Identifying New Materials with NLP with Anubhav Jain - TWIML Talk #291
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)
00:00
Unveiling Chemical Connections through Word Embeddings
This chapter explores how Word2Vec can be used to visualize the periodic table by generating 200-dimensional vectors for chemical elements. It highlights the model's ability to capture scientific relationships and predict new materials through dimensionality reduction techniques like T-SNE and PCA, ultimately revealing insights into elemental properties and research gaps in material science.
Transcript
Play full episode
