January 27, 2015 | Written by: Burak Cakil
Share this post:
In 2011, a cognitive system named after the founder and CEO of IBM, Thomas J. Watson, achieved an epic win against two human champions in a competition called Jeopardy! This was a vital technology that would have a significant importance in the future. Yes, we are talking about IBM Watson, the system that is able to answer people’s questions, not only by returning search results but also by analyzing them using hundreds of algorithms at the same time and providing evidence that the final answer is correct. Not to mention that it does all of these in almost three seconds!
After its success on Jeopardy!, IBM research worked to make this technology open for public use. Later, IBM established a separate business unit for Watson called The Watson Group and a dedicated workforce to continuously improve Watson’s capabilities. The aim is to bring the power of Watson and cognitive computing to market using cloud delivery models. And their efforts have paid off, since they recently announced some of the Watson capabilities to be available on IBM Bluemix. These are available “as a service,” meaning you can use them in your own applications and services, embed them anywhere using Watson APIs such as Watson Question and Answer API and enhance your application capabilities dramatically. This also means that soon you might be able to do all the magic behind the Jeopardy! challenge within your application, with just a couple of clicks!
I was pretty amazed when I first heard this news, so the first thing I did was open the Bluemix dashboard and start testing those services. At the time of writing, there are seven services, all in beta, and there are sample applications for most of them. So what can they do? Here they are:
• Concept Expansion: Can analyze text and interpret the meaning of some words that are used for different contexts. “The Big Apple” for “New York City” is an example.
• Language Identification: Detects the language in which a particular text is written. Simple.
• Machine Translation: Translates from one language to another. Note that you can do it in your own application using this service.
• Message Resonance: Analyzes a text or speech and measures how well it would be received by a particular audience.
• Question and Answer: Can provide answers for your questions in natural language.
• Relationship Extraction: Can analyze and map the relationships between the components of a sentence; can be used to make some other analysis services’ life easier.
• User Modeling: Computes the characteristic attributes of a person from his or her personal writing such as a blog post, tweet and so on.
All of these are useful ways to expose cognitive services. So a valid question at this point would be, “What can I do with Watson?” It’s a technology that can learn, compare, analyze, reason, justify and process human language, so there must be countless ways to create benefit from it. A nice example called Watson Films was created using the Question and Answer service and some other components. It can answer your questions about films and make suggestions based on your preferences. Another example would be a travel Question and Answer system that can answer your questions related to travel over Twitter.
Of course, what you can do is not limited to these examples. Actually, the boundaries for the capabilities of Watson are the data you have provided. This data is called a corpus, and it can be unstructured data, such as text, HTML or PDF files. The more you feed Watson, the more you will get from it. It can be trained with the provided data, and it can also receive feedback for its answers. Training and feedback are key criteria for the accuracy of the results.
In summary, it is not hard to see that cognitive systems will be the next generation of computing. This technology, if put to good use, will probably change many of the ways we do, work on and interact with things. This also brings a great opportunity to innovate your solutions, products, services and life.
If you are interested in Watson, drop a comment or connect with me at @cakilburak on Twitter.