SoylentNews is people
SoylentNews
Will Knight writes:
No one really knows how the most advanced algorithms do what they do. That could be a problem.
Last year, a strange self-driving car was released onto the quiet roads of Monmouth County, New Jersey. The experimental vehicle, developed by researchers at the chip maker Nvidia, didn't look different from other autonomous cars, but it was unlike anything demonstrated by Google, Tesla, or General Motors, and it showed the rising power of artificial intelligence. The car didn't follow a single instruction provided by an engineer or programmer. Instead, it relied entirely on an algorithm that had taught itself to drive by watching a human do it.
Getting a car to drive this way was an impressive feat. But it's also a bit unsettling, since it isn't completely clear how the car makes its decisions. Information from the vehicle's sensors goes straight into a huge network of artificial neurons that process the data and then deliver the commands required to operate the steering wheel, the brakes, and other systems. The result seems to match the responses you'd expect from a human driver. But what if one day it did something unexpected—crashed into a tree, or sat at a green light? As things stand now, it might be difficult to find out why. The system is so complicated that even the engineers who designed it may struggle to isolate the reason for any single action. And you can't ask it: there is no obvious way to design such a system so that it could always explain why it did what it did.
The mysterious mind of this vehicle points to a looming issue with artificial intelligence. The car's underlying AI technology, known as deep learning, has proved very powerful at solving problems in recent years, and it has been widely deployed for tasks like image captioning, voice recognition, and language translation. There is now hope that the same techniques will be able to diagnose deadly diseases, make million-dollar trading decisions, and do countless other things to transform whole industries.
[...] The U.S. military is pouring billions into projects that will use machine learning to pilot vehicles and aircraft, identify targets, and help analysts sift through huge piles of intelligence data. Here more than anywhere else, even more than in medicine, there is little room for algorithmic mystery, and the Department of Defense has identified explainability as a key stumbling block.
[...] At some stage we may have to simply trust AI's judgement or do without using it. Likewise, that judgement will have to incorporate social intelligence. Just as society is built upon a contract of expected behaviour, we will need to design AI systems to respect and fit with our social norms. If we are to create robot tanks and other killing machines, it is important that their decision-making be consistent with our ethical judgements.
https://www.technologyreview.com/s/604087/the-dark-secret-at-the-heart-of-ai/
What do you think, would you trust such AI even if you couldn't parse its methods? Is deep learning AI technology inherently un-knowable?
(Score: 3, Insightful) by AthanasiusKircher on Wednesday April 12 2017, @04:59AM (1 child)
But that's the problem in TFA -- how do you know for sure what is "learned"? We're very far from being able to ask an AI algorithm to explain the abstract rules it may have "learned," or even for an AI algorithm to generate abstractions.
What you instead likely have is a huge amount of weightings generated through interactions with input, effectively a bunch of numbers. What do they mean? How can you be sure what was "learned"? You may be able to abstract some general tendencies by studying what the algorithms spit out, but just because a car responded correctly to 1 unusual situation safely, does it actually " know" how to interpret inputs to respond to 10,000 real-world variants of that scenario that any human with an IQ over 80 would be able to see as similar (and thus use a similar strategy to avoid)?
You're thinking in terms of traditional explicit coding of rules where there are "bugs" that can be fixed and now you're guaranteed that the program will respond correctly to a new precisely defined class of input. TFA is about the ambiguity in AI systems and how their behavior can be a lot harder to quanitify or explain explicitly like that... Meaning you might not be able to determine how much better a software update is until it encounters 10,000 similar and unexpected real-world scenarios.
(Score: 3, Insightful) by kaszz on Wednesday April 12 2017, @05:47AM
And because AI systems lack judgement nor can be quantified. They don't belong on roads where they can maim and kill.