Watson on Watson

My namesake Thomas J Watson, IBM’s first CEO and prominent industrialist is oft quoted (perhaps incorrectly) as stating “I think there is a world market for maybe five computers” in 1943. There is scant evidence he made this exact claim but it’s somewhat interesting that in 2016 his name has been given to the most advanced technology platform in the world, IBM Watson. Although made up of many thousands of physical pieces of hardware it acts as a single ‘machine’; a platform that promises to reveal insights in data at a scale unparalleled
in the modern world.

What is Watson?

“IBM Watson is a technology platform that uses natural language processing and machine learning to reveal insights from large amount of unstructured data.” (IBM Web site). In simple terms the platform has access to many sources of data; both historical and real-time; and provides a set of tools and methods to allow applications to be built to address both old and new problems in a new way. The platform processes all of the data and builds an understanding of it in an intelligent way; to reveal insights, patterns and relationships. These insights provide the end customer with answers to the questions they want to ask of the data that previously would be highly cumbersome and difficult to extract. IBM Watson is the platform that provides the building blocks for partners to build new disruptive products. It processes at scale to understand many sources of information in the same way a human would but no human could manage; and can then be asked questions in natural language and give the answer!

Jeopardy

To highlight its abilities and to demonstrate a level of ‘intelligence’ IBM Watson took on 2 of the best humans in the game of Jeopardy, the famous TV quiz show in the US. The Watson system had over 200 million pages of information from a variety of sources given to it to provide the basis of knowledge to make it competitive in a live game scenario. These huge volumes of unstructured data required analysis and processing by the machine learning algorithms within Watson. The quirk of Jeopardy is that the contestant needs to provide the ‘question’ being asked in response to the ‘clue’ provided so the Watson system needed to not just analyse the data it held but to understand the clue and its ‘natural language’. In training the system it took 20 researchers 3 years to get the ‘machine’ competitive enough to win; the tweaking of algorithms and weightings of responses was a mammoth task in 2011. However Watson went on to record an outstanding victory against the human contestants; winning $77,147 vs $21,600 and $24,000.

2016 – Machine Learning has come of age

Roll forwards to present day and the methods and processes that helped Watson win Jeopardy have been refined and advanced upon to allow Watson to address any number of industries and problems. Through its Watson Ecosystem of partners the platform is being harnessed to build smart data driven machine learning products that provide answers to challenges humans could not answer in a timely fashion. According to IBM “The goal is to have computers start to interact in natural human terms across a range of applications and processes, understanding the questions that humans ask and providing answers that humans can understand and justify." Watson is being used to create applications for information intensive industries such as healthcare, telecommunications, financial services and in government. With significant advances in hardware, processing capability and cloud computing the platform is capable of tackling challenges previously it couldn’t. The myriad of use cases is expanding all the time as partners and customers understand the potential for machine learning solutions to drive competitive advantage out of the ever expanding amounts of data globally.

Thomas J Watson may have been incorrectly quoted; but his surname (and mine!) is attributed to one ‘machine’ that could provide the world with answers that 5 years ago could only be dreamt about. Machine learning solutions are going to revolutionize much of our world in 2016 and beyond!

At Chatterbox we have developed an advanced machine learning platform that uses NLP statistical machine learning methods and processes. If you would like to know more please visit our web site www.chatterbox.co or contact our VP Strategic Alliances, Andrew Watson Andrew@chatterbox.co