Multi-agent information analysis system designed to collect, analyze and visualize open-source and limited access text and multimedia information SIA.ATTACHE
The development of information and communication technologies has led to exponential growth of information in all areas of human activity. In this regard, it becomes relevant to develop a service that allows you to interactively search for information throughout the entire knowledge base, highlight entities (author, technology, organization, etc.), and obtain statistical data upon request in real time.
Artificial Intelligence And Digital Services
Big data storage and analysis
Depends on the complexity of the planned analytical studies: from one to nine months
For today - Russia. There are plans to enter the markets of the BRICS countries
In the scientific research field, analogues are information and analytical add-ons (for example, SciVal) over the abstract databases Scopus and Web of Science. Compared with the proposed system, they have a number of disadvantages: - Information and analytical add-ons statistically analyze the available, well-structured information from scientific articles (authors, organizations, etc.) and do not analyze full texts; - These tools only work with specific reference databases and do not allow to work with your own document collections; - The systems do not have text recognition modules; - They are available only with a subscription to abstract databases (it is financially inaccessible to many universities, for a large university it ranges from several million to tens of millions of rubles per year, depending on the subscription contents); - These databases and tools are located in the clouds in the jurisdiction of third countries. Consequently, they can be used by these countries to collect information about the formation of collections by scientists and their search queries. In the analytical sphere, analogs can be systems designed to conduct OSINT research (for example, IBM I2, Palantir, Semantic Archive), solving a wide range of tasks of identifying entities, data analysis, etc. However, these systems are inferior to the proposed software product in some respects: - Since the above products are intended for professional OSINT research, they require a serious level of operator training, therefore the entry barrier for these systems is high; - The systems require significant labor costs to configure the subject area, namely, the construction of the entity tree, which is also the “bottleneck” of the software, since it is necessary to find a balance between the detailing of information and the level of its abstraction; - The systems do not have text recognition modules; - The implementation of professional StandAlone OSINT systems, due to their specificity, is expensive for the customer. For example, a corporate concurrent license for IBM I2 with an acceptable set of tools will cost the user from 3.5 to 6.5 million rubles (not counting the costs of implementation and personnel training). Some systems are cheaper to license and implement, but the costs of customizing the subject area are higher; - At the moment, access to Western OSINT systems for some countries is limited for political reasons. In addition, information in cloud systems (for example, Palantir) can be used in the interests of third countries.
SIA LLC has been awarded with and successfully implemented FASIE (Foundation for Assistance to Small Innovative Enterprises) grant to create and develop a product within "Start" program
know-how: Component architecture of a multi-agent system for processing and presenting unstructured data arrays. Three PhD theses. More then 15 scientific articles.
