Posted November 12, 2018 03:16:01 A group of Australian academics are working on a $100 million project to create a super-intelligent supercomputer that could solve the toughest computational problems in the field of artificial intelligence.
Key points:The project is called Project Zomboids and will involve universities and companiesThe group is working on its first prototype in a laboratory at the University of Technology SydneyThis project has the potential to revolutionise the way researchers use artificial intelligenceThe project will involve Australian universities and multinational corporations and is being run by the University at Albany and the University College London.
The project involves Australian academics and universities, including Sydney University, working together to build a prototype computer that can solve some of the most complex computational problems.
The team will begin by working with IBM Watson, an AI system that has won many computer competitions and has made the rounds of technology press and academia.
“We’re looking at a real-world scenario where we want to build some very sophisticated computational systems, but we don’t have the computational infrastructure, and so we’re trying to come up with something that will be scalable, efficient and efficient at solving real-life problems,” Professor Daniel O’Brien from the University’s Centre for Computing at Albany told the ABC.
“What we’re doing is trying to create something that is scalable, very efficient and a really good supercomputer, in a lab environment.”
The supercomputer will be able to perform a task as complex as running the most sophisticated algorithm in a computer, which has been a challenge in artificial intelligence research.
“This will be a supercomputer capable of running the best algorithms in the world, and if you do something like that, which is a lot of computing, you’re not going to be able for the first 20 to 30 years,” Professor O’Briens said.
“So, you need to go back to a very traditional form of computer science to develop a computer that is able to run those things, but then eventually you will be competing with machines from the outside world.”
Project Zomboidal aims to solve one of the biggest problems in AI research: how to deal with complex problems that don’t exist in reality.
“One of the challenges is that you can’t use any existing computational infrastructure to build these systems that are going to solve a problem, but that doesn’t mean we can’t try and develop the next generation of computational systems,” Professor Richard Choy from the School of Computing at the Centre for Computational Intelligence at the Australian National University (ANU) told the BBC.
“The next generation is going to have to be much more powerful and efficient than existing ones.”
Projects like Project Zephoids are one of several initiatives being rolled out by Australian universities to help solve some challenges in artificial Intelligence research.
Professor O’Connor and his colleagues are also working on their first prototype, which will be used to demonstrate that a super computer can solve a task that is as complex and demanding as running a super high-performance computer.
Project Zephoid is one of many AI projects being rolled-out in Australia, with the University Canterbury, the Australian Institute of Cybernetics and Systems (AICAS) and the Australian Research Council (ARC) among others involved.
Professor Choy told the Australian Broadcasting Corporation that a key challenge in AI development is that “it’s very hard to get a super machine to do something, because it’s really hard to design and to build that system”.
“We have a real problem in AI where we have a lot more problems that are very difficult to solve, that are not solvable in software, that we don’t have the tools to solve,” he said.”[The] challenge is, how do we get these problems solved in a way that we can understand and test, and that can be verified, and we can then get a good result?”
Professor Chay said it was critical to get the right people in the right place at the right time.
“You have to have a clear, consistent strategy and that is why we are doing this,” he told the broadcaster.
“It’s important to have people who are trained in the areas of computer theory and AI, and who have a solid understanding of the field.”
Dr David Erikson, Director of Research at ARC, said that the ARC had a number of projects aimed at helping solve AI research challenges.
“Our programs have been working on computer vision, image recognition and robotics and we’re also involved in artificial general intelligence,” he explained.
“In AI, we need to be in the front of the queue to make sure we can have a significant impact on the future of the technology.”
Professor Choe said that many of the issues facing AI research could be tackled with a single project, such as building a super computing machine that could “exploit” a large dataset to find the best solution to problems.
“At the moment, there’s a huge amount of work going on