Anatoly Starostin. Using ABBYY Compreno technology for solving various NLP-tasks
Bio: Anatoly Starostin is the Head of Semantic Analysis Group at ABBYY Russia. He holds a Master’s degree awarded by the Moscow State University (Faculty of Computational Mathematics and Cybernetics). He is currently working on a PhD thesis in the field of natural language syntactic parsing.
The presentation provides an overview of ABBYY Compreno, a linguistic technology on which ABBYY has been working for the past 15 years. Originally the technology was geared towards machine translation, and it is now being successfully applied to MT tasks (the technology is used to translate texts from English into Russian and from Russian into English). In developing ABBYY Compreno, ABBYY relied on detailed descriptions and modelling of a large number of natural language phenomena. The technology provides tools for creating formal morphological,
syntactic, and semantic descriptions.
The key feature that sets ABBYY Compreno apart from other MT and, more generally, NLP solutions is its scope—the technology employs an unprecedented number of integrated formal tools. The main objective of this presentation is to demonstrate how this set of solutions can be used as an all-purpose tool for tackling NLP tasks. It should be noted that some of the ABBYY Compreno components, such as algorithms and formal languages, have counterparts in the
literature. References to those will be made during the presentation.
The following ABBYY Compreno components will be dwelt upon:
- Language-independent hierarchy of semantic classes (notions)
- Syntactic-semantic analysis mechanism
- Use of non-tree links in syntactic-semantic analysis
- Handling word combinations
- Handling ellipsis
In the final part of the presentation, we will explore the possible applications of ABBYY to text processing tasks other than machine translation.