: HIMSS keynote
Ginni Rometty shares learnings from 5 years of deployments
The Healthcare Information and Management Systems Society (HIMSS) annual event is the biggest of its kind covering all aspects of healthcare and technology.
In her keynote at the event, IBM CEO Ginni Rometty explained why the health industry is so foundational to IBM’s strategy and how cognitive computing brings within reach ‘moonshots’ like curing cancer and other major illnesses we face. Over the five years that IBM Watson has been working in the health industry (on projects like that with Memorial Sloan Kettering and Manipal Hospitals in the oncology space), there have been a number of learnings, which Rometty shared:
1) Success requires a range of cognitive services
The systems now being built incorporate a number of services from natural language processing to visual recognition to deep learning, to name just a few… longer list on the Watson Developer Cloud.
2) Transparency is key
It’s important to share with the client who trained the system (including whether machine learning is used for the system to train itself), what data does it use and how much confidence is there in the insights.
3) Systems are domain-specific
As opposed to general AI, we’ve seen more success by having systems that are vertically trained. So for healthcare, Watson has been focused on that domain (eg. fed the 1,000s of new oncology research papers published daily) and has been trained by healthcare physicians.
4) Systems are cloud-based
A cloud architecture is necessary to allow the systems to be ubiquitous and deliver maximum impact. A cloud needs to be built that can handle big data, and given the sensitive nature of our healthcare information, the system needs to have a high level of security. On this note, Rometty points out that data and analytics have a pivotal role to play in spotting the patterns that surround security leaks. These also need to be hybrid cloud environments spanning organizations and potentially the public cloud, so the right data is shared at the right time with the ecosystem, but when data needs to be retained, it remains within the firewall.
5) An open platform is required
To achieve innovation, it’s necessary to share information across the ecosystem. The system should be open and interoperable with other systems to derive maximum value. Using open standards makes it easier to bring in partners and build an ecosystem more effective than the sum of its parts.
You can see the full keynote and an interview by HIMSS CEO Stephen Lieber in this Youtube video:
To see some of these solutions in action:
- Atrius Health is using Watson to help patients and physicians devise joint health plans based on data
- Barrow Neurological Institute used Watson to find genetic links to the deadly ALS disease
- Jupiter Medical Center adopts Watson for Oncology
- Illumina uses IBM Watson in DNA tests for cancer