At what point does the macrophenoma of a MAS appear as a swarm like behavior? Then in a large multiagent learning problem could the agents be viewed as swarms instead of agents? Can an algorithm be created to discover the swarm behavior of a large system composed of MARL agents?
Learning from/by Demonstration what are the limitations?
I’ve been reading up on this area a bit and I haven’t yet seen a paper detailing the shortcomings or limitations of this approach to AI. I’ll read more to see if someone has researched this…
Adaptive websites
In AI terms the user agent’s environment when browsing the internet is their browser. So, if the browser could construct a model of the user’s behavior, when the user enters a website the model that has been constructed could be sent to the website in order to construct a personalized layout etc. Websites could possibly manipulate the user based on the provided model. I think that google already pretty much does this already with its search, but why stop there?
Usually when I go to a restaurant’s website I am looking for their menu. The browser could learn that is what I click when I go to those types of sites.
I could create a website that acts as a proxy to websites and could dynamically change the website for them! So, people go to my website to view other websites. I would create a plugin for chrome/firefox that would develop the profile when they search regular websites when they aren’t…
No, forget the website, just make the plugin! When they look at a website with the plugin running the website content will be rearranged and possibly home page will be displayed! Like in my personal example above when I go to a restaurant website it could show the contact info and menu!
This would mean that the user’s profile could be stored in the cloud or on there computer. If in the cloud I could compare the graphs of users visits and when a user goes to a website they have never been to, people that have that are similar to this person could help to reformat it! Sort of like Amazon or Netflix.
Of course the user could turn off the changing of the website and leave on the learning of the user.
We already have plugins to remove the ads why not do this? Is it possible to do this quick enough that the user doesn’t experience too much of a lag?
F-statistics for NEAT
To determine the number of genomes within the population automatically could use F-statistics (http://en.wikipedia.org/wiki/F-statistics). It doesn’t seem like it has been explored in genetic algorithms, rather for human population studies. Another way that subpopulations could be culled is to use the inbreeding coefficient (http://en.wikipedia.org/wiki/Coefficient_of_relationship).
Culture
Could I use a model or description of the culture of an agent to determine how well it would adapt to working with others in a group. Sort of a similarity index of cultures. This would help an agent to decide if he wants to collaborate with a group. Sean Luke has published two papers on how the use of culture has helped genetic programming (search culture http://www.cs.gmu.edu/~sean/papers/).
Automatic Subject Line Generation
I sometimes have a hard time thinking of what would be a good subject line for an email. It seems like the subject could be automatically generated based on the message with a combination of a word cloud, thesaurus, and learning based on the user’s inbox.
Skill Decomposition
In a cooperative multi-agent learning environment if a new agent enters the system the others can teach the new agent the high level task which it can then decompose to learn other useful skills. In the single agent case an example would be teaching a robot to get from a to b which requires going through a door and the robot could learn the simpler activity of walking through a door by exploring its options space. I think maybe task decomposition type algorithms could be augmented to decompose the option space.
Importing posts
Well I was able to export the wp_posts table from phpMyAdmin at freehostingcloud. Then I had to change some of the id’s in the exported sql so that they didn’t conflict with those that were already used. I imported the data into wp_posts. Then I ran this sql
UPDATE wp_posts SET guid = replace(guid, ‘http://drew.freehostingcloud.com’,’http://drew.heliohost.org’);
so that the posts would actually work!
Old site
Hi,
I used to be at drew.freehostingcloud.com and I had to switch because for some reason freehostingcloud took my website and a bunch of other people’s sites on Sunday. I even got on the forum to see what was up (http://www.freehostingcloud.com/forum/thread-can-t-access-website). Thankfully I was able to still access the backend and get all of the data. So, I switched to heliohost. They are much faster, but occasionally I get a 500 error and I just have to refresh the page.
Research Idea
“Less work has been done on formal methods of decomposing tasks (and behaviors) into subtasks (sub-behaviors) appropriate for multi-agent solutions, how agents’ sub-behaviors interact, and how and when learning of these sub-behaviors may be parallelized.” (An Overview of Cooperative and Competitive Multiagent Learning)
I think that this would be an interesting problem to explore. I found that the problem of decomposition is also a major problem in engineering. They have come to the point were the number of variables in a problem are in the 1000s and an expert is not always available. Based on the paper, The State of Problem Decomposition in Engineering Design, solving this problem is useful. So, I am going to research how others are solving the decomposition problem and combine that with multiagent learning to try to solve real life engineering problems.