You'll get invited to our Meetups as soon as they're scheduled!
Review algorithms, statistics, SVD, SVM, etc. Maybe Octave or SciPy.
And PLEASE sometimes slightly further south in Santa Clara County for those of us using Outreach. They do go to Palo Alto.
I had a bit of prior experience about AI. Now I think it may provide a means to my problem--helping to diagnose and to suggest fixes for solar power plants.
I've defined some fault & conditional probability matrices. I need to know if there are software that I can try to build inference engine incorporating domain knowledge, and a warehouse of measurement data. The preferred environment is L.A.M.P(HP), Java.
Most AI researchers are impressed when you get an AI to use an AI, or get an AI to use a different AI to train a 3rd AI... How deep has it been done? What are some effective systems that use AI recursion that way?
Let's have a demo day: Instead of giving someone the floor for an entire meetup; those who are working on AI projects could bring in a laptop and run a 15-20 minute demo. Nothing needs to be pretty or shiny.
Unmoderated group discussion about the real absolute limits (a la Turring and Godel) on intelligence and intelligent systems. Are these limits different than the limists that have already been identified for computing? What are the implications for the future of AI and evolution itself?
Working Example /Demos/Tools of AI application in Social Media.
From PhD students to hackers to corp engineers, there's no shared PC like platform for AI. Thus like a broken record, we keep starting over an over at the same place. Certainly PHD students spend 2-3 yrs building a robot to study AI. Williow Garage has project to share a Python based platform for robotics, would be cool to learn more about the project and it AI intent.
This is a game some of us can play and others can watch and learn.
Each round, maybe 20 people take 2 minutes each at the computer, save the file, get up and let the next person continue your work.
Bring a few computers (or just 1?) and a repeating timer.
It should be simple, like an AI in Javascript that runs in a web-browser, to predict where points are in a 2-dimensional space, given other points, or various other simple math tasks.
Each round of the game, we can start with an old AI from a previous round or write new code.
Before each round, we are given the task, but not the specific data points until the AI runs. Does the AI work?
In the last round, the task is the same style as the previous rounds, on average, but a little different, and nobody (except the meeting organizer) should know what the last task will be. Our final challenge is to assemble the various solutions to previous problems into 1 AI that will solve that problem without knowing in advance exactly what the problem is.