Disease testing isn’t perfect. Illustrating how the prevalence of a disease, including COVID-19, can be estimated using an imperfect test and how probability depends on the prevalence, this blog highlights how a single test result should only be one piece of information used in determining an individual’s infection status.
I’m on sabbatical from Cardinal Peak right now, teaching at Carnegie Mellon’s Africa campus in Kigali, Rwanda. Aside from the fun I’ve had being in another culture and in an academic environment, it’s been great to see how a world class university like CMU educates graduate students in ECE and IT overseas. CMU’s Rwanda program… View Article
I’ve been working on a fun problem lately that involves estimating a scalar parameter from a set of repeated observations. It turns out that in certain circumstances, the presence of noise in the system can actually make the estimate more accurate, which is a little counterintuitive and also kind of cool. In my case, I… View Article
I’ve got a new blog post up at EE Times, talking about how to build video devices for the Internet of Things: Video cameras are uniquely compelling sensors because vision is our dominant sense. Video is invaluable for applications such as license plate recognition, robot navigation, and quality monitoring. Unfortunately, video is also one of… View Article
Recently we tweeted an interesting article on big data, from the Financial Times. The author’s key point is that sampling bias and sampling error are possible even with large data sets. As illustration, the author discusses a classic case where the Literary Digest incorrectly predicted that Alf Landon would beat FDR in the 1936 election…. View Article
I came across the article “Why the “Next Silicon Valley” is Always Silicon Valley” during my lunchtime reading today and found it really interesting. I would summarize it — obviously not justly! — as follows: Excellence is a snowball rolling down hill. Excellence attracts more excellence in a virtuous circle. Getting the snowball started is… View Article
We’re a little late in posting this, but I wrote a blog entry for EDN last week that discusses how to choose the correct video sampling format. An excerpt: To process signals digitally, they must first be sampled and quantized. Sampling refers to measuring the light intensity at discrete space-time points, while quantization is the… View Article
In Part One of this blog post, I discussed how to state an experiment in the form of probability spaces. Determining the sample space and the event space is necessary to be able to talk intelligently about probability measures, which is the topic of this post. Approach 1: Counting We’ve figured out the sample space… View Article
Cardinal Peak’s big data practice is expanding as we continue adding data scientists to our staff. In a recent discussion regarding a data set we’re analyzing, a probability problem conceptually equivalent to the following arose: In a room filled with N people, what is the probability that none of them have the same birthday? In… View Article
Last night I went to the Denver IEEE meeting of the Signal Processing Society. I was particularly interested in this talk because it was given by Gary Sullivan, the co-chair of the recent international standardization effort to create the High Efficiency Video Coding Standard (HEVC).