this lecture. Godfather of AI, Geoffrey Hinton, recently gave a talk at the prestigious Royal Institution and while it is fairly introductory as far as AI goes, you really do need to see it.

The talk is called Digital Intelligence verses Biological Intelligence and if you hold the mainstream view of AI you no doubt imagine that he is going to point out how much better bio intelligence is than artificial intelligence.

But this is just the narrow view of where we are.

Hinton wants to talk about where we are inevitably going. No, this isn't Sci Fi - it Is a reasoned account of cause and effect as he sees it.

The talk starts off with some history - the two approaches to AI - logic v biologically-inspired. In the early days the logical approach seemed like the only one worth following; so much so that the people following the neural-network idea were mostly classified as cranks and denied funding. Hinton was one of the few who persisted with the idea that biology beats logic. If you know about neural networks you can skip to just short of five minutes in where he tells us that neural networks beat the logic approach to vision and then did the same to language.

This part of the talk is particularly interesting as it looks at the way language is a model for the world. To quote:

"But what about language? So there's a whole community that studies language, I think they're called linguists, and they have a very strong idea about how you should study language, and particularly, the Chomsky school. They were very sceptical that neural networks could do anything with language. They were completely convinced that it's all about symbolic expressions. They didn't really have the idea that the real function of language is to give you words which are bricks from which you can build models. Language is a modelling medium. 

They were focused on syntax. And syntax isn't the main point. The main point is, language is a wonderful way to build a particular kind of complicated model, as we'll see. Okay. They also thought that knowledge of language was innate, knowledge of syntax was innate, which is just stupid. (audience laughing) It's the sign of a cult, that in order to join the cult, you have to believe something that's obviously silly, like language isn't learned. (audience laughing)

Hinton goes on to contrast two theories of the meaning of a word. He first refers to symbolic AI theory which goes back to Swiss linguist Ferdinand de Saussure whose central tenet was that the meaning of a word isn't inherent to the word itself but rather is to to do with its relationships to other words and, Hinton suggests, requires something like a relational graph to capture its meaning.

Moving on to the alternative he continues:

But then the psychologists, who particularly from the 1930s, I think, who thought the meaning of a word is a big set of features. So "Tuesday" has a big set of active features, and "Wednesday" has a big set of active features that are almost the same. So the idea that the meaning of a words, a set of active features is very good for saying which words means similar things to which other words. These look like two very different theories of meaning. Now, what I want to do is show you that these two theories can be unified. They're not two different theories."

Yes this is a sarcastically humorous but deep take on where we are and I hope these quotes makes you want to view the rest of the talk - it;s well worth the effort.

What about the headline about not sleeping tonight if you understand what he is saying?

I'm not going to spoil the talk for you, but the clue is that Hinton is sure that while our current AI isn't as good as biology, the fundamental mechanisms that we have invented are so much better. They learn faster, communicate better and can be replicated. It is not the Skynet type of threat. It is not that AI is going to rebel and take us over - they might but that's not the real worry. The real worry is that we have invented something that has the potential to be better than us at thinking.

More Information

The Royal Institution

Related Articles

Geoffrey Hinton And The Existential Threat From AI

Geoffrey Hinton Shares Nobel Prize For Physics 2024

Geoffrey Hinton Awarded Royal Society's Premier Medal

Hinton, LeCun and Bengio Receive 2018 Turing Award

Geoffrey Hinton Says AI Needs To Start Over

Neural Turing Machines Learn Their Algorithms

Hinton Explains His New Fear of AI

To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on Twitter, Facebook or Linkedin.

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info