Jeffrey A. Bilmes

Also known as Jeff Bilmes, Jeff A. Bilmes

Jeff Bilmes is a professor of electrical and computer engineering at the University of Washington whose machine learning work centers on submodular optimization and graphical models. After a Berkeley B.S. in 1989 and an MIT S.M. in 1993, he completed a Berkeley Ph.D. in computer science and joined the Washington faculty in 1999. Starting in 2003, he developed submodularity within machine learning, and he wrote the Graphical Models Toolkit (GMTK), software used in speech, language, bioinformatics and human-activity recognition. He won best paper awards at ICML and NIPS in 2013, and ICML named Deep Canonical Correlation Analysis, his 2013 paper with Galen Andrew, Raman Arora and Karen Livescu, a Test-of-Time runner-up in 2023. In 2023 he wrote the Submodular Optimization chapter of Kevin Murphy's Probabilistic Machine Learning: Advanced Topics, and in 2024 he coauthored the NeurIPS spotlight paper Deep Submodular Peripteral Networks.

Now at: University of Washington

Timeline

1989–1999

Degrees from Berkeley and MIT

He earns a B.S. in electrical engineering and computer science at UC Berkeley in 1989, an S.M. at MIT in 1993, and a Ph.D. in computer science at Berkeley in 1999. He also works as a researcher at the International Computer Science Institute in Berkeley.
1999–2002

Joining Washington's electrical engineering faculty

He joins the faculty of the University of Washington's Department of Electrical & Computer Engineering in 1999. In 2001, he leads a group at the Johns Hopkins summer workshop on automatic speech recognition and wins an NSF CAREER award. In 2002, he becomes a CRA Digital Government Fellow.
Since 2003

Submodular learning and GMTK

In 2003, he starts developing submodularity within machine learning. He also authors the Graphical Models Toolkit (GMTK), a dynamic graphical-model software system widely used in speech, language, bioinformatics and human-activity recognition.
2008–2010

Max Planck sabbatical, UAI chair

He spends the 2008/2009 academic year on sabbatical at the Max Planck Institute for Biological Cybernetics. He is program chair of the Conference on Uncertainty in Artificial Intelligence in 2009 and its general chair in 2010.
2013–2014

Best papers at ICML and NIPS

In 2013, he receives a best paper award at ICML and another at NIPS for work on submodularity in machine learning. In 2014, the International Conference on Supercomputing gives him a most influential paper in 25 years award.
2016–2017

Visiting professor at Google Research

He spends 2016–2017 as a visiting professor and research scientist at Google Research. In 2016, he also gives a tutorial at the Computer Vision and Pattern Recognition conference (CVPR) on optimization algorithms for subset selection and summarization in large data sets.
2023–2024

Test-of-Time runner-up at ICML

ICML 2023 names his 2013 paper Deep Canonical Correlation Analysis, written with Galen Andrew, Raman Arora and Karen Livescu, a Test-of-Time runner-up; the method learns nonlinear transformations of two views of data whose representations are highly linearly correlated. In 2023, he also writes the Submodular Optimization chapter of Kevin Murphy's Probabilistic Machine Learning: Advanced Topics, published by MIT Press. In 2024, he coauthors Deep Submodular Peripteral Networks with Gantavya Bhatt and Arnav Das, a NeurIPS 2024 spotlight.
As of October 2026

Professor at the University of Washington

He is a professor in the University of Washington's Department of Electrical Engineering, with adjunct appointments in Computer Science & Engineering and Linguistics. He is the founder of the MELODI (MachinE Learning for Optimization and Data Interpretation) lab there.