How I use Artificial Intelligence to find the bad guys in White Thaw

October 15, 2013 2:52 pm Published by Leave your thoughts

Until 2001, I worked for the Naval Research Laboratory (NRL) as a research meteorologist. I retired to write fiction. White Thaw: The Helheim Conspiracy is the third thriller in a trilogy involving my protagonists Drs. Victor Mark Silverstein and Linda Kipling. They are Navy scientists who work at the same facility as I did. Artificial Intelligence (AI) plays a critical role to the plot. This blog will discuss how my knowledge of AI came to be.

As a research meteorologist, I worked on a variety of projects during my career with the Navy. In the years before I retired, I applied AI to Navy meteorological problems.

Here’s my definition of AI: the application of numerical (usually computer) techniques to simulate the thought or vision process of a human being. I did research in a couple of areas that fall within this definition. First, we developed several expert systems. An expert system is a program that encapsulates the thought process of a human in deducing an answer to a problem. You might imagine something akin to a decision tree, but more complicated.

In another application of AI, we applied pattern recognition or computer vision to meteorological image analysis. For example, we tackled tropical cyclone intensity and cloud typing, two areas where meteorologists can develop recognition skills by studying satellite imagery. We wanted to automate that process using a computer. One can imagine other fields in which computer vision might prove useful. In medicine, for example, there are programs to automate the analysis of images created by breast mammography, searching for suspicious areas requiring further analysis.

In a third area, we applied what’s called machine learning. More generally, we applied KDD, Knowledge Discovery from Databases, to a variety of problems. (The lingo here can be confusing and is often used interchangeably. KDD refers to the overall process of discovering useful information from data. Data mining refers to the algorithms applied to the data, and machine learning is one of those algorithms.) In one of our more interesting (and potentially rewarding) studies, we applied KDD to both numerical (from weather forecast models) and satellite data sets to discern weather parameters (for example, cloud base height) where no ground-based meteorological observations were available.

This brings me to the use of AI in White Thaw. In Chapter 2, I introduce the reader to a fictional government agency: the Federal Center for Data Analysis (FCDE). In truth, following 9/11, the government realized that there were clues that might have thwarted the airliner attacks: e.g., Colleen Rowley, an FBI agent from Minneapolis, who testified before congress regarding information she had forwarded through channels regarding Zacarias Moussaoui, a suspected terrorist. As it turned out, those reports could have led to advance knowledge of the plane hijackings. This was an example of the failure of human intelligence. What I envisioned with the FCDE was an agency dedicated exclusively to the automated monitoring of digital forms of data: for example, satellite pictures of enemy territory, seismic signals (that might point to an illegal nuclear test), newspaper articles, anything available in digital form.

How is this accomplished? Analogous to my discussions above, KDD is used to automate data analysis. Such analysis cannot be done manually because it would be too slow. The idea is to uncover a suspicious pattern, a warning sign perhaps that needs to be checked out or acted upon. In my novel, it is this technology at the FCDE that uncovers a suspicious underwater explosion. This discovery leads to a nail-biting caper, resulting in murder, international intrigue, and a potential environmental catastrophe for our planet.

Categorised in:

This post was written by paulmarktag01