I spent the last two weeks at SPARC, where together with several other instructors I taught high-performance high schoolers statistics, computer science, rationality, and self-improvement. Here is what I taught:
1. Calibration Game: a game where people practice placing bets on various trivia questions. The goal is to gain a visceral understanding of what it means to be, e.g., 80% sure of something. This was the least popular of the classes I taught, so I probably won’t run it again next year, at least not as a full unit.
2. Human Vision: went over some of the basics of cognitive vision, such as contrast, stereo, and edge detection.
3. Programming Lab 1: this consisted of a python tutorial together with two more substantial tasks: build a spam filter and build a word segmenter (i.e. given a non-delimited list of phonemes, separate it into words). It seems like this and the other programming lab were too challenging for students who had never programmed before, so I will work on making the tutorial better next year.
4. Common Probability Distributions: went over the central limit theorem and properties of normal distributions. In classes where I had some extra time, I also briefly went over the Poisson distribution.
5. Inference in Graphical Models: introduced the notion of a Markov random field (Bayes nets had already been introduced earlier) and presented the elimination algorithm.
6. Programming Lab 2: students were supposed to predict the 10th term in an OEIS sequence given the first 9 terms.
7. Natural Language Processing: this was my highest-rated class, so I will definitely run it again next year. After giving an overview of NLP and linguistics, I went over dependency parsing and CFG parsing, and explained the chart parsing algorithm.
In addition to classes there were many other activities. I took groups of ~5 kids bouldering on each of the two rest days, and taught many students how to play Yomi, which is my new favorite game.