Active Silicon AI Series part 9: IBM’s Watson – what’s it all about, and why is it watching the tennis?

  • Published July 2018
Active Silicon AI series part 9 - IBM Watson, Wimbledon and health imaging

IBM’s Watson is a cognitive computing capability which, using DeepQA software and the Apache UIMA framework, analyzes very high volumes of data and learns from previous conclusions and data processing to support decision-making. Put simply, it’s a question answering guru powered by AI, able to process data at 80 TeraFLOPs, effectively digesting a million books per second[1]. Watson first made the headlines when it was pitted against human opponents in the US TV quiz show Jeopardy, and beat the reigning champions. Following on from Deep Blue famously beating chess wizard Garry Kasparov in 1997, you might be forgiven for thinking that IBM are just in this for fun, but applications for Watson are far-reaching and promising to change the computational world as we know it.

Watson at Wimbledon
Computing and tennis have come a very long way since Atari’s Pong in 1972, and now AI is changing our experience of the world’s most famous grass court tennis tournament. IBM have partnered with the organizing committee to drive fan engagement and deliver faster, more relevant and more captivating viewing content. By recognising players’ movements and emotions, combined with crowd noise and match data, Watson can identify which moments in a match are the most exciting and is able to compile a highlights montage in just minutes as opposed to the hours that it takes a human editor. Increased, exclusive content on social media platforms and a chatbot called Fred are just some of the other ways in which Watson is growing audience participation and enjoyment. IBM cited a 25% increase in Wimbledon’s social media following in 2016-17 and expects to see this pattern continue.

Watson in medical imaging
Beyond sport, Watson is having a huge impact on the medical sector too. Watson Health was established in 2015 to focus solely on data generated within the medical sector; the platform is also open to developers to use in their own applications. In 2016, IBM announced its Watson Health Imaging division, combining the expertise of academic medical centres, health systems, ambulatory radiation providers and image technology companies. The collaborative “aims to bring cognitive imaging into daily practice to help doctors address breast, lung, and other cancers; diabetes; eye health; brain disease; and heart disease and related conditions, such as stroke.”[2]  Since its inception, the number of members has grown from 15 to 24, and we can now see Watson being trained to aid in understanding how a condition is likely to progress, what treatment should be considered, and when to intervene. Watson reads patients’ records, doctors’ reports and other peripheral material, and combines these with medical images to make its diagnoses and predictions.

Also in 2015, IBM bought the health solution provider, Merge, and used the company’s vast collection of medical images to train Watson’s visual recognition capabilities to identify anomalies and changes in patient scans.

And the winner is…
Making interesting further reading are the entries for the IBM Watson AI XPRIZE – a $5 million AI and cognitive computing competition running over four years from 2016-2020. From the initial 147 entrants, the 62 teams progressing to round two were announced in December last year. Details of the projects, and the top 10 selected at this point for a Milestone award can be seen here.

Don’t forget the football
Collaborating with FOX Sports, Watson is bringing its highlighting-creating AI tools to football too. It’s even been put to work analysing massive amounts of data to predict the winner of this year’s FIFA world cup. But we wouldn’t want to spoil the surprise….

Read more in our AI Series and sign up for our industry newsletter here to keep abreast of the latest vision news.

[1] https://en.wikipedia.org/wiki/Watson_(computer)

[2] https://www-03.ibm.com/press/us/en/pressrelease/49985.wss