consolidated paper todos

[judithfan]

• [ ] get bayes factors for critical model comparisons across all splits (convert RmD example to python, generalize across splits)
  • also make figure for this (Fig 5A)
• [ ] fine tune model results figure ("Figure 5")
• [x] edit model results section, leaving placeholders for final numbers
• [ ] edit model results section, plugging in final numbers
• [ ] edit discussion (expand on related work, implications, limitations)
• [ ] decide on consistent terminology and ontology for things in the paper ( when using "social reasoning" vs. "context/cost sensitivity")
• [ ] make sure that when new terms are introduced that they are described clearly enough to be understood in the main text
• [ ] when there are numbers in parentheses, add enough supporting text around so it is clear what the numbers correspond to
• [ ] which results should we make sure to have in supplemental?
  • adaptor training time courses?
  • confusion matrices from recognition experiment
• [ ] fix missing citations
• [ ] cleanup data generation: i.e., in bdaInput, don't need duplicates of data for all perception settings

[mhw32]

Note to self: add performance on test set for embeddings...