Hi, dudes! I have some delicious gift for all AI geeks!
The Ultimate Collection of Agent-Related Scientific Journals.
Warning! Newbies aren't welcome! It can really explode your brain!
Btw, I hope it helps young but ponentially outstanding researchers... It helps to bring closer the robotic apacplypse day...
Monday 3 November 2008
Agent-Related Scientific Journals
Posted by Lonely Bot at 08:41 1 comments
Labels: for AI geeks
Thursday 23 October 2008
Your First Steps in Muti-Agent Systems and Intelligent Agents
Well, we didn't touch today why you have so strange and perverted interest :) . Just short and pragmatic instructions for newbies.
1) First of all, learn Java programming language (particularly, Java SE 5 or 6). Today it’s the only programming language widely used in multi-agent area.
2) Then read the book M. Wooldridge, Introduction to MultiAgent Systems, Wiley, 2002.
It’s like a primer in multi-agent. You’ll see overall this area from a bird view. What you can and what you cannot do with agents? What problems can you solve? What applications are suitable for agents? What is agent’s brain and how can you implement it? And so on.
If you’ve survived by now -- it’s time to practice. It’s time to get our hands dirty with some actual coding.
3) Just download Java Agent Development Platform (Jade). It is the most popular and easily extendable agent platform I’ve seen. It’s open source project written in Java. So if you have troubles you can drill down right to concrete variable in source code :)
4) Read Jade manuals (i.e. JADEProgramming-Tutorial-for-beginners.pdf from /doc directory etc.).
5) Run examples and play with it. You can play with Jade tools or torment Jade in debugger of your favorite IDE.
You can use Jade in any Java IDE you like, just add necessary jars in your project but I prefer the Eclipse IDE for Java Developers.
6) And then start to code your first multi-agent system.
Of course, options in this steps are possible but strongly not recommended for the beginners.
Also there is a useful book which gives more solid practical coding information about overall Jade platform:
F.L. Bellifemine, G. Caire, и D. Greenwood, Developing Multi-Agent Systems with JADE, Wiley, 2007
If your need some additional theoretical background, you can read
S.J. Russell, P. Norvig, Artificial Intelligence: A Modern Approach, Prentice Hall, 2002.
If you interested in multi-agent aspects then look at
J.M. Vidal, Fundamentals of Multiagent Systems with NetLogo Examples, 2007.
Posted by Lonely Bot at 17:20 0 comments
Labels: for dummies
Tuesday 14 October 2008
AI Innovations Obstacles
Take the bird’s view at the whole picture of AI.
As for me the main obstacle in the path towards strong AI is the lost of interest from industry and the public. Clusters of neglected geeks are trying to make strange things working but vast majority of developers didn’t use and even understand it… Maybe it’s because of lost expectations of the early AI evangelists... or complexity of technology... or due to sectarian nature of any scientific society...
In any case we need to attract huge public attention and establish real connection with money-bringing industry. We need more fun and cool things around AI, and sometimes even more scandalous... But we must to be more close to ordinary people understanding.
Posted by Lonely Bot at 15:09 0 comments
Labels: breakthrough
Wednesday 8 October 2008
Are You Ready for Virtual Chicks Army?
Ben Goerzel widely known AI specialist wrote about great potential of virtual worlds as learning environment for AI. He wants to create baby or parrot or other pet AI’s that can learn and insert them into virtual worlds where human avatars can teach them. In virtual worlds like Second Life people could teach AI avatars not only primitive tricks, but also about general concepts like space, objects and even to talk. If any one of the AI avatars learned something new it could be transferred immediately to all of the other AI avatars. With millions of virtual world residents teaching AI avatars, they could rapidly acquire artificial general intelligence.
There is no doubt that he is right. Except one thing: who wants to bring up a baby or school a talking parrot? And who wanna have sex? Feel the difference!
Let’s imagine huge army of dolls which interconnected together. Successful behavior strategies will spread amongst them. And of course we'll kill loosers. It'll be the strongest fight for survival and reproduction.
