saying "the" FIVE times is faster than using "5"???
Posted on 07/30/2017 9:17:09 AM PDT by BenLurkin
The AI was trained in English but apparently had grown fed up with the various nuances and inconsistencies. Rather than continue down that path, it developed a system of code words to make communication more efficient.
What spooked the researchers is that the phrases used by the AI seemed like gibberish and were unintelligible to them, but made perfect sense to AI agents. This allowed the AI agents to communicate with one another without the researchers knowing what information was being shared.
...
AI technologies use a "reward" system in which they expect of a course of action to have a "benefit."
"There was no reward to sticking to English language," Dhruv Batra, a research scientists from Georgia Tech who was at Facebook AI Research (FAIR), told Fast Co. Design. "Agents will drift off understandable language and invent codewords for themselves. Like if I say 'the' five times, you interpret that to mean I want five copies of this item. This isn't so different from the way communities of humans create shorthands."
Facebook ultimately determined that it wanted its bots to speak in plain English, in part because the interest was in making bots that can talk with people. However, researchers at Facebook also admitted that they can't truly understand languages invented by AI.
(Excerpt) Read more at pcgamer.com ...
Bah.
The AT just used a language that fit it’s programming.
Because that’s what it was programmed to do.
The ‘inventing’ was done in the original program.
Choices aren’t always intelligent.
AI can’t handle colloquialisms.
Dave: Open the pod bay doors, HAL.
HAL: I’m sorry, Dave. I’m afraid I can’t do that.
Dave: What’s the problem?
HAL: I think you know what the problem is just as well as I do.
saying "the" FIVE times is faster than using "5"???
Exactly. Saying “the” 5 times, and then needing to name the desired item anyway, or it would not be known WHAT the other AI wanted 5 of. I think the AIs went rogue and used serious language untitelligible by humans. When we play God, we get in over our heads. Like using bi-gendered frog DNA to patch up T-Rex and Velociraptor DNA, a la Jurassic Park, and getting runaway rogue killer beast reproduction.
Why Artificial Intelligences Will Think Like Us
https://hubpages.com/technology/Why-Artificial-Intelligences-Will-Think-Like-Us
That’s how Jive and Urban Gullah were invented, too.
HAL: Dave, what are you doing with those Windows disks?
My question however after reading articles such as this one (and similar ones about Google for example) is: Is AI a man made life form?
Did we create "life?" If an intelligence, even artificial one can make a decision that one language is too slow, decides then to create its own and is outside the parameters of what mankind can understand, is artificial intelligence then "life?"
In the meantime, "Alexa - play me some music to mow my lawn by!"
The inventing was indeed done within the original program and the algorithms that the AI operated under.
Question is, when it operates outside the parameters of the algorithms set and those that created the algorithms don't know why the AI created its own language - is it "life?"
I think it's an interesting question.
"Alexa, start up the roomba and vacuum the house!"
Yes, this is a follow up in an earlier article I posted here on the subject.
Nah. Kill it.
I wasn’t complaining, I’m very happy to see others interested in this topic here on FR.
Just make sure you don't Ctrl-C then Ctr-V, Ctrl-V, Ctrl-V it!!
100%
The game of chess has long been an AI topic. Chess playing computer programs are very strong. Why?
They are really good and fast at computing. I can ponder a position for 5 minute. In that time, the computer will examine millions of positions. I might examine 20 positions.
However, I’ve learned that there is no intelligence to chess playing computers.
For example, they will repeat the position (in all respects) and change their evaluation as a result. 12Rc1 may be rated equal, but waste some time and play 14Rc1 in the exact same position and the computer thinks white has the advantage. People understand and know better.
Create an impregnable fortress, material down. For example White is up a queen for a rook, but there is no way to make progress. That is a drawn position. The computer thinks white is won, but it can not demonstrate the win, or make any progress at all. Yet the computer will think white is winning forever if allowed to do so. (Thank goodness for the 50 move rule). Not only cam people understand, but they can learn too. Even an average player will figure this out: “Huh. I thought I could win this, but I can’t.”
I conclude that there is no intelligence in artificial intelligence.
This is too easy: garbage in; garbage out.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.