We spend a lot of time talking about supervised learning when discussing ML with students, but I find reinforcement learning just as interesting and useful.
I developed a talk on reinforcement learning for high school participants in SAIL ON, the year-round diversity and high school outreach program of the Stanford AI Lab that I initiated and led, which follows the SAILORS intensive two-week AI summer camp.
We discuss how reinforcement learning works, how to make decisions given Bayesian bounds, touchstone RL problems and recent applications, and where RL tends to succeed and fail.
The talk introduces the language identification problem within AI, teaches about decision trees, and then asks students to write decision trees in small groups to distinguish between Hmong, Balinese, Zulu, and other languages. After a debrief on why computers are might be more effective than human-written rules, it briefly ties in themes of feature extraction and gradient descent via GBMs.
As the progenitor and leader of SAIL ON, the Stanford AI Lab’s year-round effort to attract and keep underrepresented minorities in the field of Artificial Intelligence, I engage high schoolers about artificial intelligence, machine learning, and the positive social impacts of our field. SAIL ON meets once a month in the Computer Science building at Stanford. Its trifold focus allows past participants in the SAILORS two-week summer camp to continue to learn about AI, to nurture strong relationships with each other, and to lead outreach projects that bring the technical, humanistic, and diversity missions of the AI outreach program to the wider community.
As the educational component of the October meeting of SAIL ON, we discussed and applied Naive Bayes modeling. Like other machine learning methods, Naive Bayes is a generic approach to learning from specific data such that we can make predictions in the future. Whether the application is predicting cancer/whether you’ll care about an email/who will win an election, we can use the mathematics of Naive Bayes to connect a set of input variables to a target output variable. (Of course, some problems are harder than others!)
By providing this basis, I hope to increase appreciation for applications of what students are seeing in their math classes, and to facilitate students moving further on their own with applied machine learning before November’s meeting.
One of my favorite parts of studying linguistics was being presented with data and being asked to find the system within it. Language data, with linguistic theory’s insistence that everything must make sense, make the most excellent data and logic puzzles.
As part of preparing for a spatial grammar-heavy meeting of the Montgomery Blair High School Linguistics Club, I developed three American Sign Language morphology problems. These problems illustrate interesting properties of American Sign Language that spoken languages do not have (non-manual markers, spatial agreement, and a rich temporal inflection system based in manual phonology).
Try your hand at doing the problems if you’re interested in any of the following:
What it means to do theoretical linguistics (or the sort of logic skills that linguists develop)
Unique properties of spatial languages
Basic American Sign Language linguistics
Similarities between American Sign Language and other world languages
Unlike most materials on ASL linguistics, the problems don’t assume that readers are fluent in American Sign Language or in linguistic theory — I developed these problems because I couldn’t find any resources aimed at an intelligent lay non-Deaf audience. The problems deliberately walk users through the steps to answer a question, whereas most theoretical linguistics problem sets jump straight to the questions at hand and assume existing familiarity with linguistic features not observed in English.
Once the club and I meet, I’ll post the answer sheet as well.
During the discussion we focused on introducing different non-voiced communication forms and on linguistic anthropology/linguistic creativity. We postponed theoretical linguistics until another time (in which we did some experiential learning on morphology). This page consists of a set of links, prepared videos, and notes designed to support real-time interaction with students at the linguistics club at Montgomery Blair High School.
The big take-away is that American Sign Language is not “English on the hands”. ASL is independent from English both in grammar and linguistic culture.
Introduction
Caveats for posterity: I’m hearing, I don’t possess native-like fluency in ASL, and I don’t have an advanced degree in this; I do have general and ASL linguistic training, I read widely, and I’m more or less aware of what I don’t know
What are some ways deaf people communicate? [YouTube]
Compare ASL structure [.avi | .ogv | .gif] with PSE structure [.avi | .ogv | .gif] with English structure [.txt]
YouTube has a variety of performances, lectures and vlogs
What’s this blog about?
Whatever is on my mind. The content has varied over the past more-than-decade, but it's always been technical. In the early years I focused on improving the fabric of the internet for some niche tools. But the internet no longer needs that kind of improving, and search doesn't really work like that anymore either. This blog is currently mostly about documenting notes for my future self, and sharing those notes with anyone who is interested.