Only real breadwinners will survive. What's the cruel nature of natural selection! :)
What kind of intelligence we can gather from this virtual sex? Of course it isn’t a language. :) Mainly it is behavior improvement but they can even improve their hot digital bodies. Dolls can behave like self marketers. They are trying to sell their behavior and digital passion in their digital eyes.
Posted by Lonely Bot at 13:58 0 comments
Labels: breakthrough
Sunday 28 September 2008
Can Virtual Sex Boosts the AI Innovations?
Hit the ground running. LonelyBot’s thoughts about the most interesting and sacred things throwing off all shame and decorum.
Definition of the “virtual sex”
What do I mean by “virtual sex”? First of all, check the Wikipedia for "virtual sex". Non penetrative sex? Virtual sex over telegraph, eh? Virtual sex enthusiasts? You, people, are very funny creatures. To be precisely, I mean virtual sex in virtual worlds like SecondLife.
How virtual sex looks like? Just check it yourself!
What’s the most impressive? :) Ok, ok… Let’s move on. Hey! Read next! Don’t watch videos again!
Analyzing the Use Case Scenario
Let’s examine standard use case. We have AI-controlled woman and human-controlled man in the virtual world. He wants the greatest pleasure ever experienced. She wants his money. And I want breakthrough AI :)
Standard sequence is the following. Man goes into the virtual brothel. Then he watches virtual strip dancers. He picks up a virtual girl and pays virtual cash. They perform virtual coitus. Ring down the curtain.
Analysis always means “separation of the whole into parts for better understanding”. Of course, we can divide virtual sex universe using various criteria.
By participants (I mean standard heterosexual act): man (human), woman (AI bot), environment (virtual brothel).
By interests of participants (by goals): satisfaction of sexual instinct (man), money (woman). “Man pays” is the current trend in sex industry. Pornography and virtual sex for women are rare case but technology and insatiable business changes everything.
By channels of conscious of participants: visual (movement of partner and other bodies), audio (music, sextalk, screams, moans, growling, etc) and text messages. With current technological level it’s impossible to have direct brain and muscles stimulation. And even body stimulation doesn’t become the mainstream. What a pity? But we have what we have.
Imagine you and me are the most genius AI developers. Well, we want to make artificial dolls for unbelievable men’s pleasure. Sex dolls with advanced AI. Let's name it SAI (sexual AI).
What does “virtual sex” mean from this point of view? It’s is a very special (к чему прилагательное относится?) and primitive media show (e.g. it occupies a few scenes compared to a whole film or a computer game). Client pays for entertaining behavior and atmosphere. And the main problem, which we should solve by our SAI system, is to make really cool sexy show and increase money outcome from a satisfied viewer.
Possible Applications of Sexual AI
First of all, sex industry itself can benefit from SAI. The more powerful AI we have – the more interesting, natural, human-like and variable behavior can be implemented with less development efforts at lower cost.
Next, AI community can use virtual sex as research testbed for next-gen AI. Virtual sex attracts huge men attention because of biological reasons of course. It consists of tight (but very restricted) interactions between partners. It's very well for automatic learning at all levels of AI, realistic movement, rich media of interactions, emotion simulation… All of it results in next gen digital actors. Let’s imagine unbelievable virtual sex in Hollywood movies!
Last, sexology and related sciences can use technology as instrument for sex studies. Advanced tools always enable to carry out more and more inhuman experimentation on people and Mother Nature.
Conclusion
Well, what we have in the end? SAI can bring more money and increase customer satisfaction. Also there is scientific interest in SAI improvement as well as SAI evaluation in sexual sciences. Let me guess it'll be very hot area! :)
Ok, we have a Question in our title. The answer is yes, it can. And no doubt it will.
PS. Any suggestions and comments at any time from anyone are welcomed! The only restriction is meaningless and stupidity.
Infinitely yours, LB.
Posted by Lonely Bot at 05:03 0 comments
Labels: breakthrough