(On the Merits of Parrots … or: “The Wisdom of Crowds” as a Strategy for Educating Young AI’s)

In this blog post I’ll enlarge upon a point I made during my recent talk at TransVision 2007 (see a recent blog post by Bruce discussing this talk), regarding the potential of virtual worlds to help in accelerating the path of AI’s toward mastery of human language.

(This is just an example of a more general point: The more I think about the direction Novamente has chosen to take over the next few years — seeking to roll out intelligent virtual agents widely throughout 3D and 2D virtual worlds and MMOG’s — the more I become convinced that it’s a very positive direction from a pure-AGI perspective as well as from a business perspective.)

As a specific example, one vision that’s been haunting me lately is a virtual talking parrot. A simple idea, of course — but very powerful in its AI implications. Imagine millions of talking parrots spread across different online virtual worlds — all communicating in simple English. Each parrot has its own local memories, its own individual knowledge and habits and likes and dislikes — but there’s also a common knowledge-base underlying all the parrots, which includes a common knowledge of English.

[Bruce Klein w/ parrot at Novamente’s Second Life HQ]

Now, suppose that an adaptive language learning algorithm is set up (based on, oh, say, the Novamente Cognition Engine), so that the parrot-collective may continually improve its language understanding based on interactions with users. If things go well, then the parrots will get smarter and smarter at using language, as time goes on. And, of course, with better language capability, will come greater user appeal.

The idea of having an AI’s brain filled up with linguistic knowledge via continual interaction with a vast number of humans, is very much in the spirit of the modern Web. Wikipedia is an obvious example of how the “wisdom of crowds” — when properly channeled — can result in impressive collective intelligence. Google is ultimately an even better example, I think — the PageRank algorithm at the core of Google’s technical success in search, is based on combining information from the Web links created by multi-millions of Website creators. And the intelligent targeted advertising engine that makes Google its billions of dollars is based on mining data created by the pointing and clicking behavior of the one billion Web users on the planet today. Like Wikipedia and Google, the mind of a talking-parrot tribe instructed by masses of virtual-world residents will embody knowledge implicit in the combination of many, many peoples’ interactions with the parrots.

Another thing that’s fascinating about virtual-world embodiment for language learning is the powerful possibilities it provides for disambiguation of linguistic constructs, and contextual learning of language rules.

Michael Tomassello, in his excellent book Constructing a Language, has given a very clear summary of the value of social interaction and embodiment for language learning in human children.

For a virtual parrot, the test of whether it has used English correctly, in a given instance, will come down to whether its human friends have rewarded it, and whether it has gotten what it wanted. If a parrot asks for food incoherently, it’s less likely to get food — and since the virtual parrots will be programmed to want food, they will have motivation to learn to speak correctly. If a parrot interpret a human-controlled avatar’s request “Fetch my hat please” incorrectly, then it won’t get positive feedback from the avatar — and it will be programmed to want positive feedback.

Yes, humans interacting with parrots in virtual worlds can be expected to try to teach the parrots ridiculous things, obscene things, and so forth. But still, when it comes down to it, even pranksters and jokesters will have more fun with a parrot that can communicate better, and will prefer a parrot whose statements are comprehensible.

What it comes down to is: A virtual parrot, learning language, will have lots of teachers, and that’s a good thing. The more customers we get for the parrot, the more teachers the AI underlying the parrot will have.

A baby AI has a lot of disadvantages compared to a baby human being: it lacks the intricate set of inductive biases built into the human brain, and it also lacks a set of teachers with a similar form and psyche to it … and for that matter, it lacks a really rich body and world.

However, the presence of thousands to millions of teachers constitutes a large advantage for the AI over human babies. And a flexible AGI framework, like the Novamente Cognition Engine, will be able to effectively exploit this advantage.

On a more theoretical level, this community of individually-acting yet collaboratively-learning parrots may be considered an example of a mindplex, a term I introduced in this essay a couple years back, referring to a collection of minds in which

• each individual mind has its own declarative and procedural memories, and sense of self
• there is also a collective declarative and procedural memory, and a collective sense of self

Mindplexes tie in interestingly with the notion of the emerging global brain, which I discussed extensively in my 2002 book Creating Internet Intelligence.

Getting back to practicalities: The rate of progress of Novamente LLC in our new business direction is difficult to estimate, as it depends on funding and other related issues. But, if all goes as we’re hoping, we may well be able to release a parrot-that-talks-and-adaptively- learns-to-talk-better sometime before the end of 2008. And that will be pretty exciting!

And of course parrots are not the end of the story. Once the collective wisdom of throngs of human teachers has induced powerful language understanding in the collective bird-brain, this language understanding (and the commonsense understanding coming along with it) will be useful for other purposes as well. Humanoid avatars — both human-baby avatars that may serve as more rewarding virtual companions than parrots or other virtual animals; and language-savvy human-adult avatars serving various useful and entertaining functions in online virtual worlds and games. Once AI’s have learned enough that they can flexibly and adaptively explore online virtual worlds (and the Internet generally) and gather information according to their own goals using their linguistic facilities, it’s hard to see limits to their growth and understanding. (And this leads to various deep and critical ethical concerns, such as I’m exploring with my colleagues at the Singularity Institute for AI.)

But, we need to get there one step at a time. What’s exciting about virtual parrots-that-talk — and the intelligent virtual agents space generally — is the way it poses an incremental path by which getting more and more customers for products is directly connected to making the AI underlying the products smarter and smarter (which in turn will attract more and more customers). This is exactly the kind of virtuous cycle one wants to see in an AI start-up company (in my never-very-humble and admittedly rather biased opinion!).

Ben Goertzel
CEO and Chief Scientist
Novamente LLC

This entry was posted on Saturday, July 28th, 2007 at 8:16 pm and is filed under Novamente update. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.