- Our Tools
- For Developers
LTfLL services make substantial use of language technologies to provide extra value to our tools. Language technologies try to "understand" human input (e.g. digital text) on a higher level. Our tools work with 'concepts' rather than words. This means that they can deal with a certain amount of inaccuracy or abstraction. Here are some examples:
Semantic concepts are more than words
Similar words can be grouped together by meaning, so-called semantic fields. For example, 'red' and 'blue' are colures; 'ducks' and 'geese' are birds. Sometimes words have a very nearly similar meaning, they are called synonyms, e.g. 'monitor' and 'screen' are synonymous, but only when referred to in a computer context.
Scholarly domains lay down their special terminology and meaning in controlled vocabularies and ontologies, which are basically collections of agreed concepts between scientists working in the same field. 'monitor' and 'screen' in a computer context covers a different concept than the same words in the medical domain (where they are verbs).
Concepts are arranged in hierarchies (ontologies). For example, 'HTML' and 'XML' fall into the same category of 'Markup Languages'.
Human language uses reference words
In human language we do not always spell out every word in full. We use internal referencing so we don't have to repeat everything in full. This is often done using pronouns: "the book is mine" instead of "the book is my book". For us this is easy to understand, but not so for a machine. Special rules are required to tell the computer that 'mine' in this sentence means 'my book' and in another sentence it means 'my dog' or something else.
Typing is often not exact
Typed text often contains misspellings, typing errors, or grammatical mistakes. Especially in electronic communication like chat or sms talk a special language has developed that abbreviates full typing, e.g. 'btw' instead of 'by the way'. Even in ordinary language, speakers reduce full sentences to fragments, e.g. in answers: "yes, I have (done this)".
Such grammatically incomplete utterances require special techniques to make them intelligible to machines for further processing.
Below you can download the related materials: