Yoshua Bengio, one of the pioneers in deep learning and artificial intelligence (AI), has developed new ways for computers to learn without much input from humans. The deepest learning has accelerated AI in the last five years changing the way how we use our smartphones and computers. It also pushes the new technologies like speech recognition and machine translation to next levels.
The world famous professor from the University of Montreal in Canada has come to the city to deliver a lecture on deep learning at the IIT Madras. In an exclusive chat with A. Ragu Raman from Deccan Chronicle, he talks about recent developments in the field of artificial intelligence. Excerpts:
Q Do you think computers are capable of making decisions in the absence of humans?
In general no, but for specific tasks where the computer is better than the human at that task then we should use the computer. We already do that.
Q In the future, how far the machines can be autonomous?
The machines have to be autonomous. A lot of our intelligence is due to our ability to learn in a very autonomous way. We are very far from achieving that. There is a lot of research to be done and a lot of researchers are interested in this direction.
Q Recently, Microsoft has unveiled a computing system which recognises the conversational speech as accurately as humans. In this direction what would be the next product?
Understanding speech here really means that the computer is translating the sound into the words. But, it doesn’t mean that it understands the words. This is where the research is right now. We have made some progress, but we are very far from having computers, which actually understand what do you mean so that the computer can answer the question or do things for you. We are moving in that direction. For example, If you look at Ok Google (Voice search) it is using some of that and very useful, but still very far away.
Q How much deeper learning accelerated artificial intelligence in the last five years?
All of the progress we have seen in the recent years is because of deep learning. Of course, combining deep learning with other things like reinforcement learning, robotics and many other technologies also contributed to the progress in artificial intelligence.
Q Can you tell some areas in which the artificial intelligence and Deep Learning research going to have the major impact?
Understanding the language better and being able to respond in national languages would be one obvious area. This will probably change a lot the way we interact with the computers. Another area which it can touch a lot of people lives is health. I think we will see a lot of uses of deep learning in health care.
Q Do the researchers in artificial intelligence try to recreate the human brain for the machines?
We are not trying to copy the human brain. We can take inspiration from what we know about the brain, because, we are trying to discover general principles of intelligence of learning. So, human brains and animal brains give us a lot of clues about how it could be. There is the big difference between planes and birds, but same underlying principles. We are trying to figure out these principles.
Q Google and Facebook are keenly investing in artificial intelligence research. Do you aware of their research areas?
They are working in many areas. One area that I know is machine translated and Google, in particular, has made a lot of progress. Facebook is trying to understand the language. They want to understand what you put on Facebook.
