Thursday, January 27, 2005

General intelligence without natural language processing is impossible

Jiri,

1) Yes, human knowledge base is limited.
2) Children's knowledge is limited even more.
That's why our children's intelligence is essentially weaker than adults' intelligence.
3) ARTCOM's knowledge would be EXTREMELY limited (because of very poor communication channel).
That's why ARTCOM's general intelligence would be EXTREMELY weak.
4) If you suggest to use new language for communication with ARTCOM then it's not user friendly already.
What is more important: knowledge base on the internet is practically not available for ARTCOME.

5) Yes, ability to understand available data is critical.
That's exactly the direction to dig in the General AI research.
And this is directly relates to natural language reading problem.

6) You can measure my intelligence even if I have no external tools. I still have knowledge database in my head.
What is important here: my intelligence with external tools (Google/Internet/Other experts/...) would be essentially higher than my intelligence without external tools.

BTW, John Searle did big mistake in his Chinese Room Argument.
Student with dictionary is different system than student without dictionary.
No wonder that "student with dictionary system" can speak Chinese, but "student without dictionary system" cannot.
Intelligence level of these two systems also differs.

7) Difference between "human search" and "intelligent calculator" is huge:
Humans already have huge knowledge base of possible solutions. Current problem activates the most relevant solution in the knowledge base.
Then human tries to apply these most relevant solutions and check the results. Solution which brings the best result is selected.
(The quality of the result is evaluated against human's goals).
In addition --- this selected solution, relevant information, and relations between problem and solution are added to the knowledge base for future use.

"Intelligent calculator" behaves in different way. Calculator doesn't use knowledge base, because it doesn't have solutions knowledge base.
Calculator doesn't have the ways to find the solution in the knowledge database either.
Calculator just applies some calculations to input data and returns the result.
You are building calculator.
Yes, you are going to implement small database based on past experience with user stories.
But the key to efficient general intelligence is HUGE database, not small one.

8) If you cover essential amount of concepts and relations between concepts that you will quickly get big database.
Why do you think it would be small?
It would be small only if your input channel is inefficient (like special language which wasn't used before).

9) Little data in knowledge base is absolutely not enough for general intelligence!
It could be little data in the question, but knowledge base has to be HUGE.

10) Since your system with small knowledge base would be inefficient --- nobody would put data in your database. Therefore the project would die.

11) Contrary to General Intelligence, HTML prototype perfectly works with one page. That's why some people learned HTML.
But even in case of HTML it took years before HTML became popular.

12) Your special language has other disadvantages aside of "nobody use it" problem.
It is less efficient than natural languages in supporting "General Intelligence Thought Process" and "General Topic Conversations".

13) If my long term memory doesn't accept any new knowledge then:
I still will be able to apply solutions from my huge database to new problems which are similar to old problems.
But this would be possible only because I already have HUGE knowledge base.

Worst part in "read-only memory" is that deliberation would be impossible.
Adapting to the changes in the world would be impossible.
Improving solution solving skills would be impossible.

Too many problems. Even with HUGE knowledge base.
Without HUGE knowledge base already in place there would be practically no intelligence.

14) If you want to keep things as simple as possible --- don't invent your own language. This is not just useless it's harmful for the system.

15) I hope I saved your research/development time
:-)

1 comment:

Jiri Jelinek said...

Dennis,

A little clarification: My AI system is called ARTEX (=Artificial Reality Explorer). It consists of several modules. ARTCOM is just a communication module (responsible mainly for the I/O format).

I'm thinking about a new definition of intelligence for the purpose of my AI research. You are not gonna like it but here it goes:

Intelligence = Ability to identify existing useful resources in the scope of available resources and use it in order to achieve a goal.

I may find a better way how to say it later.
It implies that if you cannot achieve a goal because the solution requires an information which is in that particular scenario totally out of your reach does not necessarily make you "not intelligent." In fact, you can still be very intelligent. The amount of useful data affects the visibility (not existence) of your intelligence. Without the data, you are still more than a piece of rock, even though your problem solving abilities may be the same at the moment. So the question is "what you can do with useful data," not just "what you can do."

People have various definitions. That's fine. It just needs to logically fit their AI theories (something what helps them to move forward in the direction of their theory).

So my system can use a small amount of useful data and demonstrate that it's able to find relevant stuff and put it well together to find a solution (which still may require a long chain of thoughts and demonstrate the quality). Of course that small amount of data will not be sufficient for lots of other problems but that does not make the general AI algorithms invalid. Again, it just makes the intelligence less visible. The next step is straight forward: just adding more data = a clear way how to move forward.
With NL systems, on the other hand, there currently does not seem to be any clear way how to move forward (after decades of intense research). Let me know if you see some good ways.

If you say that ARTCOM is not user friendly then you can say that nearly all software written so far is not user friendly because we do not control it using NL. ARTCOM uses intuitive English pseudo-NL. It still goes through changes so the user-friendliness goes up and down ;-)-:, but I think it will be OK when it's done.

Yes, ARTEX 1.0 cannot learn from the Internet easily. It's the price I'm paying for the system's ability to well understand the input. A good understanding is critical so I do not really have a choice, unless I want to spend (probably MANY) years focusing on NL-related research. I'm more interested in problem solving so I'll try to get the AI working with pseudo-NL and then possibly focus on NL-to-My_pseudo-NL conversion. Hopefully, some new tools will make it easier then.

>5)

Are you digging? For how long? How far did you get? Look at the others? How far did they get?
See the McCarthy's problem demonstration from 1976 and similar articles. This is a bigger problem than many AI people think.
After you get the meaning from NL correctly (which still may take decades of research), there will be lots of other work to do. I would like to see some AGI results earlier so I'm taking a different approach. When you solve the NL, we can still benefit from each other's work. Good luck with that NL mess.

Most of your points are related to the Small/Huge KB issue. I think I addressed that above.

BTW since this is a public place, I think it would be helpful for the other readers to see the following part of one of my older responses:
>>ARTEX's KB can for example contain only 20 scripts (stories), 10 of
>>which (all together) contain clues on how to solve some particular
>>types of problems. Such problem will be given to the system to solve.
>>The system will have to find those 10 relevant stories and correctly
>>combine the data in order to find a solution. A human may know all
>>those 10 relevant stories and still might be unable to see a >>solution.
>>If all works well, the system is not gonna miss it.

Thanks for thinking about my system.

Sincerely,
Jiri