C. Drake P.

Poster Report.

cdrakep — Sun, 13 Dec 2009 21:50:21 +0000

Tomorrow is poster day for CSUMS. Here is the PDF for my and Chris’s poster.

Poster for Numerical Characterization of Entropy Data

The main focus of our project was verifying that the Entropy Key was returning data that was rigorously random, or entropic, as the name would suggest. The algorithms we implemented and tested on various data sets, derived from both the entropy key and other sources of supposedly random data, suggest that both the programs themselves are working correctly, and that the Entropy Key returns data at least as random as we can test for.

Most of our time was spent on coding and verifying that our code was returning useful information. We both implemented methods of further visualizing the data. Further time was spent making sure the data was in a format which our programs could parse. So in the end we have performed lots of tests on large sets of data, the longest taking approximately eight weeks of computational time (1 week real time) on the computational math server.

So basically our conclusion, in statistical terms, yes the Entropy Key is random. It has not failed the null hypothesis test. That is about all that can be said given the nature of randomness.

i win

cdrakep — Wed, 18 Nov 2009 21:08:19 +0000

I win

Weekly writeup for 10/27

cdrakep — Sat, 24 Oct 2009 22:27:13 +0000

Well, waited on some results from all my terminal monkeying. Wrote a bunch of programs and scripts to manipulate data. Made a lot of graphs. A big thanks to the 20 cores over 5 computers that resulted in roughly 3 months of computation over the last two or three weeks.

Most of the work over the last couple weeks relied heavily on Chris’s OpenMP Auto-Correlation code. Into this code was fed random data obtained wholly from Random.org. In addition to this known periodic data was also fed into it, all of which resulted in some pretty graphs, some computer lock ups, and some ugly graphs. In reverse of the stated order.

All the stuff I promised I would link to.. I lied. I’ll do it tomorrow or later tonight, maybe Monday.

For Your Consideration, Write Up the Third.

Weekly Update 10/6

cdrakep — Wed, 07 Oct 2009 00:33:08 +0000

This week I’ve mostly been going through the C code for the entropy keys and the code supplied by the NIST, as well as updating my M code to C to better understand their code and come up with methods of combining and automating several tasks to shoe certain properties of random numbers. As a result, I don’t have much in the way of code or results to show.

Weekly Update SomethingSomething

cdrakep — Wed, 30 Sep 2009 02:31:11 +0000

This week I did some testing of random numbers to try and establish some properties to look for when we get Entropy Keys.

Write Up for this post

Links to the code.

Chris’s Auto-Correlation Code
Random Xor Mix Code

Random Sets Xor’d, Non-Random Mix Code
All 1 set a Xor’d Random Mix Code

Alternating 1s and 0s for a, random b Mixed Randomly Code
Alternating 1s and 0s for a, random b, Mixed Non-Randomly Code

Quick Graphs Code

Change the value to j for larger random sets. Multiply that by 3 for the lower loop. In randomxorcorr change the sample size to whatever is desired as long as it is less than the loop size.

Weekly Update 09/20

cdrakep — Mon, 21 Sep 2009 04:10:56 +0000

I started out the week playing with the predator prey function. I explored how to translate my existing programs into ‘auto-test’ programs. I finished the week training neural networks. Read more about it in the write up, if the link works. It doesn’t seem to for me.

Weekly Write Up 9/20

When I go to media, and insert the PDF it comes out like this ::

Weekly Write Up 9/20

Oh thats better. I didn’t realize I had to click “file url” or whatever it was that I just clicked.