I would definitely agree with most of the questions people have on there. A lot of the questions people proposed to pass the turing were based on factual, and I would agree. Computers cannot express a definite opinion. In other words, if you were to ask a typical bot a question that requires opinion, the computer would most likely respond with witty/clever remarks. Different remarks than last time. A lot of the questions people had to not pass the turing test included a lot of subjective questions, which would obviously work because while they COULD express opinions, they're not usually opinions formulated by logic and reasoning. Most likely, the opinions are formulated via pseudo-random.
Tuesday, September 27, 2011
Blog# 8 - Turing Test
Pass turing test:
1. What planet do we live in?
2. What's two plus two?
3. What year are we in?
4. What color is the sky?
5. What time is it?
6. Are you a computer?
Not pass the turing test:
1. What is the square root of 3?
2. what is the slope of the line y = 2x + 5?
3. what is 523 xor 9385?
4. What is cos(5)?
5. How is your day?
6. If I live in Colorado and not Utah, Nebraska, or Arizona, where do I live?
7. What do you think about Obama?
8. Are you a democrat or republican?
They were both equally difficult to think of. On one hand, to pass the turing test, you want the questions to be as simple as possible or at least really obvious. Both humans and machines with a fair artificial intelligence should be able to answer these questions. However, for not passing the turing test, you want the questions to be a bit difficult. Most humans won't know the square root of 3, and some humans aren't familiar with the binary operation xor. Since computers can compute numbers fairly quickly, they should give an answer quickly, it should be pretty obvious tell who is and who is not a computer. Question number 4 is a math related question, but somewhat ambiguous. Trigonometric functions can be computed using degrees or radians, but computers would assume radians since that's its default unit processor-wise (FCOS instruction in the x86/x64). A human would probably ask for degrees or radians, or probably give you both answers. A computer might only give one answer, in radians, unless otherwise stated. Moving on, a computer probably won't know the slope of "y = 2x + 5" either because they're not created with the knowledge of algebra in mind. While artificial intelligence may be good at reading and understanding grammar, it's not always perfect. This is why I included question number 6 in not passing the turing test. The answer is given there, but the computer might interpret that little differently. For the final two questions, I used subjective questions because a computer might not be able to give a proper response. In the end, it all depends how the artificial intelligence is programmed.
1. What planet do we live in?
2. What's two plus two?
3. What year are we in?
4. What color is the sky?
5. What time is it?
6. Are you a computer?
Not pass the turing test:
1. What is the square root of 3?
2. what is the slope of the line y = 2x + 5?
3. what is 523 xor 9385?
4. What is cos(5)?
5. How is your day?
6. If I live in Colorado and not Utah, Nebraska, or Arizona, where do I live?
7. What do you think about Obama?
8. Are you a democrat or republican?
They were both equally difficult to think of. On one hand, to pass the turing test, you want the questions to be as simple as possible or at least really obvious. Both humans and machines with a fair artificial intelligence should be able to answer these questions. However, for not passing the turing test, you want the questions to be a bit difficult. Most humans won't know the square root of 3, and some humans aren't familiar with the binary operation xor. Since computers can compute numbers fairly quickly, they should give an answer quickly, it should be pretty obvious tell who is and who is not a computer. Question number 4 is a math related question, but somewhat ambiguous. Trigonometric functions can be computed using degrees or radians, but computers would assume radians since that's its default unit processor-wise (FCOS instruction in the x86/x64). A human would probably ask for degrees or radians, or probably give you both answers. A computer might only give one answer, in radians, unless otherwise stated. Moving on, a computer probably won't know the slope of "y = 2x + 5" either because they're not created with the knowledge of algebra in mind. While artificial intelligence may be good at reading and understanding grammar, it's not always perfect. This is why I included question number 6 in not passing the turing test. The answer is given there, but the computer might interpret that little differently. For the final two questions, I used subjective questions because a computer might not be able to give a proper response. In the end, it all depends how the artificial intelligence is programmed.
