:: Monday, May 26, 2003 ::
:: Sunday, May 11, 2003 ::
I finally received a letter in the mail confirming my admission to the University of Southern California. I was accepted as Masters student. I was rejected as a Ph.D. student, but all the credits I earn as a Masters student can be applied towards a Ph.D. if I get admitted later at the same school.
It certainly was a great deal of stress on my part. They were the last school to make any notification about my status. I was calling the department office about 3 times a day trying to get into contact with someone. No one seemed able to answer the phone, and when they did, I got promises of information that never materialized.
My future was uncertain, and I didn't know if I should follow up on the offer I received at University of South Florida. I was neglecting it because I would have rather attend USC instead. Now I get to attend a university that's very strong in robotics. I'm very excited about going, and I can't wait to meet some of the professors of whom I've read about in books and papers.
Although, I'm not too enthusiastic about living in Los Angeles. I suppose I can get used to the heat and smog. It's a fair trade.
:: everist 6:15 PM [+] ::
:: Tuesday, May 06, 2003 ::
Ideas for My Graduate Thesis Work
These are some ideas I've been throwing around for a while that I might choose to do my graduate research on. Many of them are just superficial ideas that may or may not have much depth to them. All of them are related to robotics and AI.
Modular Reconfigurable Robotics: This involves collections of robotic units that are all interconnected. With a collective effort, they can metamorph themselves into complicated and useful forms such as snakes, walking legs, chairs, velociraptors, etc.. The challenging part of this new field is how does the robot make decisions on how to move its parts around to achieve these new forms? Additionally, if some of these individual parts would break or go bad-- as they are guaranteed to do if you have units up into the thousands-- how do you deal with these dead limbs? Do you treat them like tumors and eject them from your body or do you just work around them?
Robopsychology (or Robotic Psychological Disorders): First pioneered by a severe female scientist in one Isaac Asimov's famous science fiction novels, this field doesn't actually exist. Rather it's a hodgepodge of research of AI researchers and psychologists examining what exactly it is to be a thinking being. I would like to take it a step further by examining the AI systems in use today in academia and the real world and catalog all the various psychological disorders and insansities these systems have suffered and research their impact on people and other AI systems around them. I think it would be fascinating to contrast this with human and animal psychology theories and see what emerges. Currently most AI systems are delusional since their entire world view is represented internally and are incapable of comparing it with reality.
Robotic Sexuality: This idea popped into my head after I saw the Tickle Salon. It's basically a robotic brush that massages your body autonomously. The map-making and kinematics of the project are very interesting, but the sexual undertones of the demonstration are unmistakable. It occurred to me that there's a whole field of knowledge to be had that's been relatively untouched for decades in AI and robotics research. Sexuality is a very broad and complicated term in and of itself with things like gender roles, gender identity, sexual orientation, fetishes, biology, and a whole mess of other things that I can't even begin to list. It would be fascinating to see how these things might be applied to disembodied or robotic entities. Of course, there's a reason a lot of this research hasn't been done before. Funding for research of sexual-related topics is scarce to come by and publishing papers on the topic could give you a reputation that you would prefer to live without.
Robotic Asteroid Mining: Something I've been peripherally studying for a couple years now. The basic idea is that there are lots of valuable resources to be had in space such as metals or hydrocarbons. What makes them so valuable is that they are already in space instead of on the surface of the Earth. The cost of launching materials into space on rockets is enormous and the effort involved in procuring resources off-world may be worth it. The problem is in the risk, time, and cost of the operation. Believe it or not, it's far more cost effective to travel to an asteroid and bring resources back to the Earth over a period of 1-2 years, then it is to travel to the moon and bring resources back in 2 months. But it's still too high of cost in general. One of the possible options of making a feasible mission would be to send a robotic prospector and excavator. The cost-savings in not sending a human are several orders of magnitude, but they increase the overall risk of the mission. The key is how do you make an effective and reliable semi-autonomous robot (or robots) that will carry out the mission in an environment you have no control over and little initial knowledge. It's not as easy as it sounds. There's a whole slough issues such as dust in the gears from the asteroid, not falling off the asteroid in a microgravity environment, weathering the harsh change in temperatures as the sun comes over the horizon every 30 minutes (or 10 hours, depending on the body), and other numerous hazards that I won't go into.
:: everist 2:31 PM [+] ::
Wow! It's been a long time since I wrote on here. Approximately 3 months. Well, you're probably all wondering what the heck I've been doing lately.
For one, I'm finishing up another year of school. The interesting thing about this year is that it hasn't been an incredibly important set of classes. About half of the classes I took this year wont count towards any degree I'm pursuing. A partial reason for this is that I changed my mind halfway through the year about what I was going to do. I had fun though! I don't regret taking any of the classes I did, and I learned so much.
Next year I'm going to graduate school to begin my pursuit of a Ph.D. in computer science. Specifically, I am going to study robotics. I've been accepted to the University of South Florida and I am waiting to hear from the University of Southern California. Though I was rejected for the Ph.D. program at USC, it is very likely that I will be accepted as a Master's student instead, though I will be able to apply all my earned credits to a subsequent Ph.D. program. If accepted, I will definitely be going to USC.
The current classes I'm taking this term are Artificial Intelligence, Computer Data Structures, and Modern Physics. I'm really psyched about taking physics since it's something I've wanted to do for about 4 years, but never had the time up until now.
Work is progressing on my robot project, and I'm very excited that things are starting to be put together. For a long time, things seemed like they were going nowhere. However, now that we have parts to work with, I've been able to delegate duties to the team members and get things moving. I've found that one of the better strategies is to serve as a supporting role in other people's tasks because it keeps me involved in the design process and keeps the person I'm working with motivated. The website hasn't been updated in a month but you can visit it here.
:: everist 10:33 PM [+] ::