MIT’s Spark Lab Ponders Better Self-Driving Systems And More

How do we advance artificial intelligence to the point that it can think more like a human?

Well, first of all, we need to solve some of these problems because they can be deadly…

As for the ‘how’ of the problem, one thing that MIT groups are looking at is developing beyond current models that have demonstrated the limitations of modern AI.

One of the main problems is local minima – having these valleys in which systems can get stuck or blinkered, not understanding the global scenario, and as a result, not responding appropriately…

Over at the MIT Spark lab, Luca Carlone and others are trying to apply these solutions to get us closer to better AI responses. “These capabilities will unlock a huge number of practical applications,” Carlone says, citing improvements to systems aiding first responders, growing crops, and landing space vehicles, as well as improving self-driving cars.

Receive Our Newsletter

Get the latest Safety21 news sent to your inbox.