Natural Language Processing
Introduction
Natural Language Processing (NLP) Engine analyzes natural language and converts it into a semantic structure that machines can understand.
It understands the meaning of a sentence and precisely grasps the intention of the speaker.
NLP Engine takes input directly from the user of an application, or obtained through ASR Engine, and analyzes the text in three steps: morphemic analysis, syntactic analysis, and semantic understanding.
- Morphemic Analysis: A morpheme is the smallest unit of a word with a meaning, and is the primary reference unit for analyzing natural language. The NLP engine first analyzes the input at the morphemic level.
- Syntactic Analysis: The NLP engine analyzes the syntactic structure of a sentence and structures the sentence based on parts of speech, such as nouns (subject, object) and verbs.
- Semantic Understanding: The purpose of this step is to draw the meanings needed for a machine to understand the human language. In this step, meaningful expressions are categorized to generate structured data that will be easy for the machine to understand.
NLP Engine provided by ThinQ.AI has the following features:
| Feature | Description |
|---|---|
| Hybrid method | A hybrid method of natural language intention analysis that combines machine learning and rule-based methods. Machine learning methods provide probabilistic high-intent analysis results, and rule-based methods provide a way to fine-tune the service. |
| Natural language intention analysis | Intention analysis technology consists of intention classification and object name recognition. It classifies intent codes from speech data, while object name recognition extracts important keywords used as attributes of intentions.  |
Engine Structure
All features of NLP Engine are run on a server. NLP Engine takes text and JSON data as input, and outputs the results of its natural language intention analysis.
Examples of Use
NLP Engine is used in a variety of areas that require voice-based services. Natural language recognition can also be applied to devices that require interaction, such as robots.
- Voice agent service while driving
Drivers can perform specific actions using voice commands.
- Natural language recognition in home appliances
Equipped with an NLP engine, home appliances can perform requested actions by voice. For example, a user can request information about food stored in the refrigerator using voice commands.
- Natural language recognition in-vehicle infotainment systems
In-vehicle infotainment systems equipped with the NLP engine can perform the specific action through voice. For example, drivers can play songs using voice commands.
