Researchers using machine learning algorithms have successfully deciphered a collection of fragmented 8th-century manuscripts written in the Kutila script, found in a remote temple archive in Himachal Pradesh. The texts contain lost Sanskrit treatises detailing sophisticated observations on soil chemistry, including a functional understanding of nitrogen fixation through the use of specific leguminous crops. This suggests a highly systematic approach to agricultural science in early medieval India.
The AI-driven reconstruction of the text, titled 'Kshiti-Vigyan', outlines protocols for soil revitalization that mirror modern ecological practices. The decipherment was made possible by a new neural network model designed to recognize the ligatures of transitional northern scripts, marking a major breakthrough in the study of ancient agrarian knowledge.