Using a customized neural network designed for episodic paleography, researchers have reconstructed a series of charred birch-bark fragments from a site in the northern Himalayan foothills. The fragments belong to a previously unknown 7th-century Sanskrit treatise titled Jala-Gati-Sutra, which details advanced models for hydro-dynamic flow and sediment management in large-scale canal systems.
The text provides mathematical rules for calculating the velocity of water through varying channel geometries to prevent siltation, a problem that plagued ancient irrigation projects. The AI reconstruction revealed that the authors used a precursor to modern fluid dynamics equations, suggesting that ancient Indian engineers possessed a formal theoretical framework for the complex hydraulic systems found in historical city planning.