Artificial Intelligence

IBM Watson Turning into Better-Rounded Genius

The "cognitive computing" at the heart of IBM's Watson is being adopted by a growing number of industries. IBMs house fair on the "World of Watson" gave its creators the chance to show which industries their AI is likely to conquer next.

23 Nov. 2016 Mark Schröder

World of Watson 2016

Universalgenie_Watson1
John Kelly III: Watson co-founder at the IBM house fair (Photo: IBM)

A little over four years ago IBM's Watson AI won the game show Jeopardy. For one of Watson's creators, John Kelly III, that victor of man over machine represented a turning point. From that day, 16 February 2011, the Director of IBM Research began believing in the breakthrough of "cognitive computing" using Watson technology. This was one of the insights revealed before a record-breaking 17,000 spectators at IBM's "World of Watson" exhibition in Las Vegas.

IBM has worked hard these past five years improving their AI technology. While Watson could previously understand, read and speak in English, the system is now multi-lingual — at least in terms of verbal interaction with the user. The AI can also now process images and detect patterns. Researchers in Kelly's team are currently working on giving Watson the ability to comprehend video. It's a clever strategy, market researchers say, since by 2020 over 80 percent of new online content will consist of films. Beyond this, augmented and virtual reality are both poised to break through to the mass market.

Watson for the iPhone and Facebook Messenger

Universalgenie_Watson2
Finally mobile: David Kenny wants to bring the IBM Watson to the iPhone. (Photo: IBM)

IBM has been working with Apple to bring Watson technology to the iPhone. The partnership was established two years ago and is now slated for expansion, Watson General Manager David Kenny of IBM told conference attendees. Together they have developed around 100 MobileFirst apps for iOS, now to receive Watson APIs. Developers then have the option of using Watson functions for their own applications. As such, the company is tying itself to one of the most widely distributed platforms anywhere, with over one billion Apple smartphones sold.

Facebook Messenger also has a billion users. Watson will be more present on that platform in the future as well. And in more than one place: "Watson ads" will allow customers to place their own ads within the chat software. The technology will then behave like a proper chat partner, answering questions about promotional topics. Weather is a perpetually popular topic, so one new bot will always have an answer ready, courtesy of IBM subsidiary The Weather Company. It also interacts with the user, if desired, says Cameron Clayton, CEO of The Weather Company.

Watson, Private Tutor

Pearson Education intends to expand digital learning platforms using Watson. The company is considered the world's largest publisher of instructional materials. IBM technology will on the one hand help test learning progress, and on the other allow for individualized preparation of instructional materials, says Tim Bozik, President of Global Product at Pearson. Students at both the school and university level will then receive materials with an emphasis clearly on the areas where they are weakest.

A digital learning assistant will also be able to explain the material — both to students and teachers preparing for class. Pearson hopes that its natural language solutions will also help mitigate the impact of teacher shortages, Bozik claims.

SAS right on IBM's tail

Universalgenie_Watson3
The competition isn't resting SAS founder Jim Goodnight has clearly had a change of heart about "Cognitive Computing." (Photo: IBM)

"Cognitive Computing" had previously been a term used solely by IBM. The company uses it to describe what Watson does. The system processes structured and unstructured data — primarily using advanced statistical methods. Watson accepts questions regarding that data in a user's natural language, either in writing or orally. Watson can also provide back answers verbally if desired. Big Blue sees it as a huge potential field and is investing billions in the technology. A lab was recently set up in Munich with 200 million dollars of equipment.

In many ways, IBM's Watson is without peer. When it comes to the analytics business, however, SAS is Big Blue's biggest competitor. The US company has been growing steadily since its founding in 1976. Big Blue by contrast reported falling sales this past mid-October — for the 17th straight quarter. Where the SAS business model is primarily focused on analytics solutions, IBM still maintains software for various industries, businesses and infrastructures, as well as a hardware portfolio. The latter areas are the ones running at a loss for years. Analytics, and Watson in particular, has been profitable.

SAS once called Watson little more than a "marketing campaign." That derogatory assessment came from no less than SAS founder Jim Goodnight himself. He knows of what he speaks. He also seems to know when it's better to hold his tongue. The market clearly wants "cognitive computing." SAS recently announced that it will also be delivering solutions for "cognitive computing" – to fight against cybercrime, for energy utilities and for price dumping on the retail market.

Analysts keeping quiet

Market analysts Gartner see two market leaders for Advanced Analytic Platforms: SAS and IBM. Compared with a year ago, SAS's lead has shrunk, with IBM almost caught up in the leader's quadrant. Critics have questioned why Watson technology isn't included in that analysis — at present the SPSS statistics solution is the only one included in Gartner's market overview. Watson appears to be classified as a 'special solution' by Gartner, one that doesn't fit into any existing category.

Others take a different approach: Forrester's current overview of Big Data Text Analytics Platforms sees IBM Watson as having a leg up. The Watson Explorer product is joined at the top of the market by SAS Contextual Analytics and Intelligence Platform from Clarabridge. The IBM solution claims to be one of the few applications on the market capable of processing data in Hadoop clusters, and as such can handle more than linguistic and statistical procedures.

Artificial Intelligence CEBIT RSS Feed