According to the creators of 20Qs, "20Qs A.I. is a neural network-based artificial intelligence; it learns by repetition. When people play objects that the A.I. already knows, it adjusts the weightings within the neural network. If 20Q performs poorly with a particular object and you play that object multiple times, you will see the A.I. learn and get better at guessing that object." The program has been running and building its system for over 17 years now.
Here is the link: "20 Questions" Game
Instructions:
1. Click "Think in American"
2. Enter your gender and age (optional)
3. Choose "Classic 20Q" (if you want to play other games, you can, but to keep it consistent for my project, choose Classic)
4. When the program gets the answer right (or wrong), before you click "Right," Copy & Paste the page into your blog post so myself and others can see the "thought process" of the program.
5. Answer the response questions below:
a. Did the program get the right answer, and after how many questions? If id didn't get it right, what was your word and what do you think are the reasons the program couldn't guess it?
b. Do you believe in artificial intelligence/ Can it exist?
c. This 20 Questions program has grown its database each time it is played (it stores the answer and information in its memory) for 17 years. Would you say that the program is "learning"? Or in general, is it possible for a machine to learn, without being programmed to do something?
Thanks very much for your cooperation!
6 comments:
It got it right!
Q19. I am guessing that it is a pineapple?
Right, Wrong, Close
18. Does it come from a plant? Yes.
17. Is it human powered? No.
16. Does it have a handle? No.
15. Is it straight? No.
14. Would you give it as a gift? Yes.
13. Can you drink from it? No.
12. Is it larger than a pound of butter? Yes.
11. Do you make pie from it? Yes.
10. Is it found on a desk? No.
9. Can it be used in a pie? Yes.
8. Is it round? Sometimes.
7. Does it have a shell? No.
6. Does it have a sense of smell? No.
5. Is it normally planted in gardens? No.
4. Is it hard? Yes.
3. Does it break if dropped? No.
2. Do you hold it when you use it? Yes.
1. It is classified as Vegetable.
I chose "truth" as my word, and remarkably the program got it right within twenty-three questions. I'm impressed.
23. I guessed that it was truth? Right.
Is it spontaneous? Unknown.
22. I guessed that it was a voice? Wrong.
21. Is it used for communications? Yes.
20. I guessed that it was an idea? Wrong.
19. Is it used in travel? No.
Can it be stolen? Unknown.
Is it used by a baby? Unknown.
18. Is it used by the police? Yes.
17. Would you be lost without it? Yes.
Is it man made? Unknown.
16. Can you sit on it? No.
Can you control it? Unknown.
Can it help you find your way? Unknown.
15. Is it usually warm? No.
14. Is it considered valuable? Yes.
13. Does it use electricity? No.
12. Is it comforting? Yes.
11. Does it move? No.
10. Is it flexible? No.
9. Does it sweat? No.
8. Does it have writing on it? No.
7. Is it hot? No.
6. Is it a pure-bred dog? No.
5. Does it get wet? Irrelevant.
4. Do most people use this daily? Yes.
3. Does it have a hard outer shell? Irrelevant.
2. Does it usually have four corners? Irrelevant.
Would you use it in the dark? Unknown.
1. It is classified as Other.
to answer the discussion questions... I do believe that artificial intelligence exists simply by the definition of artificial intelligence. In my opinion AI means that artificially produced by humans to appear/function as intelligence but not actually intelligence.
I would not say that the program is learning. I think it is retaining information but learning requires understanding and actual meaning which the computer program does not have.
I don't think that artificial intelligence can ever exist because i believe that intelligence is restricted to only a human characteristic. In relation to rachel's project, the ability to think, create, have emotions and evolve are human traits that cannot, no matter what we invent, be replicated or programmed into a machine.
Whoops, I forgot to answer the other two questions before.
I believe that artificial intelligence does exist--just the name "artificial intelligence" implies a replication of human-like decision-making, regardless of its accuracy. As long as a program attempts to make decisions calculated off how it understands human thought processes to be, it has artificial intelligence. Of course, there are varying levels of complexity: the computer player in a game of Pong is simpler than a modern robot.
Is it possible for a machine to learn without being programmed to do something? This is a difficult question because I continue to believe that our brains are very complex mechanisms similar to computers. Humans learn how to do things by instinct, such as walk, speak, and build a fire. Are we programmed to do these things? I would argue that since we have the capability of learning such skills, we are programmed to learn. If something isn't programmed to do anything, then it cannot learn.
An extra issue is what "programming" means. In biological-environmental terms, any form of selection is a form of programming, hence evolution, which brings us more complex creatures. In this way, it's almost impossible for something not to be programmed. Whether the programming is done by humans or not should be totally irrelevant.
It was hailed as the most significant test of machine intelligence since Deep Blue defeated Garry Kasparov in chess nearly 20 years ago. ai security
Post a Comment