Textbook of AI

1973–, Computer scientist, statistician

Also known as: David M. Blei

David Meir Blei is an American computer scientist and statistician whose 2003 paper with Andrew Ng and Michael Jordan, Latent Dirichlet Allocation, introduced LDA, a probabilistic generative model for collections of documents. LDA represents each document as a mixture over latent topics, each topic as a distribution over words, and provides a principled way to discover thematic structure in unlabelled text corpora. The paper has been cited tens of thousands of times and shaped a decade of NLP research.

Blei has been a leading figure in probabilistic machine learning: variational inference (his work on stochastic and amortised variational inference is a central tool in modern Bayesian deep learning), nonparametric Bayesian methods, exponential family embeddings, and the broader cause of probabilistic programming. He spent his career at Princeton and now Columbia, where he directs research that consistently bridges statistics, computer science and applied domains including political science and history.