Rural Development Administration, first step in ‘AI for agricultural research collaboration’… Study of 36,000 papers, etc.

The Rural Development Administration announced on the 9th that it had succeeded in developing a large language model (LLM) specialized in agricultural life using a graphics processing unit (GPU) supported by the Ministry of Science and ICT.

Based on this, the Rural Development Administration began developing ‘artificial intelligence (AI) for research collaboration’.

The researchers trained AI with approximately 20,000 academic papers collected through agreements with five agricultural and life academic societies, approximately 13,000 papers from international academic journals from the Rural Development Administration, and approximately 3,000 research reports. The Agricultural Life LLM developed in this way presents evidence and sources along with the answers.

The Rural Development Administration plans to install the developed LLM in the Agricultural and Biotechnology Information Center (NABIC) and run it on a pilot basis within the agency since last October, and to make it public starting in January of next year.

We are also working on developing AI for research collaboration with the goal of completion in 2028.

While the LLM specializing in agricultural life is the stage of answering researchers' questions with evidence, AI for research collaboration is with researchers throughout the entire research process, from analyzing agricultural big data to deriving research hypotheses and analyzing and interpreting results.

Lee Tae-ho, head of the Supercomputing Center, said, "With support from the Ministry of Science and ICT's graphics processing unit, we were able to complete an LLM specialized in agricultural life that learned agricultural life data directly within the institution. Based on this, we will also speed up the development of AI for research collaboration that designs and solves research together with researchers."