Hinton is a retired distinguished professor of computer science at the Faculty of Arts & Sciences Department of the University of Toronto. He has been studying artificial neural networks since the 1970s. His goal is to create machines that can think and learn by modeling the structure of the human brain. At that time, most researchers still rejected the artificial intelligence method of the neural network, but Hinton and his team persisted, and later achieved great success. Over the past decade, their deep learning neural networks have surpassed traditional artificial intelligence methods on almost every benchmark. In 2013, Google acquired Hinton’s neural network startup DNN research. Recently, he has also been nominated as one of the “Connected†2016 Global Top 100 influential figures. His learning machine has been proven to have great practical value. These technologies can make a self-driving car safer, help us to translate another language effortlessly, and will gradually take over human work and even labor in our work and family. Their ability to discover patterns in huge data sets will also help us improve genetic medicine and develop new treatments for disease. Hinton's wish is simple: "I want to understand how the brain calculates." Although Hinton’s research has had a huge impact on the systems that billions of people use every day, the neural network revolution has only actually begun. Hinton recently accepted an interview with Jennifer Robinson, a writer at the University of Toronto, about his career in artificial intelligence and his future expectations for this erupting field. Q: In your current study, what do you think is most exciting to you? In the entire field of artificial intelligence? A: Deep neural networks have performed very well on some important tasks such as speech recognition, image interpretation, and machine translation. As our computers get faster and data sets get bigger, I'm convinced that such rapid progress will continue. But I don't think the type of artificial neural network we have developed so far is the best. There may also be better types of artificial neural networks that can learn from far less data and provide more insight into the way the true brain thinks. Finding new types of neural networks is the most exciting direction I feel now. Q: Why is the brain the best model for creating artificial intelligence? How much time does it take for the machine's computing power to reach or exceed the level of the brain? Or does Google's AlphaGo already prove that the human brain has been exceeded? A: Until recently, the brain was much better than any computer in interpreting images or understanding natural language. So if we ignore the incredible computing power of the brain, it seems very stupid. Recently, the artificial neural network inspired by our understanding of the brain and greatly narrowing the performance gap between humans and machines, I think, this seems to have proved that the idea of ​​drawing inspiration from the brain is correct. I think computers will eventually be able to surpass the capabilities of the human brain, but transcending different abilities will also require different lengths of time. It may take a long time for computers to understand poetry, jokes or satire as well as humans. Fiber Optic Fast Connector Sijee Optical Communication Technology Co.,Ltd , https://www.sijee-optical.com
Artificial intelligence will eventually surpass the human brain but it is difficult to understand jokes
Geoffrey Hinton of the University of Toronto is one of the world’s leading computer scientists and Google’s vice president-level engineer. His approach to artificial intelligence will dramatically change the role of computers in our lives.