I like this quote at the beginning of the chapter on regression models in my stats book:
So startling would his results appear to the uninitiated that until they learned the process by which he had arrived at them they might well consider him as a necromancer.
Dr. Watson speaking about Sherlock Holmes — A Study in Scarlet
We had freezing rain in Fairfax overnight and this morning and the ice expanded the seam of two panels on the back drivers side of my car which cracked the bottom panel!
https://docs.google.com/drawings/d/1VRNvbyjAiI8JXjfe7sgV21VzoN2LI3OiBFmPuZRzE-g/edit?usp=sharing — for pic red box is where it happened. Ugh…
When I’m working on a problem, I never think about beauty. I think only how to solve the problem. But when I have finished, if the solution is not beautiful, I know it is wrong.
Buckminster Fuller
Came across this quote as I was looking into whether there exists N-dimensional hierarchies.
Found this paper http://www1.ceses.cuni.cz/benacek/3D-Apr05-full_version.pdf not the most useful, but interesting (haven’t read the whole thing).
This year started off quite well:
1. My friend Indranil passed the CS quals
2. I became a Graduate Research Assistant (GRA) and got an early 2013 15″ macbook pro!
3. I might get to go to Brazil for Robocup (as long as we qualify)
4. I might get to go to Prague to present a paper at ECAI as long as I can take the quals the week before.
5. I have two paper worthy ideas that I am researching! (well one is a survey paper on heterogenous multrobotic systems and their applications and the other is Dynamic auction protocols for task allocation in multirobotics (this could lead to bigger things))
6. Today I found out that one of my ML ideas for creating new recipes was just implemented at IBM! I couldn’t pursue it because I didn’t have the data, but am glad they did.
7. Learned how to solder (really easy) and make the connector cables for robots (hard)
and probably a lot more I can’t remember.
Should learn stochastic control theory because I think that some of their results may apply to multiagent learning.
Need a game that people play that simulates human trafficking without it actually being human trafficking (since that would just be wrong). It could be a MMO strategy game that requires the players to distribute mail in a war zone. So, you can choose to be the good guy, the mail distributor (which corresponds to the human trafficker and the mail is like the humans). Or the bad guy and try and stop the mail men.
We can then use machine learning to identify trends in the strategies, social networks and roles that form as the game progresses. Using this data along with real world human trafficking data we can create a multiagent simulation that can predict the effects of various decisions.
They are pretty cool. This post was driven (pun intended) by the need for determining whether a 2003 BMW 325i can be safely driven with regular gas rather than higher octane gas.
The compression ratio of the bmw is 9:1. This ratio tells us the volume of the combustion chamber at its largest capacity to its smallest capacity. Essentially, the larger the ratio the more heat we use to actually do work the better. However, this requires higher octane gas in order for knocking (the gas detonating) not to occur. Basically higher octane gas has a higher activation level (meaning it can stand a much higher temp/pressure before it combusts). So, when a lower octane gas is put into a higher compression ratio engine the fuel will self detonate (knock) before the piston compresses the whole way possibly causing damage to the piston. When a knock sensor detects knocking it adjusts when it ignites the fuel so that knocking won’t occur. However, fuel will be wasted.
Knock sensors are required on vehicles after 1996. So, it seems like it is ok to put lower octane gas into the bmw. Maybe keep an eye on the mpg and compare it to what the manual says it should be getting.
My mom taught all this to me when I was homeschooled! Wish I had a better memory.
http://upload.wikimedia.org/wikipedia/commons/d/dc/4StrokeEngine_Ortho_3D_Small.gif
https://app.box.com/s/fi1c3wccxjrwtan9hq69 is a very interesting article I’ll write a bit about it later.
http://www.math.hawaii.edu/~dale/godel/godel.html descriptions of a couple paradoxes and theorems important to computer science and mathematics.
I posted every month this year! Of course this is a pretty small one….
Lua is a pretty neat scripting language. It can be embedded in C programs and is used in some set top boxes, games and the UPenn robocup code http://j.mp/1c9DX09.
Note to self, print(package.path) will show where Lua searches for Lua code.