Recently, during its 60th-anniversary conference, DARPA revealed its $2 billion investment plan to spearhead artificial intelligence. “We think it’s a good time to seed the field of AI,” the agency’s Deputy Director of Information Innovation Office John Everett said to CNNMoney. “We think it’s a good time to seed the field of AI. He added: “We think we can accelerate two decades of progress into five years.”
Known for allowing machines to undertake tasks that were previously done by humans, artificial intelligence (AI) is a popular subject in both business and technology circles.
For instance, Google recently alarmed and delighted observers by demonstrating to them how an AI system can contact a restaurant and make a reservation, all while sounding like a human.
In the last decade, technological innovations have triggered companies to hire top artificial intelligence (AI) talent outside the academic circles. In fact, machines are currently more accurate when it comes to identifying speech, processing words, and understanding images.
As a result of this breakthrough, various products such as Waymo’s self-driving vans, Apple’s Siri, and Amazon’s Alexa have emerged.
Currently, the nation’s largest and most innovative entities depend on artificial intelligence to remain ahead of their competitors. Waymo’s self-driving cars have driven over 9 million miles, especially on US roads, all thanks to this revolutionary technology. On the other hand, national governments like France, India, China, and Canada have begun to give AI priority.
They see it as a vital aspect of improving their economic performance in the 21st century. China has gone ahead to declare plans to become an AI global leader by 2030.
DARPA’s $2 billion investment plan is set to focus on developing systems with better efficiency, common sense and contextual awareness. With such improvements, the government could be in a better position to accredit software systems, make AI systems explain themselves, and automate security clearances.
Machine learning depends on algorithms that learn from massive data sets. A computer may be shown millions of cat images and after some time it would identify when another picture includes a cat. However, such AI systems mostly rely on thousands of computer chips, which process data for weeks, before they can learn anything.
Everett said: “machine learning is remarkably efficient. It can do amazing things, but it’s also remarkable what it can’t do.” In this case, he said that the next wave of AI is needed in a bid to support complex in-home robots.
Everett added that if you request a robot to “pick up the living room,” it would not know what to do. The robot would struggle to recognize the items that require being picked up and which ones do not.
DARPA intends to adopt AI techniques that are similar to the way humans learn. In some cases, humans can learn a particular thing from simply watching one example.
According to Everett, the agency may eventually invest additional money into AI development. He said: “if we get positive results and they’re important, and they’re relevant to the military and national security, we’re not going to stop.”