RCO research history

RCO core team has 10+ years of background in computational linguistics, specializing in natural languages processing. First steps with statistical and logical modeling of natural language have been made when studying in Moscow State University (Departments of Computational Mathematics and of Theoretical and Applied Linguistics).

In 1996 first publications appear. That time the statistical approaches based on neural networks have been a main subject of RCO team scientific interest, resulting in "associative semantic networks" designed to extract essence out of unstructured text and to find its key topics and relationships. These networks have been considered as a computer model of information processing by the right hypersphere of human brain. Corresponding approach to text processing was mostly statistical and language independent.
Later some applications based on this approach have been developed: automatic categorization and clustering, search query expansion, text summarizing.

By 2001 the evolution of algorithms on text processing scene resulted in RCO team decision to re-design some of its core technologies from scratch, and to create several new ones, implementing latest known algorithms and approaches as well as adding own.
This way, by 2002 the new set of technologies appeared, performing complete text processing - featuring morphological, syntactic and semantic analysis, performing anaphora resolution and many other options.

Since 2003, RCO is one of co-founders and active participants of Russian Information Retrieval Evaluation Seminar.

In 2005-2006 RCO had successfully adapted its algorithms to support English language. Resulting product, RCO Fact Extractor English edition, exceeded expectations demonstrating the level of performance and quality among the best competitors on market.

The scientific record made by RCO team so far is described in more than 50 publications.

RCO products history

The first product had been announced in 1997 - Russian ConText Optimizer for Oracle database. This product initially was considered to be a kind of extension to native Oracle Text option, adapting its facilities for Russian language. In past 10 years 4 new versions of RCO for Oracle had been released. Recent versions of RCO for Oracle offer far more features than only localization for Russian language.
In 2006 RCO for Oracle was adopted to work with Ukrainian language and this way UCO (Ukrainian Context Optimizer) for Oracle appeared.

Following success of RCO for Oracle, the quite similar product for Microsoft platform had been developed, first announced in 2000 and also adopted for Ukrainian language in 2006. This product is called RCO for MS Backoffice, reflecting its ability to optimize text search features not only in SQL Server, but also in other server products - Microsoft SharePoint Portal, Microsoft Exchange Server.

Starting in 2001 and later RCO announced a set of products targeted at text analysis itself, stepping beyond search-only applications. These were products performing full-featured syntactic and semantic analysis of text, including anaphora resolution, for Russian language. This set of products was addressed to developer’s community and was available as SDKs: RCO Morphology SDK, RCO Pattern Extractor SDK, RCO Semantic Network SDK, RCO Entity Extractor SDK and RCO Fact Extractor SDK. This approach was successful - many ISVs had built-in RCO products into their own products and solutions.

In 2005-2006 two major releases announced - RCO KAOT and RCO Fact Extractor. Both addressed to customers seeking for complete text analysis solution, available both in desktop and server versions.