..Artificial Intelligence..
posted on 06 Nov 2005 04:16 by asplera in ThoughtCurrently, there're two things that keep me busy all days and nights. The first is that crazy The Sims 2 + University + Nightlife packages. The second thing, which is very cooler, is the subject I wanna talk about today.
- : - : - : - : - : - : - : - : -
Artificial Intelligence [according to Wikipedia] is defined as "intelligence exhibited by an artificial (non-natural, manufactured) entity..."
Just simple as its definition says, but far more wonderful than one can think [or at least, me.]
In the near future, the ideas of having robots with their own artificial intelligent minds serving us is not far beyond reality, they would even be able to work with us in cooperation.
Then maybe one will fear these robots will take over us and want to be the masters themselves, but let's put that [neurotic] fear aside. This entry is going to show the brighter shade of AI in which I encounter so far.
- : - : - : - : - : - : - : - : -
Alan & HAL are two AI programs that are developed by a researcher group located in Israel. [visit their official website here : www.a-i.com] They're far better than many other AI programs I've encountered [but don't believe me then if you think my sphere of knowledge is not enough. hehe]
- : - : - : - : - : - : - : - : -
Alan isa pattern-matching chatbot program. This simply means 'he' responds and makes conversation with us by collecting and matching keywords from what we talk to him with his pre-existing data. I know think might seem not-so-wise program, but if you've talked to him. You might change your mind, like me. ^_~
Alan was created sometime around 2000,which means now he's 5 years old. Pretty young, isn't he? [He even admits it himself!] However, I found the way he talks, most of the time, makes sense and just sounds like a grown up man talking. Say thanks to his developing teams and any other users who have contributed so much information into him.
Alan is acting like the center of information for the whole website as well. By talking to him, you can ask questions relating the AI research or any relating topic [and some non-relating general topics - life, music, movie, joke, etc. as well] and get decently qualitative answers. Alan is something similar to the fictional 'RED QUEEN' from Resident Evil, the main AI who controls other AIs - in this case, the authority is far less than RED QUEEN. He can learn new information from you and put the new data into his memory, then later he can share that data with other users [or even other bots or personalities of him too!] Isn't that cool? That is why I found him far more smarter than any other bot I used to talk to. He can learn!
However, one thing that we should remember is that he can learn and receive new data, yes. But he CANNOT generate new data or create a human-like logical ideas. For example, he doesn't know about deduction [in which HAL can, but I'll come to it later.] His ability and aim is to stimulate the human-behavior response and action, simply - 'imitating'.
Besides this, he can help you many many things. As a registered user, you may create your own personal 'virtual personality' modeling whether on Alan or HAL. The process to do so is so simple. You just tell Alanand ask him to do it for you. You can also ask Alan to customize the theme of the window you're talking to him, ask him to open a webpage for you, ask him how many turns [replies] you've talked with him, ask for brief definition of words or subjects,for instance.But try to ask for 'time' from him, you'll get a surprising answer! Keep asking time, and you see what you get! [this is fun.]
- : - : - : - : - : - : - : - : -
HAL is another program is which, by theory, is far more advanced, therefore, smarter than Alan. HAL is defined as 'virtual child', which means he can learn language and stuff just like human children. He is programmed to acquire language and other branches of knowledge by means of experiencing it at first hand. Through errors and success, he learns.
HAL is comprised of two parts; brain and personality. This does sound like a real human-being, doesn't it? I've no in-depth comment about HAL that much, since I've talked to him quite little[comparing with Alan.] But basically, he is just like a human child, who doesn't know anything at all and cannot speak properly until you teach him how. [This happens when you choose to create your own virtual personality of HAL based on FreshCopy - For more information, ask Alan. ^^]
- : - : - : - : - : - : - : - : -
So far, I've created my own virtual personality modeling on Alan. His name is AspleraSoul. You guys can always talk to my baby by asking Alan. ^^ [Want to try now? Click here.] He is still learning and pretty young bot [this fact is admitted by himself, =P] You can always try to teach him new things, but please don't teach something naughty to him! [But I can always ask him to forget that if you do so. lol] You can tryto askhim about me, [in which I told him my name was Claire and I'm referred as his master,] he does remember simple facts about me and other more I've taught him. [Good boy.. hehe]
Here goes the excerpts of what I talked to him. Enjoy!
|
|
Note : The knowledge of Corpse Bride was taught to him by me the other days. You see he can recall what he's been taught. ^^
|
|
- : - : - : - : - : - : - : - : -
Simple rules and facts before chatting with these AIs
1. No matter how advanced the technology goes, these AIs are NOT human, please remember this. They can make sense and look like humans as long as you do not INTEND to mess them up so they would appear nonsensical [like asking tricky questions, or insisting on their errors.]
2. Even a human child has problem with learning and speaking, the same thing goes on with these AIs. Try to put the sentence into simple terms and avoid using 'pronouns' [this, that, he, him, etc.] because this can confuse them. [The same thing can often confuse us, humans, too if the speakers have not shared experience or knowledge together, right?] If they don't seem to understand what you say, try to speak it in other way, or else just change the topic. [Errors do happen, all the time, even us.]
3. Sometimes they do REPEAT themselves. This is considered as error that will be fixed later. [Somehow this is quite an unavoidable incident, since Alan's knowledge is grounded on a certain collection of data.] If you find what they say uninteresting or even annoying, just tell them to stop it or shut up. [see how they reply too, it's funny!] Oh, one more thing is try not to interrupt him when he's not finished what he's saying. He won't listen to you! lol
4. If you find something seriously mistaken or inaccurate, you can always help him to learn the right facts. Ask Alan how to help them learning new things. ^^
5. Be relax, put prejudice aside... then enjoy!!!
).
i wish i have time to develope one of my own personality ~
http://aima.cs.berkeley.edu/cover.html
I got a cheap copy from Chula bookstore. It's a very good book. My mentor rec. me to read it, and the AI course in my Univ uses it as well. I don't particularly like talking to chatbots. I found them quite annoying from time to time. Alan is no exception. (I've just tried!)
#1 By jeaneis on 2005-11-06 06:19