tl;dr Debugged and tested; trained NLU; gave small demo
More debugging and testing (and consequent fixing) for the demo; Alankar, Yoichi, and I also worked to run the full system with all of its improved components.
And I retrained the NLU based on new data, which involved a side trail of inexplicable compilation errors but was fun once I was able to feed sentences in (and test out the new model).
Finally, our team gave a small demo (I spoke with Sara!) for a couple of inpromptu (but important) visitors, which was a good practice run for next week's big demo--afterwards, we all exchanged feedback/suggestions on our respective parts (Yoichi and Alankar, for instance, both presented slides) as well as on our question-answering.
On Thursday, our lab started a week-long game of Assassin (which is basically stealth-tag) and I hustled home Baltimore-style (see rightmost tab here) for fear of getting "out."
[Update: I tagged five people this week, which involved one phone call; lurking in two stairwells; and charging at Yoo Jin full-speed.]
Also: Yoo Jin, Shruti, and I grabbed brunch over the weekend and went shopping for movie marathon (see next week) snacks.