Can we teach computers to think? Systems that reason and learn may be closer than they appear
Cognitive systems will help people and computers partner to reduce complexity and enhance our ability to reason through difficult decisions
systems is based on creating a partnership between people and machines to enhance human cognition. Back in the days of the mainframe, operators accessed machines to help with massive calculations. The era of personal computers introduced a new kind of interaction built around optimizing tasks and information. Our mobile era has again changed the way we interact with computers to something up close and personal, where machines function as a type of personal assistant. But with cognitive computers, our interaction will be an entirely new experience. Computers have always been used to enhance human capabilities – whether our physical limitations, productivity, or connectivity. IBM is proposing new systems that will branch out from the path of calculations and move on to learning and reasoning.Cognitive computing is a branch of computing that draws inspiration from the human brain. We usually think of computers for automating tasks or calculations – but cognitive systems are an entirely new approach. The driving principle of these
On November 21, IBM Research – Haifa held a unique colloquium focused on cognitive systems and this new era of computing. The event brought together world experts, industry leaders, and academics from Israel and abroad to share a peek into the future of computing systems and see how future architectures are being inspired by the most efficient reasoning tool we know – the human brain.
Creating a confluence of humans and computers
"Cognitive systems are based on a partnership with humans to provide reasoning and learning systems, with a completely different design," explained Dr. Dario Gil, Director of Cognitive Computing at IBM Research. "As opposed to working towards automation, they are driven towards helping us manage complexity." Examples abound of complex problem areas where humankind needs real help. Such areas include cyber security, which accounts for $400 billion in losses annually; certain cancer treatments that have a 44% rate of misdiagnosis; or the fact that only 22% of students worldwide graduate from high school. Gil believes that we only have a limited window of opportunity to solve these problems.
What can we learn from the human brain?
World-renowned expert Prof. Idan Segev brought the audience up to date on new research into the human brain. "With 100 billion neurons, we can do things no computer can even come close to," he explained. "We want to learn from the brain--but first we need to understand it. This will ultimately help us both repair the brain and build machines that can enhance our capabilities." According to Segev, there are several international organizations allocating hundreds of millions of dollars towards better understanding the brain. He strongly believes that we are at a historical juncture of brain research, in which groups of scientists are working together to make progress in this area.
"We all expect computers to evolve and change, but the progress will happen by strengthening the integration of people and computers," noted Oded Cohn, VP and director of IBM Research - Haifa. "To what extent is brain research relevant to building the computers of the future and are all elements of the brain transformative? Some aspects are relevant and some are not – but it's certainly a good place to start."
Can a blind person understand our world?
One of the highlights of the event was a talk by IBM Fellow Dr. Chieko Asakawa, who spoke about how innovation has been applied to improve accessibility. IBM has a long history of helping people with disabilities, including the production of a Braille printer in 1975 and a Home Page Reader in 1997, which allowed blind people to use the Internet to read newspapers, shop online, or even check their bank balance. Blind since the age of 11, Asakawa was one of the first people to use that device.
Computers are bringing people the power of information and knowledge. But can they be used to help a blind person sense all the complexity going on around them? Asakawa feels that with cognitive computing, things that are impossible today will become possible for the blind. For example, systems can now recognize faces and emotions and give that information to a blind person. New mobile apps, like the personal shopping assistant developed in Haifa, can help recognize products on store shelves and advise which is the best choice. And new smart table technology can provide audio information about the food on it, including calories and diet advice.
IBM Israel country general manager Rick Kaplan recounted being present during a recent conversation between Israeli President Shimon Peres and IBM CEO Ginni Rommety as they discussed cognitive computing. "By the end of the conversation," he noted, "they even had me convinced that cognitive computing could be used to solve the problem of terrorism in the world."
The event included talks on current research into the human brain and the latest work on surrogate bodies and thought-controlled robots. There were also talks on how the brain deals with uncertainty as well as the vast opportunities for diagnostics, brain monitoring, neural feedback, and brain computer interfaces. Demos for a number of cognitive computing projects from the scientists at IBM Research – Haifa also starred at the event. These included decision support for radiologists, a tool for cervical cancer prevention and treatment in Africa, and search and retrieval enablement for multimedia and video content.
Event participants were enthusiastic about the messages presented. "The colloquium talks had a futuristic and optimistic message," one noted, "leaving us with a feeling that we are privileged to be living in this era when so much progress is taking place so quickly." Perhaps the future and cognitive systems are nearer than we think.