Introducing gecko2vec

gecko2vec is an embedding software based on mol2vec. It is an application of the word2vec algorithm to atmospheric molecules representation.

What is gecko2vec?

The project builds a large and unique database of atmospheric molecules, where embedding representations retain information on:

• Molecular structures (functional groups distribution)
• Chemical compositions
• Inter-atomic connectivity

Why molecular embeddings?

Traditional molecular representations often fail to capture the rich structural information needed for accurate property prediction. By borrowing techniques from natural language processing, we can create dense vector representations that encode molecular similarity in a mathematically useful way.

Applications

These embeddings allow further investigation of molecular properties via machine learning algorithms, opening new possibilities for:

• Property prediction
• Similarity searching
• Reaction classification
• Aerosol chemistry modeling

This project laid the foundation for subsequent work on glass transition temperature prediction and mechanism synthesis.