IBM betting the future of AI is in quantum computing
There is no stopping artificial intelligence, but IBM and MIT figure they can speed up its development with a healthy infusion of cash and a brand new research center.
On Thursday, IBM and the Massachusetts Institute for Technology (MIT) announced a new $240 million research center called the MIT-IBM Watson AI Lab that’s focused, at least in part, on how new hardware can help energize and grow AI technologies.
“The field of artificial intelligence has experienced incredible growth and progress over the past decade. Yet today’s AI systems, as remarkable as they are, will require new innovations to tackle increasingly difficult real-world problems to improve our work and lives,” said Dr. John Kelly III, IBM senior vice president, Cognitive Solutions and Research in a statement.
One of those nascent technologies is quantum computing. It’s a technology that excites AI researchers because it holds the potential of far greater computing power, thanks to its ability to hold bits in multiple states at once. Traditional computing can only have bits that are on or off (1s and 0s).
Machine learning, which researchers have increasingly used to train AI systems, could become even faster and more powerful with quantum computers and materials.
IBM’s been quite active in the field for some time. Last year, it unveiled the world’s first cloud-based quantum computing experience.
The new lab will also look at materials sciences and how new computing materials, potentially for processing and storage, could enhance computing power (likely with quantum computers) and further turbo-charge AI development efforts.
The physics of AI is a critical area of study for the lab, said MIT-IBM Watson AI Lab co-chair Dario Gil in an email exchange with Mashable.
“Today, it takes an enormous amount of time to train high-performing AI models to sufficient accuracy. For very large models, it can be upwards of weeks of compute time on GPU-enabled clusters….Our teams will explore new materials, devices and architectures for analog AI computation, as well as the intersection of quantum computing and machine learning. The latter involves both using AI to help characterize and improve quantum devices, and also researching the use of quantum computing to optimize and speed up machine-learning algorithms and other AI applications,” he said.
In addition, the MIT-IBM AI Lab will investigate existing and new AI algorithms and focus on how AI can help cybersecurity and healthcare, though Gil told me they will look at professional use of AI in other industries, as well.
Gil will co-chair the lab with Anantha P. Chandrakasan, dean of MIT’s School of Engineering. They plan to ask MIT and IBM researchers to submit research proposals for joint study.
According to IBM, this 10-year collaboration will involve 100 researchers, professors, and students across IBM research centers in Cambridge Massachusetts and MIT’s campus.
The new AI lab is a continuation of along partnership with IBM. “Our ability to solve complex technology problems with MIT has given us the confidence to commit to a partnership of this magnitude,” he said.
As for the safety of AI and how this lab’s work might help us avoid (or bring us closer to) SpaceX and Tesla CEO Elon Musk’s free-floating fear of an AI-related global conflict, Gil pointed to how the MIT-IBM lab plans to embed rules into their systems while not killing what makes AI special.
“How we embed ethical values into our machines is two-fold – we create a set of rules that the system must comply with, while also giving AI systems machine-learning freedom. Artificial intelligence technology is intrinsically proficient at identifying nuances and applying these insights.
To adapt values from humans to machines we must first define these tenants – before we can embed them into our systems. This is an active topic of conversation among ethicists, and something we will shine greater light on through this research program with MIT,” he wrote.