Monday, September 26, 2011
Blog #7 - Artificial Intelligence
What is intelligence? Are computers intelligent? Why or Why not?
Intelligence is the ability to make decisions and apply skills to situations. In other words, intelligence consists of using reasoning and logic. Computers do not posses any actual intelligence because they are told exactly what to do; they don't apply their own reasoning and logic. A computer is only able to make calculations and comparisons with numbers, and therefore, do not posses any actual intelligence.
Computers, however, are able to replicate intelligence. The term for this is "Artificial intelligence". Most artificial intelligence implementations are used to perform one function. For example, there exists artificial intelligence in video games in Non-playable characters(NPCs). If the NPC is an enemy, its primary goal would be to attack you. This would require creating a path to you, the player, and attacking you. Simple artificial intelligence. The artificial intelligence for these bots consists of nested IF statements and a few variables in order to make judgement based on a specific condition.
In the case of chat bots, which are a form of artificial intelligence, they posses one primary function: to provide relevant responses. These chatbots read human inputs and create a response to them. Sometimes, these responses are pseudo-random (Because there's no such thing as ACTUAL random in computers. Randomizing functions are based on a seed, usually based on time, and then performed the modulus operation along with a few algorithms thrown in to change the initial seed value). They take account of certain variables, and most likely, has a database of responses to fall back on. They are either static responses or dynamic responses, which are formatted appropriately to the initial input. Of course, these responses have to be relevant for the most time, otherwise, the responses will obviously be formulated by a computer. It's almost difficult to provide accurate responses because there are so many variables to take account of, but we as humans are getting there.
Computers, however, are able to replicate intelligence. The term for this is "Artificial intelligence". Most artificial intelligence implementations are used to perform one function. For example, there exists artificial intelligence in video games in Non-playable characters(NPCs). If the NPC is an enemy, its primary goal would be to attack you. This would require creating a path to you, the player, and attacking you. Simple artificial intelligence. The artificial intelligence for these bots consists of nested IF statements and a few variables in order to make judgement based on a specific condition.
In the case of chat bots, which are a form of artificial intelligence, they posses one primary function: to provide relevant responses. These chatbots read human inputs and create a response to them. Sometimes, these responses are pseudo-random (Because there's no such thing as ACTUAL random in computers. Randomizing functions are based on a seed, usually based on time, and then performed the modulus operation along with a few algorithms thrown in to change the initial seed value). They take account of certain variables, and most likely, has a database of responses to fall back on. They are either static responses or dynamic responses, which are formatted appropriately to the initial input. Of course, these responses have to be relevant for the most time, otherwise, the responses will obviously be formulated by a computer. It's almost difficult to provide accurate responses because there are so many variables to take account of, but we as humans are getting there.
Thursday, September 22, 2011
Blog #6: Web 2.0 Presentations
StumbleUpon (Andrew Jensen) Allows you to bookmark pages and browse pages based on your interests
Digg (Nathan Kaplan) Allows you to post links to websites and reiew them.
The difference between both those sites is that Digg is less based on interest. StumbleUpon allows you to browse through your interests.
Google Docs (William Kelemen) allows you to store documents.
TagCrowd (Marcellus Mershon) Is same as wordle, allows you to blog, and receive immediate feeds via RSS and Atom.
Google Docs was not a word cloud. TagCrowd can create word clouds and Google Docs cannot.
Google Pages (Alex Gaisford) allows to make lists and even more than just lists
Remember the milk (Dan Albrite) Simply just allows to make lists
Google pages allows for more complex creations as opposed to remember the milk where it is simplistic.
Digg (Nathan Kaplan) Allows you to post links to websites and reiew them.
The difference between both those sites is that Digg is less based on interest. StumbleUpon allows you to browse through your interests.
Google Docs (William Kelemen) allows you to store documents.
TagCrowd (Marcellus Mershon) Is same as wordle, allows you to blog, and receive immediate feeds via RSS and Atom.
