ILO forms are first recognised using regular expressions, then modelled using methods from lexical semantics.
In order to aid the correct determination of the meaning of an ILO, a natural language parser is used to identify constituent phrases and clauses.
Stripped of the introductory clause, like After completing this module a student will be able to..., ILOs take the form of imperatives (orders): Discuss this or Explain that, for example. The verb might be modified by some adverb or adverbial phrase.
It is interesting to note that another domain in which there are similar constructions for conveying information is - kitchen recipes!
State Boyle's Law and use it in simple calculations.
Chop the filet mignon and add it to the skillet.
It is also interesting that there seems to be have been more research into the natural language processing of recipes than of ILOs. Let the cooks show the way.
For example, there is an article that describes the use of semantic role labelling to extract meaning from recipes [1].
There is also a computer cooking contest:
The goal of the CCC is attract new people (e.g., students) to work with AI technologies such as case-based reasoning, semantic technologies, search, and information extraction.One of the distant aims of the formalisation of ILOs is the comparison of their expression in different languages. Very Bolognese. It is a little known fact that the Hungarians do not use the imperative mood when writing recipes. They use the first person plural: We chop the filet mignon and add it to the skillet.
I wonder if the same is true of the Hungarian ILO?
We state Boyle's Law and use it in simple calculations.
We are amused.
1. Agarwal, R. and Miller, K., Information Extraction from Recipes, http://nlp.stanford.edu/courses/cs224n/2011/reports/rahul1-kjmiller.pdf
