RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2
Project

A prototype of a system for analyzing regulatory documents for Roscosmos based on the Preferentum AI platform has been developed

Customers: Roscosmos State Corporation for Space Activities

Moscow; Space industry

Product: Preferentum. Unstructured Information Analysis Platform

Project date: 2024/05  - 2024/11

2024: Completion of the project to create an intelligent control system for the body of regulatory documents

TsNIImash and SL Soft completed the first stage of the project to create an intelligent management system for the body of regulatory documents (NSD) - international, all-Russian, industry standards. Softline announced this on December 6, 2024.

A prototype of a system for analyzing technical standards for contradictions and inconsistencies - SATSPIN was implemented. This tool is already being tested by employees of the central office of Roscosmos and at the enterprises of the corporation. The prototype provides a quick search for relevant technical information, checking the correctness of terminology, analyzing the compliance of new regulations and terms of reference with current standards and requirements. The introduction of a fully functional system will significantly improve the quality and speed of design, minimizing the risk of errors at all stages of the design cycle.

SATSPIN is implemented on the basis of the intellectual document analytics platform and the construction of the corporate knowledge base "Preferentum." The digital tools of the platform easily adapt to the specifics of the tasks of any industry, ensuring high accuracy of processing of subject area information through automated construction of linguistic and functional ontologies.

The prototype provides a centralized, intelligent repository of NSD's interrelated domain-specific requirements (knowledge base) and has the ability to integrate with external information systems such as PLM systems or CAD/CAE systems. Using a virtual assistant, Roscosmos specialists can enter requests in a natural language and receive reliable and up-to-date information in a matter of seconds. The system effectively copes with inaccuracies and errors in queries and supports search in any morphological forms, cross-language search, can automatically expand the query with synonyms and associatively close concepts. All this ensures high quality information availability and accuracy.

When new NSDs are loaded, the prototype automatically analyzes the documents, identifying the requirements and definitions of concepts contained in them. Along with this, the system forms an automatically replenished glossary and builds a "mental map" based on it, which greatly simplifies the work with terminology. SATSPIN automates the management of the knowledge base, adapting to changes and additions in regulatory documents.

The key ability of the system is the ability to compare individual requirements for inconsistencies and contradictions, as well as check whether the draft document meets the requirements of NSD. Thus, the implementation of SATSPIN will eliminate contradictions, inaccuracies and inconsistencies in the regulatory documentation and will serve as the next step towards the formation of uniform and clear standards for the design and operation of space complexes.

File:Aquote1.png
The SATSPIN system clearly demonstrates how artificial intelligence can transform the analysis of regulatory documents and its standardization. This is not just a saving of time, but also an increase in design accuracy, which is critical for the rocket and space industry. We expect that the introduction of the system will radically reduce the labor costs of Roscosmos specialists and its subsidiaries for design, increase the consistency of the requirements of technical documentation, - explained Dmitry Chernov, Deputy General Director for Digital Transformation of TsNIImash JSC.
File:Aquote2.png

File:Aquote1.png
The AI platform "Preferentum" provides exceptional processing accuracy, which data is achieved through a flexible combination of language models (LLM), (machine learning ML) and intelligent, algorithms working in unified analysis processes. information The systems implemented on the basis of our product guarantee predictability of work and transparency of decision-making. artificial intelligence The introduction of Preferentum allows companies to effectively manage knowledge by providing employees with quick access to the necessary information and automation, business processes for example, checking technical documentation for NSD compliance or the accuracy and relevance of the reference information base, "explained SL Dmitry Romanov Soft, Managing Director of Preferentum.
File:Aquote2.png