What I did today:
1. (1 hour) Met with Percy to discuss the setup some more
2. (1 hour) Planned out how to code the factorial HMM inference algorithm
3. (1 hour) NLP reading group
4. (2 hours) Wrote code to check hypotheses about whether a given approximate likelihood would work well (answer: seems likely to work well).
5. (1.5 hours) Start to implement the inference algorithm.
I’m not yet done with this last part but hope to finish it up tomorrow.