While I understand it digresses from learning Korean, I certainly have developed an interest in dictionaries (and the structures a computational linguist would use).
The Merriam-Websters Dictionary API is available free for non-commercial use, with certain limitations.
Their Developers center describes how to access their API.
XML: http://www.dictionaryapi.com/api/v1/references/learners/xml/apple?key=%5BYOUR KEY GOES HERE]
It provides a list of the XML tags learners-tag-description and a sample XML response.
Further information can be found in the OED Guide. While I knew about headwords, entry words, parts of speech, definition, synonyms, and cross references, there were other parts of dictionary entries I did not know before reading this article.
Learning Korean has turned into a sub-project of creating my own mini dictionary of Korean words I am learning. I find I need things like HEADWORDS and homonym numbers, alphabetizing letter (in both English and Korean), and cross references. Not to mention parts of speech, categories, and sources.
For my purposes, all I would want is the English definition for my Korean words I am learning. However, looking into the data returned is an interesting peek under the hood of how dictionaries are put together.