Friday, December 11, 2015

Where is the boundary of IBM’s super-smart computer Watson? – IDG.se

Artificial intelligence has created two camps in the computer world. On the one hand, Elon Musk and Bill Gates and raises a warning finger portrays technology as the destruction of mankind. Other hand, sitting in an armchair in a hotel in Stockholm, in the form of Guru Banavar. One of the world’s leading researchers in artificial intelligence and responsible for IBM’s research into cognitive systems. The system known as Watson.

Guru Banavar
Guru Banavar, IBM, visit Stockholm. Photo: Sophia Nilsson


– There is such a hype about sci-fi, Hollywood makes a big deal out of it. And people are worried there is always more material for fiction. But that operates in the area I do not buy the hype. I would like to turn the question around and wonder if we can risk not investing in artificial intelligence, he says.

Read also: IBM makes it easier for companies to use Watson

2011 Watson won a race in the quiz show Jeopardy against two people. Watson had access to 200 million pages of data that took up four terabytes of memory and included the entire Wikipedia database.

 

 - The inspiration actually came from Jeopardy and games have historically been a source of inspiration for the research conducted at IBM. As a way to reach new ideas. The last five to ten years we have developed a whole toolbox of techniques of statistical and probability-character. The probability of something going in one direction or the other, he says.

 

 The supercomputer had learned to track the patterns of the language. See which words most often combined, if it were a noun or a verb. Question were then matched against the words and phrases in dictionaries as well as against databases. It can see how often a response has occurred and which response was probably the right thing. All based on statistics and probability.

 

 - Questions and answers processes, as this is common in many professions. You could say any area and I can find a way where Watson can be useful. As in health care, where we are currently testing the technology, says Guru Banavar.

 
 

IBM Watson
IBM Watson. Photo: Clock Ready (CC)



 Doctors and healthcare professionals are often very stressful days. Many decisions must be made quickly and the physician must quickly gain a picture of the patient’s background but also keep abreast with current research. With the help of Watson, the doctor may ask questions about the symptoms, which clinical studies in this area or find similar cases.

 
Also read: IBM’s supercomputer is already the most part. What happens when Watson gets even smarter?

– I want to make it quite clear that Watson does not set out a definitive diagnosis, it gives options. The doctor has to evaluate and make their assessment. There are many factors that can not be digitized, which the doctor knows that lifestyle choices and social conditions.

 

 Guru Banavar see that there are problems with relying too much on cognitive systems. There are always factors that are not computer can calculate. The solution the researchers presented was that Watson report how secure it is the answer and can come with several options. It says what it does not know, that might only be 75 percent sure. The entry will be a call to the doctor to self-assess. Man and machine complement each other.

 

 For it is something Watson will never make as it is to try to be a man. Where are the limits of supercomputer. It will not, as in the films dystopian world, act on their own initiative.

 

 - There is something called a general artificial intelligence, designed to resemble a human. It will not Watson do. People and machines complement each other, it learns to interact with natural language. But do not be human, says Guru Banavar.

 

 He believes that we can not afford not to invest in cognitive systems. Problems of the world and humanity is facing are so large that they can not be solved without computers Watson.

 

 - We pay a high price because we do not know how to take care of our environment, or our children, and we do not know how to cure a patient. And we do that because we can not understand all the knowledge available to solve the problems. There are areas we need to focus on. Everything else hajpas takes the focus from the important issues, he says.

 
Also read: It is now possible to build smarter applications with IBM’s Watson

Today technology which scientists built into Watson in the cloud, where anyone can build apps on the platform. It has been IBM’s way to get the cognitive systems to a wider audience. In the current situation there are apps such as health and training.

 

 - Watson makes the app understands and gathers data from reports and blogs, and turn them into advice for the user, says Guru Banavar.

 

 He also believes that this type of smart computers will serve as assistants in the future. Pharmaceutical Scientists can use the platform to sift large amounts of data and ask for new studies. Journalists to ask questions in the big study materials.

 

 - My vision is that all knowledge workers have a cognitive assistant that helps you. That is how I think of the future experts will look like. We force people to learn things by heart, we let people do many things that machines can do. It is not productive. People do what people are good at.

 

 People will ask questions, defining goals. The machine can build models, sort the data, perform calculations.

 

 - Together they can be better than each one separately to solve a problem.

LikeTweet

No comments:

Post a Comment