A "new Siri" has been developed by Brian Williams and his team at MIT that could eventually help NASA plan their missions in space. This new and improved artificial intelligence in not only conversational, meaning that it can smoothly interact with the user, but it is also risk-assessing. This means that the algorithms used can work by reasoning within a set of goals and constraints given by the user, while taking into account different probabilities which allows for a certain level of risk. For example, if you tell this AI that you want to catch a bus with a 90% chance of success, the AI is going to assume that you're fine with a 5% chance of taking some other form of transportation. The same type of AI can be used with NASA, but since the missions are more delicate and offer many different obstacles, the algorithm would just need to take into account more constraints and calculate more alternatives. But this would be really cool to have in cars, because it could function as the core of autopilot, which would be greatly beneficial in many ways.
More information along with video here: http://motherboard.vice.com/read/the-problem-solving-ai-of-the-future-will-be-on-your-phone
Why do you believe the "new siri" could eventually help NASA?
ReplyDeleteHow is risk quantified? As a mere percentage?
ReplyDeleteThis is highly interesting. What I'm wondering is how would a scientist be able to program every possible scenario in a car for example for the computer to be able to asses the risks associated with something.
ReplyDeleteThis is highly interesting. What I'm wondering is how would a scientist be able to program every possible scenario in a car for example for the computer to be able to asses the risks associated with something.
ReplyDelete