I have to give a short lecture in some artificial intelligence topic i choose and i'm looking for an interesting one. we've already had "self organising maps" "chatterbots" and "ai in fpp games" - to give you a taste of what i'm looking for. i've been searching the internet for last two days and i just can't find anything special. any ideas?
+1
A:
The Netflix Prize seeks to substantially improve the accuracy of predictions about how much someone is going to love a movie based on their movie preferences.
Paul Dixon
2009-03-12 22:50:08
You took my idea! I've been working on the prize on and off for a little while. =)
Jess
2009-03-12 22:53:50
Million dollar prize +1
bias
2009-05-29 13:42:39
+1
A:
Hmmm.... I was originally going to say Dr. Sbaitso, but apparently that falls under 'chatterbot' =)
So instead, what about AI in robotics? Maybe highlight the AI necessary to make something like Asimo work? Or the rather bizarre, but believable, 2-year old humanoid robot. Lots of interesting AI projects in Japan right now.
Jess
2009-03-12 22:52:39
A:
I guess netflix is working with an advanced version Collaborative Filtering. And I guess we can call that some kind of AI.
Emil C
2009-03-12 22:54:48
+3
A:
There is some interesting AI research in Poker Bots. The University of Alberta Computer Poker Research Group is doing some interesting opponent modeling.
Colin
2009-03-12 23:00:38
A:
Robotics...such a vast area made up of so many different research groups that you could probably talk for days on just one of them.
Some examples:
- Walking - how can we get a robot to maintain balance on uneven ground: Current applications: humanoid robots.
- Speech Recognition - not only looking up what we say, but understanding and acting on that speech - Current applications: Helping the disabled retrieve objects, various search and rescue scenarios, User interface design.
- Emotional response - Getting robots to respond and react to our emotions: Children's toys, helping autistic people communicate
There are many more.
Jamie Lewis
2009-03-12 23:03:00
A:
Something that have always interested me is the "gamer" topic:
"When do the enemy AI get to smart and the user get bored?"
Other then above, how about the Terminator AI:
My CPU is a neural net processor, a learning computer. The more contact I have with humans, the more I learn.
// The Terminator
I guess The Arnold is a neural network of some sort.
Daniel T. Magnusson
2009-03-12 23:05:54
+4
A:
First narrow down what genre of AI you're interested in - the term "AI" is sorely over-generalized. There is human AI - seeking to make computers behave human-like - under which eg video game NPC behavior would fall, robotic facial response (being studied at MIT), the Netflix problem, etc. There's also computational AI - solving difficult but well-defined computational problems - such as using an A* (A-star) algorithm for pathfinding, or using a genetic algorithm to solve nonlinear equations. Of course, there's plenty of crossover.
Speaking of, genetic algorithms are always a fun topic, widely researched, widely applicable, and always making for a good presentation since the basic concepts are very easy to grasp. Have you looked into those?
Not Sure
2009-03-12 23:10:52
generic algorithms are a regular lecture i have at my course at the university. and neural networks as well. i'm looking for something "fun", any kind of ai actually
agnieszka
2009-03-12 23:13:33
Thanks, not only is AI "sorely over-generalized" the second term 'intelligence' isn't well defined ...
bias
2009-05-29 13:41:38
+1
A:
Demos are always dazzling and serve to illustrate what you're trying to say. Here are some:
Breve Simulation Environment uses genetic algorithms to create walking creatures out of spastic building blocks.
EvoCAD also uses genetic algorithms, but this time to build Lego structures given a set of criteria and goals.
This site has a bunch of Java applets to demonstrate how artificial neurons and backpropagation networks work.
OCHRE is a demonstration of a neural network for optical character recognition.
GamesCrafters uses game trees to solve a variety of abstract, two-player games such as tic-tac-toe, connect-4, and mancala.
PBS has video of their older Scientific American Frontiers program. Some of the episodes focused on AI, including these:
Robots Alive: See the "Almost Human" segment.
Life's Really Big Questions: See the "Robot Independence" segment.
Barry Brown
2009-03-12 23:23:57
The Mona Lisa example isn't GP. You are adding to everyone's confusion by calling it that. -1
bias
2009-05-29 13:36:46
@Noosphereious: You should have at least clicked on the link before leaving your comment.
J.F. Sebastian
2009-05-30 00:38:42
+1
A:
I'm reading Jeff Hawkins's "On Intelligence" right now. It's changing the way I think about AI.
duffymo
2009-03-12 23:45:42
Let's see - the title doesn't give you any hint? What's more pertinent to AI than how the brain works?
duffymo
2009-03-15 21:45:06
""He who asks a question is a fool for five minutes. He who doesn't, stays a fool." - how does this square with your comment?
duffymo
2009-03-15 21:46:02
WHERE do you suggest any topic? you've just written about a book you read - this is not an answer to my question
agnieszka
2009-03-18 17:23:53
+1 for the Book on AI by the inventor of Palm. A vague answer for a vague question is fine in my book.
bias
2009-05-29 13:44:39