Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Training AI to Do Everything in the Digital Universe

singularityhub.com

Post reply on HN
Print view View original

11–20 of 23 posts

Re: Training AI to Do Everything in the Digital Universe

#11

Post by moxious »

ramzyo wrote: post #4

Interesting read, thank you for sharing. The author summarizes the thesis of the OpenAI team to be that by exposing AIs to a variety of experiences, the AIs will learn flexible problem solving skills. I'm not convinced this is true. Without imparting upon the AI some mechanism for reasoning across experiences, (e.g. reasoning through analogy), the AI will simply be trained on many specific experiences. How does the A…

I just don't trust general, high level descriptions. People have been going at this goal for decades without much proveable success on general intelligence. So when I read generalities I think they're not any further, just still trying. But when someone gives you a gory technical breakdown of a subproblem, that's when you know you've got something and they're serious.

Re: Training AI to Do Everything in the Digital Universe

#13

Post by wubbfindel »

marvin wrote: post #9
sdenton4 wrote: post #6

Earlier quoted context omitted.

I'm curious about the total number of image recognition neutral networks that have ever been trained...

Each instance of an AI only has to do it once, which is an important distinction, but still relevant. E.g. once we get to the point where a human-level housekeeping robot is possible, you don't have to wait 10 years after purchasing it for it to do useful things in your house :)

If an AI is duplicated, so that a new instance is created (pro-creation), then surely the new instance already knows what the parent instance learnt?

Re: Training AI to Do Everything in the Digital Universe

#15

Post by ge96 »

robochat wrote: post #14

This reminds me so much of Ted Chiang's story "the lifecycle of software objects". http://subterraneanpress.com/magazine/fall_2010/fiction_the_...

Damn I looked it up to see if that was a real game. haha, still reading

edit: holy christ look at the size of that scroll bar nvm

Re: Training AI to Do Everything in the Digital Universe

#16

Post by Jack000 »

ge96 wrote: post #15
robochat wrote: post #14

This reminds me so much of Ted Chiang's story "the lifecycle of software objects". http://subterraneanpress.com/magazine/fall_2010/fiction_the_...

Damn I looked it up to see if that was a real game. haha, still reading edit: holy christ look at the size of that scroll bar nvm

it's a great short story, well worth the time

Re: Training AI to Do Everything in the Digital Universe

#17

Post by maerF0x0 »

ramzyo wrote: post #4

Interesting read, thank you for sharing. The author summarizes the thesis of the OpenAI team to be that by exposing AIs to a variety of experiences, the AIs will learn flexible problem solving skills. I'm not convinced this is true. Without imparting upon the AI some mechanism for reasoning across experiences, (e.g. reasoning through analogy), the AI will simply be trained on many specific experiences. How does the A…

I think the concept of reasoning through analogy could kind of be like "call a friend" on game shows. If an AI had a directory of task specific agents, it could call on those agents to provide "advice".

Re: Training AI to Do Everything in the Digital Universe

#18

Post by ramzyo »

maerF0x0 wrote: post #17
ramzyo wrote: post #4

Interesting read, thank you for sharing. The author summarizes the thesis of the OpenAI team to be that by exposing AIs to a variety of experiences, the AIs will learn flexible problem solving skills. I'm not convinced this is true. Without imparting upon the AI some mechanism for reasoning across experiences, (e.g. reasoning through analogy), the AI will simply be trained on many specific experiences. How does the A…

I think the concept of reasoning through analogy could kind of be like "call a friend" on game shows. If an AI had a directory of task specific agents, it could call on those agents to provide "advice".

This is a really interesting idea, thanks for sharing!

Re: Training AI to Do Everything in the Digital Universe

#19

Post by thallukrish »

Humans have the drive to go about because of a body that has various demands. And with every doing, humans accumulate Ego which again helps their drive to learn, explore, do stuff or define their behaviour. Humans are self-learning systems. The rewards and penalty are also self-created.

Unless general AI systems start to model the above which creates self-learning, they will be just specific systems with a tiny aspect of human world ingrained.

Re: Training AI to Do Everything in the Digital Universe

#20

Post by emmab »

thallukrish wrote: post #19

Humans have the drive to go about because of a body that has various demands. And with every doing, humans accumulate Ego which again helps their drive to learn, explore, do stuff or define their behaviour. Humans are self-learning systems. The rewards and penalty are also self-created. Unless general AI systems start to model the above which creates self-learning, they will be just specific systems with a tiny aspec…

You don't need to program Ego to program a drive to learn new things, just a procedure to calculate the expected value of new information from engaging in some candidate training activity.
Post reply on HN