Slashdot Mirror
DARPA Wants To Build an AI To Find the Patterns Hidden in Global Chaos (techcrunch.com)
A new program at DARPA is aimed at creating a machine learning system that can sift through the innumerable events and pieces of media generated every day and identify any threads of connection or narrative in them. It's called KAIROS: Knowledge-directed Artificial Intelligence Reasoning Over Schemas. From a report: "Schema" in this case has a very specific meaning. It's the idea of a basic process humans use to understand the world around them by creating little stories of interlinked events. For instance when you buy something at a store, you know that you generally walk into the store, select an item, bring it to the cashier, who scans it, then you pay in some way, and then leave the store. This "buying something" process is a schema we all recognize, and could of course have schemas within it (selecting a product; payment process) or be part of another schema (gift giving; home cooking).
Although these are easily imagined inside our heads, they're surprisingly difficult to define formally in such a way that a computer system would be able to understand. They're familiar to us from long use and understanding, but they're not immediately obvious or rule-bound, like how an apple will fall downwards from a tree at a constant acceleration. And the more data there are, the more difficult it is to define. Buying something is comparatively simple, but how do you create a schema for recognizing a cold war, or a bear market? That's what DARPA wants to look into.
Although these are easily imagined inside our heads, they're surprisingly difficult to define formally in such a way that a computer system would be able to understand. They're familiar to us from long use and understanding, but they're not immediately obvious or rule-bound, like how an apple will fall downwards from a tree at a constant acceleration. And the more data there are, the more difficult it is to define. Buying something is comparatively simple, but how do you create a schema for recognizing a cold war, or a bear market? That's what DARPA wants to look into.
>> a machine learning system that can sift through the innumerable events and pieces of media generated every day and identify any threads of connection or narrative in them.
This sound like a marketing question. As in, "how well are the talking points from various agencies and political groups represented in the media." There are communications firms that perform this type of analysis today on the messages they try to get out into the public (e.g., "this statistic we created - that's just a little bit off the official one so we can track it - has been republished in 228 news stories in the past 6 months").
Not freakonomics but psychohistory.
They're either buying into the same marketing and media hype for the half-assed excuse for AI everyone keeps trotting out, or they've got something nobody else has, meaning general AI. The latter is highly unlikely, if they did we wouldn't be hearing about it at all.
If they had general AI, they wouldn't need to build a super computer powered expert system just to tell whether someone is buying something. This is a complete waste of time. You have a better chance of reaching the moon by building longer and longer ladders than you do reaching general intelligence by hardcoding facts.
It is neither.
This is not exactly a new, they put out these proposal requests all the time. They select some problem of interest, or more often a whole bag of problems and post a request for proposals, then see what various groups think they can accomplish along those lines. The actual research is a lot less dramatic than pieces like this suggest, and really just represent DARPA giving seed grants based off some theme to a bunch of teams and seeing what they come up with.
They're either buying into the same marketing and media hype for the half-assed excuse for AI everyone keeps trotting out, or they've got something nobody else has, meaning general AI. The latter is highly unlikely, if they did we wouldn't be hearing about it at all.
Or the option you didn't mention -- they know that existing ML-based systems are really good at pattern-matching, and that many forms of pattern-matching don't require full general AI in order to be cost effective, and they're applying it (like the headline and article says) to another domain for pattern-matching.
I was thinking more 'Pi'...
"Listen to me. The Ancient
Japanese considered the Go
board to be a microcosm of
the universe. Although when
it is empty it appears to be
simple and ordered, in fact,
the possibilities of game play
are endless. They say that no
two Go games have ever been
alike. Just like snowflakes.
So, the Go board actually
represents an extremely complex
and chaotic universe. That is
the truth of our world, Max.
It can't be easily summed up
with math. There is no simple
pattern."
"I bless every day that I continue to live, for every day is pure profit."