Articut NLP AI
  Free Trial: 2,000  [Reset Hourly]
Key (full word)
Value (alias or abbreviation)
Should you have any questions on how to use Articut or the correctness of the result, please come to our Facebook page for further discussion: Articut Facebook Fans' Page. All languages are welcome.
  Free Trial:2,000  (Reset on the hour)

Droidtown NLP/NLU Solutions

Natural Language is the most sophisticated behavior a human can perform. With massive amount of calculation on contexts, syntax, semantics, morphology, phonology, and phonetics, it is an extremely difficult task to enable computers to process natural languages.

Luckily, with the advances of modern linguistic researches, Droidtown Linguistic Tech. adopts Professor Noam Chomsky's theories and implements them into NLP/NLU software solutions that can process various languages — just like what Universal Grammar promised.

Unlike other data-driven machine learning solutions, Articut does not use "space" as word boundary. On the other hand, Articut utilizes X-bar theory to preserve meanings (semantics) of words on a syntax tree node. In this case,

Don't give up

would be processed into

Do / not / give up

instead of

Don't / give / up

which only uses space to segment the sentence and thus drops the meaning of "give up."

What's more, Articut also attaches Part-of-Speech (POS) tags to the words/phrases basing on the syntax tree nodes. Therefore, "Don't give up" is then tagged as

<MODAL>Do</MODAL> <FUNC_negation>not</FUNC_negation> <VerbP>give up</VerbP>

and provides much richer semantics for further natural language applications.

Word Segmentation

Find the minimal meaningful words with linguistic rules

Named Entity Recognition (NER)

Recognized specific named entities in the text (e.g., person names, place names, dates, time and other nouns...)

Part-of-Speech (POS)

Each word plays a syntactic role in the sentence, and it's called 'Part-Of-Speech' (e.g., noun, verb...). In addition to common POS, Articut has the ability to deal with the POS-shifting problem. For example, the 'record' in the sentence 'I got this idea to record a record here.' belong to verb and noun. Articut can identify the POS of the two 'record' as verb and noun according to the syntactic structure of the sentence and give them proper tags.

Location-Event Extraction

With the POS marks provided by Articut, we can extract 'TIME', 'LOCATION', and 'EVENT' from the text. For example, in the sentence 'We plan to go to a restaurant in downtown Boston for lunch tomorrow,' Articut can extract 'tomorrow' as the time, 'Boston' as the location and 'having lunch' as the event in the sentence.

Question Detection

Articut automatically detects interrogative expressions in the sentence. In this case, whenever a 'Yes-No' question or 'Wh' question is proposed, your NLU applications can interact with humans more naturally.