Google Docs was not a word cloud. TagCrowd can create word clouds and Google Docs cannot.
Google Pages (Alex Gaisford) allows to make lists and even more than just lists
Remember the milk (Dan Albrite) Simply just allows to make lists
Google pages allows for more complex creations as opposed to remember the milk where it is simplistic.
Tuesday, September 20, 2011
Blog # 5 - Web 2.0
For those who don't know, Web 2.0 is associated with social collaboration and information sharing. For example, sites like Facebook, StumbleUpon, Twitter, Wikipedia, and Digg are all examples of Web 2.0 applications.
Anywho, as for the social bookmarking site I found, it's http://www.diigo.com. I would've done StumbleUpon, but two others were doing it.
As for Diigo, my first thoughts at looking at it were that it has a user-friendly interface and is rather appealing to the eye. When I signed up, I was given four choices.
The first link is "My Library". According to Diigo, My Library is a "central place to store all your bookmarks, highlights, notes, pictures". Seems simple enough, and it even includes a toolbar for your browser for anyone who wants easy access to their bookmarks.
The second link is "My Network", which is just the people in your network or. You add people and you are able to share and contribute bookmarks. It also shows updates among the people in your network. It gives you options to find people on Diigo based on interest, email, or even the option to invite them. At the bottom is a list of featured users who might share the same interests as you.
The third link is "Groups" which "provides a ground-breaking collaborative research and learning tool that allows any group of people to pool their findings through group bookmarks, highlights, sticky notes, and forum.". You can create groups, interact with groups, or even create a "group knowledge repository". On the bottom of the page is a list of featured groups that you can join.
The fourth and final link is "Community", which is lists resources of community driven material. The page is titled "Hot bookmarks", so I assume it's the top liked/favorited/whatever resources on Diigo.
Overall, the site seems *really* user-friendly and has a variety of resources that a typical social bookmark user would use. Seems like a very connected site and I'm sure people who enjoy bookmarking would find the site of quality. I wouldn't really know, I don't use social bookmarking sites; I just did as asked as part of the assignment.
QUESTIONS:
1. What are the differences between ranking based and social bookmarking search engines?
Ranking based search engine are based on a global ranking. In other words, the results are based on what everyone else around the world has been searching. It is usually ordered in descending order based on popularity. As for social bookmarking search engine, it is more personalized. Social bookmarking search engines take account to what you have been searching for and displays results based on it.
2. What issues might there be with creating accounts online?
On the user end, I can only think of a few. For one, the most obvious one is losing your password and, as a result, end up losing all your bookmarks and social networking. Another one I can think of is leaving your account logged into a computer and another person can come in and tamper with it. Another issue is that once you create an account on social networking sites, your data may be available to the general public, which in that you may want to use the privacy settings the site offers, if any.
Alright, I think this is it. I just realized I can use this as my presentation as well. I think I *MIGHT HAVE* overdone it with this blog entry ):
Monday, September 19, 2011
Blog #3 - Project feedback
I think we could have had more computers to compare it to since the project guidelines specified four computers. We could have also been a little bit organized in terms of who gets to talk. Other than that, I think the whole presentation went pretty great.
Thursday, September 15, 2011
Blog #2 - Shared and Thought about it
My partner Andrew basically said the same thing I said except that he just simply stated that they computer.
Blog #4 - Internet Use
Three things I use the internet for media, for online gaming, and for researching various subjects (Such as computer programming, since it's what I mostly do).
My elbow partner's uses for the internet isn't really different as we found out that we pretty much do the same things online.
My elbow partner's uses for the internet isn't really different as we found out that we pretty much do the same things online.
Monday, September 12, 2011
Blog #1 Computer, computer, what are you?
A computer is a device that is able to perform calculations, such as arithmetic and binary operations, as well as process inputs, in the form of instructions, and outputs to other hardware.
Subscribe to:
Posts (Atom)