Interviews: Ask Dr. Andy Chun About Artificial Intelligence

samzenpus (5) writes "Dr. Andy Chun is the CIO for the City University of Hong Kong, and is instrumental in transforming the school to be one of the most technology-progressive in the region. He serves as an adviser on many government boards including the Digital 21 Strategy Advisory Committee, which oversees Hong Kong's long-term information technology strategies. His research work on the use of Artificial Intelligence has been honored with numerous awards, and his AI system keeps the subway in Hong Kong running and repaired with an amazing 99.9% uptime. Dr. Chun has agreed to give us some of his time in order to answer your questions. As usual, ask as many as you'd like, but please, one question per post."

Comments

Broader implications

[Iamthecheese]

What real-world problems are best suited to the kind of programming used to manage the subway system? That is to say, if you had unlimited authority to build a similar system to manage other problems which problems would you approach first? Could it be used to solve food distribution in Africa? Could it manage investments?

Hubert Dreyfus

[MAXOMENOS]

Have you read Professor Dreyfus's objections to the hopes of achieving "true AI" in his book What Computers Can't Do? If so, do you think he's full of hot air? Or, is the task of AI to get "as close to the impossible" as you can?

Narrow down to one thing that needs improvement

[gunner_von_diamond]

If you had to narrow it down to one thing that needs the most improvement in the field of AI, something that we could focus on, what would it be?

Will we know when we create it?

[wjcofkc]

Considering we have yet to - and may never - quantify the overall cognitive process that gives rise to our own sentient intelligence, will we have any way of knowing if and when we create a truly aware artificial intelligence?

Current progress

[JoeMerchant]

Dr Chun,

What area of AI development is currently making the most progress? In other words, where are the next big advances most likely to come from?

Question for Dr. Andy Chun

[ArcadeMan]

Dear Dr. Chun,

why do I have this terrible pain in all the diodes down my left side?

Slashdot, please don't...

[wjcofkc]

Slashdot editors,

Please don't ruin this by turning it into a video interview where you don't actually ask anyone's questions like you did the last one.

Sincerely,
Speaking for a lot of us.

Where do you see A.I. in 5,10,20, and 30 years?

[Maxo-Texas]

And what's the latest date you see A.I. that is conscious and self aware in the human and animal sense?

Re:Where do you see A.I. in 5,10,20, and 30 years?

[Maxo-Texas]

Actually, we are pretty close to discovering what consciousness is physically.

They've found one spot in the brain that when stimulated electrically, you don't go asleep but your "conciousness" turns off. When the stimulation stops, you recover conciousness without an awareness of any time passing.

The particular part appears to be acting like a conductor of multiple streams of information from the rest of the brain. For some reason in 70 years of this type of research, they'd never explored that particular part of the brain yet.

If it is the seat of consciousness then it's physical configuration may lead to new theoretical and machine implementations within a 30 year window.

---

To be blunt, everyone I've known personally that felt machine consciousness was impossible had a religious basis for that belief. Basically, despite all evidence to the contrary from brain research and animal research, that consciousness resided in a "soul" that was independent of the human body or with that as a basis posited that consciousness was a quantum effect (i.e. god of the gaps) which humans would be unable to duplicate.

Definition of AI?

[Okian Warrior]

Can you explain to us exactly what AI is?

As a definition, the Turing test has problems - it assumes communication, it conflates intelligence with human intelligence, and humans aren't terribly good at distinguishing chatbots from other humans.

Also, using a test for a definition works well in mathematics, but not so much in the real world. Imagine defining a car as "anything 5 humans say is a car" and then trying to develop one. Without feedback or guidance, the developers have to trot every object in the universe in front of a jury, only to receive a yes/no answer to the question: "is this a car?"

Many AI texts have a 'kind of fuzzy, feel-good definition of AI that's useless for construction or distinguishing an AI program from a clockwork one. Definitions like "the study of programs that can think", or "programs that simulate intelligent behaviour" shift the burden of definition (of intelligence) onto the reader, or become circular.

One could define a car as "a body, frame, 4 wheels, seats, and an engine in this configuration", and note that each of these can be further defined: a wheel is a rim and a tire, a tire is a ring of steel-belted rubber with a stem valve, a stem valve is a rubber tube with a schrader valve, a schrader valve is a spring and some gaskets...

With a constructive definition, one could distinguish between a car and, say: a tractor, a snowmobile, a child's wagon, a semi, and so on. Furthermore, it would be conceptually straightforward to build one: you know where to start, and how to get further information if you are unsure.

Compare with a group from mathematics: a closed set plus an operator with certain features (associativity, identity, inverses), and each feature can be further defined (an identity element is...). Much of mathematics is this way: concepts constructed from simpler concepts with a list of requirements.

The study of AI seems to be founded in mathematics. At least, all the AI papers I've read are heavy with mathematical notation - usually obscure and very dense mathematical notation. It should be possible to determine with some rigor what the papers are talking about.

Can you tell us what that is? What *exactly* is AI?

What do your think of Fergus et all paper

[L'Ange Oliver]

What do your think of Fergus et all paper : Intriguing properties of neural networks. Is this a phenomenon you came across before?

How similar is your AI boss to the fictional Manna

[Ken_g6]

Dr. Chun,

Have you read a short story about an AI boss called Manna? (I'll include relevant quotes if you don't have time.) How does your system for the Hong Kong subway compare? It's clearly similar to your subway system in some ways:

At any given moment Manna had a list of things that it needed to do.... Manna kept track of the hundreds of tasks that needed to get done, and assigned each task to an employee one at a time.

But does it micro-manage tasks like Manna?

Manna told employees what to do simply by talking to them. Employees each put on a headset when they punched in. Manna had a voice synthesizer, and with its synthesized voice Manna told everyone exactly what to do through their headsets. Constantly. Manna micro-managed minimum wage employees to create perfect performance.

Does it record employee performance metrics and report them to (upper) management like Manna?

Version 4.0 of Manna was also the first version to enforce average task times, and that was even worse. Manna would ask you to clean the restrooms. But now Manna had industry-average times for restroom cleaning stored in the software, as well as "target times". If it took you too long to mop the floor or clean the sinks, Manna would say to you, "lagging". When you said, "OK" to mark task completion for Manna, Manna would say, "Your time was 4 minutes 10 seconds. Industry average time is 3 minutes 30 seconds. Please focus on each task." Anyone who lagged consistently was fired.

And how have employees reacted to their AI boss - if, in fact, you have been able to get honest evaluations from employees?