We’ve already covered Vision and Speech API – now’s the turn for Language API.
This part consists the most features that we could try. We can spell check with Bing, use Language Understanding Intelligence Service (LUIS) for understanding the commands from the users, or analyse the text with Linguistic Analytics API. What is also available is Text Analysis (i.e. sentiment) or just simple a translator API. The last available service is Web Language Model based on data from the web.
Skipping the translator, as this is probably not very interesting service, let’s check what is…
LUIS or Language Understanding Intelligence Service offers the ability to add text understanding to our application without coding a bunch of if-statements. The actual page for the LUIS service is luis.ai and there we need to create an account in order to start using the service. There we can provide examples of the phrases that LUIS should learn from and extract relevant information of the things we want to talk to it. If we just want to play with it, there’s a nice example on MS Cognitive Services website.
Text Analysis API
Another interesting services that located under the Language part of Microsoft Cognitive Services. It allows to extract useful information from the text.
It can detect the language, extract key phrases from the text and also provide information about sentiment. Currently works with language like English and Spanish. For Polish, it can only detect that the text is written in it.
The last described service is..
Web Language Model
Services that we can use there are ranging from detecting the spaces in a long string. So we can get “some long text without spaces and anything” from “somelongtextwithoutspacesandanything”.
It allows also to calculate probabilities of words/phrases being put in a sentence together as well as to get the percentage of two words/phrases are combined with each other. Finally we can use it to predict what is the best match for group of words that is commonly present.
If you are dealing with text – Microsoft Cognitive Services Language API might be handy to add additional layer of intelligence to your app. Have you used it already?