Below are a variety of pieces and papers that I wrote. Many of them are from college, but I hope they hold up. I have also written for:
- How to Burn Down the Kitchen Without Really Trying
- Spawning Tool Blog
- The Unofficial Stanford Blog
- Yelp Reviews
- Goodreads Reviews
- Zanbato Blog
PDF – Spies Like You
I wrote and published a one-shot Dungeons & Dragons adventure. The pitch is below with more details about the process in this blog post.
Mission Impossible meets D&D. It’s your first mission as royal spies, and you will have to be crafty, sneaky, and clever to get through it. It is up to you to use your imagination to come up with the craziest ideas that might just work and improvise your way through dangers you definitely weren’t trained for. This 4 hour adventure is designed for both first-time players and first-time Dungeon Masters. It includes the adventure itself, background, basic tips on starting, several maps, monster stat blocks, and 6 1st-level pre-generated characters.
PDF – Engineers for Designers
A presentation I have done for Whitespace a few times to teach product designers how to work with software engineers. You can read corresponding notes here.
A group of high school students play their ongoing game of Dungeons & Dragons. As they complete quests and approach the end of the game, they find their insight into problems in their real lives as well.
PDF – You Use This Washroom Often?
Fun essay for a creative nonfiction class in college. I wandered around washrooms on campus.
PDF – Ali, K., Leung, K., Konik, T., Choi, D., Shapiro, D. (2010). Knowledge-Directed Theory Revision. Inductive Logic Programming 2009, Lecture Notes in Computer Science, 5989, 1–8.
A published paper on a theory revision module for the ICARUS cognitive architecture, based on work from a summer internship during college. It is artificial intelligence using logical reasoning to develop better football plays.
PDF – Identifying Structure in Gradable Adjectives Through Corpus Analysis
Research paper for a college class CS224N, “Natural Language Processing” that I did with Dan Lassiter. Here’s the summary of it.
Many of human concepts can be expressed in varying degrees, and in English, we can express these using gradable adjectives. Gradable adjectives accept degree modifiers, which determine the strength of the concept being expressed. For example, “tall” is a gradable adjective because one can be “very tall” or “kind of tall.” On the other hand, “dead” is a non-gradable adjective because one cannot be “very dead” or “slightly dead.” Different gradable adjectives, however, accept different degree modifiers, and Kennedy and McNally offer a theory that types of gradable adjectives are determined by their scale structures.
This paper was a corpus study to see how frequently adjectives were paired with different modifiers. For CS224U, we did a subsequent experimental study on Mechanical Turk to get human judgments as well. I think it turned out well, but I can’t seem to find that report. Unfortunately, I graduated that term and never followed up to publish any results on it.
Related code is in this GitHub project.
PDF – Identifying Actors in Political Activism over Twitter
Research paper for a college class CS378, “Phenomenological Foundation of Cognition, Language & Computation.” I was interested in the use of Twitter for political activism (such as the Egyptian protests) and thought that there were interesting questions about identity and commitment in that context. I take some ideas from network models to understand the roles that individuals have in terms of concrete actions, and I connect that to some empirical work on types of actors in activism on Twitter to understand where identity comes from.
PDF – Evolution of Internet Information Consumption Through Bookmarking
Project writeup for a college class CS224W, “Social and Information Network Analysis.” The question I had going into this was whether we could quantify how information overload might be reflected in changes in internet usage. Particularly, Neal Gabler here complains that we can’t grapple with big ideas and are stuck in the constant flow of unimportant data. I tackled this hypothesis by looking at Delicious data and seeing whether the distribution over bookmarks has changed from year to year. If he’s right, we should see more bookmarks happen sooner and less of a long tail. The basic result is that over 3 years, things look pretty much exactly the same, and I also try to come up with a model to explain the data.
Code is in this GitHub project.
PDF – A Connectionist Model for Visual Search via Evidence Accumulation
Project writeup for the 2nd version of the model in the paper below for a college class Psych 209B. Again, there are some pretty serious problems, but it does kind of work, too.
PDF – A Connectionist Model for Visual Search via Excitation
Project writeup for a college class Psych 209A, Jay McClelland’s PDP class, where I tried a very simple model for visual search. The model largely doesn’t work, but it was great for me to find out all of the things I don’t know about visual search and need to research more in-depth.
PDF – Robinson, D., Leung, K., Falco, X. Spoken Language Identification With Hierarchical Temporal Memories.
Project writeup for a college class CS229, “Machine Learning”, where we tried using Numenta’s HTM technology to improve classification of different spoken languages.
PDF – Does Word Order Affect Processing?
Minor research paper for a college psycholinguistics class where I conduct an experiment to analyze the importance of Behaghel’s Law on processing. (Hint: it’s not very important)
PDF – A Review of Predictions from the Dynamic Routing Circuit
Survey paper for a college class on models of the neocortex. In it, I review some stuff from an earlier Olshausen model. It’s a literature review of neuroscience papers, so pretty far out of my comfort zone, but it was fun.
PDF – Direct and Indirect Consequences of Chosen and Alternative Actions
Final paper for a college class CS222, “Intelligent Interaction and Rational Agency” where I extend Propositional Dynamic Logic to include some ideas of consequence. The paper leans heavily on material from class that I don’t really explain, and I’m not sure how original the ideas are.