Wed Nov 20 14:27:03 UTC 2024: **IISc Researchers Develop New Language to Predict Nanopore Properties**
**Bengaluru, November 21, 2024** – Researchers at the Indian Institute of Science (IISc) in Bengaluru have developed a novel language, called STRONG (STring Representation Of Nanopore Geometry), to describe the shape and structure of nanopores. This breakthrough, published in the Journal of the American Chemical Society, allows machine learning models to accurately predict the properties of nanopores across various materials.
STRONG uses a unique alphabet where different letters represent different atom configurations at the edge of a nanopore. This allows for a concise representation of complex nanopore structures. The system significantly reduces the amount of data required for analysis by identifying functionally equivalent nanopores, even those related by rotation or reflection.
The IISc team utilized a neural network, similar to those used in natural language processing, to interpret the STRONG language. This network was trained on nanopore structures with known properties, enabling it to predict the properties of new, unseen structures. This approach offers a significant advancement over traditional programming methods, allowing for the prediction of properties like energy of formation and gas transport barriers.
This development also opens the door for reverse engineering nanopores – designing structures with specific desired properties, a crucial step in applications like gas separation. The researchers’ work represents a significant contribution to materials science and nanotechnology.
First Piper 