Doctor of Philosophy and the AI Wars

My return to Texas A&M University brought with it new opportunities and challenges. Given we were going to be here for three years, it made sense to buy a house. We were very fortunate in getting a house in a great school district in a cul de sac five minutes from campus. The kids could walk to school. Eileen and I knew the area and I knew the university very well. On the slightly negative front, I had taught for three years at West Point and formally trained to teach well. I could detect poor teaching and professor apathy much faster. I was going to have to adjust back into the role of student.

I would pick up new colleagues and friends like Jeff Humphrey, Buck Surdu, Dan Ragsdale, Rocky Gay, Michael Miller, Jim Jansen and John Hill.

The Dissertation

My faculty advisor would be Dr Udo Pooch and I would continue to focus on cybersecurity and artificial intelligence. Specifically, I would examine the use of intelligent agents to automatically defend and respond to computer attacks. At the time, it was a very controversial topic. While the exponential growth of computer attacks and associated sophistication was well recognized, our defenses and responsiveness in responding remained primitive. Artificial intelligence could help.

I took a month to write the 25,000 line of code necessary to build the simulation to test my thesis and provide the results portion of the dissertation. Writing 25,000 lines of code in a month has the following schedule:

  • Get up at 0700 and make coffee
  • Start loading all of the variables and implications into your brain
  • By 0900, start coding and continue coding until 11 PM or midnight. Food magically appeared and preferably food that can be eaten with one hand.
  • Extra points for talking to yourself. Double extra points if you mumble because you are sleep deprived. Acting as if you do not see or hear anything around you works. Taking over the kitchen table has a multiplicative effect.

After a month, Eileen had had enough and fortunately I was finished. The rest of the dissertation process went smoothly and I wrote the document in Microsoft Word and then transferred it into LaTEX which was the only format I could submit the document. It took five weeks to get the “approved” dissertation through the editors in the university library but this was a well known huddle at Texas A&M.

My class work beyond Masters degree was 12 graduate courses and then of course the dissertation. A couple of vignettes from those 12 classes:

  • Dr Pooch offered a graduate level advanced security course with an interesting premise. The course would have only one graded project: offer services on three different operating systems and defend against attacks from the other teams. Sounds simple except you could not earn a passing grade unless you had at least one successful attack. My team passed the course which was not interesting. What was interesting is that every other team was completely locked out of all of their machines by the end of the first lesson. Some teams created really complex passwords that no on could remember. Some teams decided to patch multiple components at the same time because it was boring to do them one at a time. With great creativity, all of the systems were so completely screwed that they had to be rebuilt from scratch. It was an enjoyable class.
  • Perhaps the most interesting was John Leggett’s Information Retrieval class. One of the course assignments involved implementing canonical Huffman encoding with a large data set. We were given five weeks to complete the assignment. My team jumped on the project immediately and built a solution. We then watched the program work but work so slowly as to be impractical. Three weeks of constant coding and improvement and the program ran in less than 5 seconds. On the week the assignment was due, one of the teams asked Dr Leggett if he cared what programing language they used to complete the assignment. He responded he didn’t care and that the team was going to fail the assignment because they had waited too late to start programing. We concurred and Dr Leggett was right. The other team submitted about a week late with a 70% reduction in grade just due to tardiness.
  • I dropped only one class in my academic career and it was due to transparent and rampant cheating that the full professor was just not going to address. Of all things, the professor was related to me.