Miso Information Technology Wins Contract to Develop AI for Public R&D Budget Review Based on Its Own Foundation Model
TECHWORLD ·
✦ AI Summary
Miso Information Technology announced on the 3rd that it has won the Ministry of Science and ICT's project to develop an AI system for public research and development (R&D) budget review based on an independent foundation model.
The project aims to build a "National R&D Budget Review Support AI" system that analyzes data from the government R&D budget allocation and adjustment process with AI to support the preparation of review materials and deliberation documents.
Miso Information Technology plans to connect the entire budget review process into a single integrated workflow by linking RAG and LLM technologies with its own platforms, and to expand the model into the public and policy sectors.
Miso Information Technology announced on the 3rd that it has won the Ministry of Science and ICT's project to develop an AI system for public research and development (R&D) budget review based on an independent foundation model.
The project aims to build a "National R&D Budget Review Support AI" system that helps produce review materials and deliberation documents through AI analysis of data used in the government R&D budget allocation and adjustment process. Public R&D budget review requires a comprehensive examination of a wide range of data, including budget requests by project, project information, past review materials, meeting materials, and policy documents.
Miso Information Technology will build a system that links multiple documents, review histories, and external information, and can search and analyze the evidence needed. The system will apply retrieval-augmented generation (RAG) and large language model (LLM) technologies.
Based on document processing, chunking, embedding, and reranking integration, the system will search for relevant information and help produce budget review documents such as review opinions and adjustment result reports based on the search results.
For this project, Upstage's "Solar" will be used as the independent foundation model, and Miso Information Technology's in-house connected technologies — the multimodal big data platform "SmartBig," the data review solution "SmartDRB," the VLM-based OCR solution "Bigs," and the LLMOps and AI orchestration platform "Actic" — will be combined to connect the entire budget review process into a single integrated workflow.
The areas to be connected are data collection, management, search, analysis, AI review, document generation, user review, approval, and results management. The implementation feature is analysis of project-level similarity and duplication.
The linked data will include budget request details, past review history, and external R&D information. Using these together, the company said it expects to reduce repetitive document review work, and the support goal is to let staff focus on key issues and policy judgments.
The company plans to use this project as a steppingstone to expand its public AI business. The expansion target is support for budget, policy, and administrative work at government agencies and public institutions.
Nam Sang-do, CEO of Miso Information Technology, said the project goes beyond using generative AI as a simple work support tool and brings it into specialized tasks for national R&D budget review, adding that it will become an example of improving the efficiency and expertise of decision-making. He also said Miso Information Technology will build a trustworthy AI-driven work innovation model in the public sector based on its data platform and AI full stack, and expand it into the public and policy sectors.
Source: TECHWORLD · Kim Seung-ki
Original: https://www.epnc.co.kr/news/articleView.html?idxno=406482
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.