ABBYY Compreno Technology: Key Components

Technology & Features

ABBYY Compreno uses three major components

  • Semantics - in the form a language-independent hierarchy of concepts
  • Syntax - i.e. ability to understand how concepts relate to one another within one or more sentences)
  • Statistical data - which is used for combining words into natural-sounding sequences and as an aid in sense disambiguation.

1. Semantics: getting to the very meaning

Semantics is a branch of linguistics that deals with meanings. Universal Semantic Hierarchy - a hierarchy of meanings believed to be common for most natural languages - builds the core of ABBYY Compreno Technology. While analyzing the text, ABBYY Compreno Technology identifies the exact position of each word in the hierarchy, thus defining its meaning and the set of concepts semantically connected to it (e. g. for “Labrador”: “dog”, “animal”, “dachshund” etc.). This way, the quality of machine text understanding is language- and style-independent: Compreno Technology can identify identical and similar meanings in texts using different vocabulary, as well as in multilingual documents or document sets.

2. Syntax: reveal context-dependent relations between the meanings

A sophisticated Syntactic Parsing Technology defines the grammatical role of each word in a sentence and analyzes all interrelations between words within one sentence and between sentences. Syntactic parsing complements Universal Semantic Hierarchy adding context-specific non-hierarchical connections between the concepts within and between the sentences in the text.

Syntactic parsing allows to resolve multiple ambiguity cases which cause challenges in machine text understanding, such as:

  • Homonymy / polysemy disambiguation
    Example: can (Verb) and can (Noun)
  • Anaphora resolution
    Example: After Peter had waited for 30 minutes, he (Peter) went home
  • Sentence part omission (ellipsis) resolution
    Example: He says he has lost the money, but I don't think he has. (lost his money)

3. Statistics: relevant support for the language knowledge

Last but not least, statistical analysis enhances the quality of context-dependent synonymy/homonymy disambiguation.

Further Info

This website uses cookies which enable you to see pages or use other functions of our websites. You can turn off such cookies in your browser’s settings. If you continue to use these pages, you consent to the use of cookies.