Steve Gold came up to talk about how IBM Watson. Watson, of course, started with the focus on analyzing unstructured data to present hypotheses and learn from the results. Watson brought together lots of IBM projects and technology and Watson Analytics is similar – a combination of various technologies and projects to solve a new problem. This is happening in the context of a tidal wave of data, both structured and unstructured. The power of Watson Analytics and Big Data in combination is to turn this tidal wave of data into something more actionable he says.
Steve repeated some of the standard analytics talking points
- 1 in 2 CEOs say they don’t have the data they need
- It’s hard to use data to change decision-making in a timely fashion – 1 in 5 patients are misdiagnosed for instance
- Firms estimate that they analyze only 12% of the data they already have
As well as dealing with this data challenge, Watson focuses on using cognitive computing to change the way expertise is being used. IBM sees this as a new era in computing, replacing the generation of programmable computers with a generation of cognitive ones. Two key elements:
- Harnessing the expertise embedded in recordings, documents, publications, blogs and more into something usable. Teaching systems rather than programming them and getting progressively smarter and learning faster as the knowledge base grows.
- Probabilistic not deterministic, identifying multiple possible outcomes and weighting them rather than driving to a single outcome. Multiple outcomes based on contextual meaning or because there are various possible solutions to a problem.
Watson cognitive computing acts very like humans do:
- It observes and consumes very large amounts of information of lots of different types in a board context.
- It interprets all these observations to find connections and frame choices or options.
- It forms and then evaluates various hypotheses to see how they might work
- It decides what to recommend with weights and suggestions.
IBM sees Watson scaling the expertise that companies are acquiring, bringing it together in a more systematic way. It learns faster too, accelerating the speed at which this expertise can be put into practice.
Watson Analytics is built on top of the modular Watson architecture, alongside Watson Explorer for instance. These two Watson products (and future ones) allow the development of Watson Advisers and Watson Industry focused solutions. Watson provides a core set of capabilities to Watson Analytics that can be combined with other analytic technologies to deliver a solution.