Improvements
Discovered Errors
- In the functions base_cat_subj, base_cat_obj, and base_cat_indobj, KE checks to make sure that the noun she is categorizing is basic level, but then returns all possibilities in result.
- In the functions
some_cat_obj and some_cat_indobj, KE writes the same statement twice. I think that the
second one should read objects1 and "ARE". Also, in some_cat_obj, she uses agent-, instead of object-.
Algorithm Comments
- Although she uses the terms "cause" and "effect", KE really means
results and enabling conditions, respectively.
- Are verb types classified through general pattern-based rules, or
through an explicit classification?
- If through rules, transitive and reflexive verbs need to be
identified better in her background knowledge as it stands in her dissertation.
- Perhaps a better approach to the verb algorithm is to use an
algorithm more like the noun algorithm. For example, verbs have
synonyms, antonyms, and maybe a physical/abstract nature.