Apply general AI+ technology to various scenarios for business solution.
There are two types of text information: unstructured and structured. The former type, such as a CV file, a passage of judgment, is not comprehensible information for machines to learn and analyze. With Text Info Extraction, it is converted to structured data which machines could read, understand and analyze. This is how unstructured information becomes valuable.
Natural language description
Structured information
Text Classification extracts features from descriptive text, semantically comprehends it, and finally classify it. For example, using Text Classification, machines can tell occupation according to text of work experience, can determine case category according to text of case description, can distinguish architecture type according to text of architectural description.
Natural language description
Classification information
With OCR (Optical Character Recognition) technology, Visual Information Extraction extracts print letters as image letters and converts them to text format for machines to process. It could be used to extract information from picture CV, picture business card and picture of listed items, etc.
Picture of text
Text information
Semantic Comprehension is the ability to understand meanings of words and sentences in the context. For example, "Apple", in some context, means a kind of fruit, while in others means the famous tech company. In addition,Semantic Comprehension allows machines to associate and comprehend ideas, understanding meanings behind the surface. For example, in the sentence "hiring UX designer with big IT company experience", Semantic Comprehension can tell what kind of company is "big IT company" and know "UX designer" also means user experience designer or UE designer. Traditional keyword searching is not comparable to this stage of intelligence.
Relevant Search can retrieve and rank the most relevant text information from a collection of natural language documents with certain trained models.
Applied scenarios examples:
- Recommend the most suitable candidates for a job description
- Recommend the most suitable job openings for a candidate
- Find the most similar cases, judgments and clauses
- Recommend the most suitable lawyer to win a case
