Researchers using advanced neural network models have successfully deciphered a series of fragmented 9th-century birch-bark manuscripts in the Sharada script recovered from the Gilgit-Baltistan region. The texts, translated into modern Sanskrit and English, are part of a previously unknown collection detailing avian physiology and the seasonal migratory patterns of waterfowl across the Pamir Mountains. The manuscripts describe internal anatomical structures and suggest that ancient Indian naturalists were monitoring bird migrations as indicators of high-altitude weather changes.
The AI model utilized a deep-learning architecture to bridge gaps in the damaged scrolls, revealing detailed protocols for the medicinal treatment of birds used in royal falconry and agriculture. This discovery emphasizes the breadth of early Indian biological sciences, extending beyond traditional human medicine into specialized zoological and ethological observations that were systematically documented over a millennium ago.