Kevin P. Murphy
Also known as Kevin Murphy
Kevin Murphy is a machine-learning researcher whose work connects Bayesian networks, probabilistic learning, and machine perception. He studied computer science at Cambridge and Pennsylvania, moved from doctoral study at UC Davis to Berkeley, and developed inference methods for dynamic Bayesian networks in his 2002 thesis. He released the Bayes Net Toolbox for Matlab in 2001, joined UBC in 2004, became a tenured associate professor there, and joined Google as a research scientist in 2011. Machine Learning: A Probabilistic Perspective was a 1,100-page textbook that won the 2013 DeGroot Prize; he co-led JMLR from 2013 to 2017 and presented Google's DeepLab semantic-segmentation system. He published Probabilistic Machine Learning books in 2022 and 2023, spoke about agentic AI in 2025 and 2026, and is an adjunct professor at UBC as of October 2026.
- Now at
- University of British Columbia and Google
- Works
- 93
- Citations
- 53,545
- h-index
- 54
Publication numbers from OpenAlex, as of 15 October 2026.
Timeline
-
1989–1997
Cambridge and Pennsylvania degrees
He studies computer science at Cambridge and completes an MSE in computer science at the University of Pennsylvania, then begins doctoral work at UC Davis before transferring to Berkeley. His early research focuses on probabilistic Bayesian networks for temporal sequences; in 1997, he coauthors “Space-efficient inference in dynamic probabilistic networks” with John Binder and Stuart Russell.
-
2001–2002
Bayes Net Toolbox and thesis
He releases the Bayes Net Toolbox for Matlab in 2001. His 2002 UC Berkeley thesis studies dynamic Bayesian networks and introduces factored-frontier approximate inference.
-
2004–2011
From UBC to Google
After a postdoctoral appointment at MIT, he joins the University of British Columbia in 2004. Before joining Google in 2011, he becomes a tenured associate professor in computer science and statistics; at Google, he works on AI, machine learning, and computer vision.
-
2012–2013
Machine Learning and a DeGroot Prize
In 2012, he publishes Machine Learning: A Probabilistic Perspective, a 1,100-page textbook. It wins the 2013 DeGroot Prize for best book in Statistical Science.
-
2013–2017
Co-leading JMLR and DeepLab
In March 2013, he becomes a co-editor-in-chief of the Journal of Machine Learning Research and steps down in spring 2017. At Google, he presents his team's DeepLab system for semantic segmentation alongside object-detection, instance-segmentation, and pose-estimation systems.
-
2022–2023
Probabilistic Machine Learning books
In March 2022, he publishes Probabilistic Machine Learning: An Introduction, which connects classical statistics and probability with deep learning. In 2023, he publishes Probabilistic Machine Learning: Advanced Topics, covering inference methods, generative models, and decision making.
-
2025–2026
Agentic AI talks
In 2025, his talks cover generative AI, agentic AI, and sequential decision making. In April 2026, he presents “Code Synthesis for improved agentic decision making” at the ICLR World Models Workshop.
-
As of October 2026
Adjunct professor at UBC
As of October 2026, he is an adjunct professor in UBC's Department of Computer Science, with machine intelligence and machine perception among his research areas.
Sources and credits
Photo: We have no published photo of Kevin P. Murphy yet. Send us a free-licensed photo.
- Kevin Patrick Murphy CV
- Dynamic Bayesian Networks: Representation, Inference and Learning
- Publications by Kevin Murphy
- Data Science Institute Colloquium: Kevin Murphy | Research Scientist, Google (Keynote)
- DeGroot Prize | International Society for Bayesian Analysis
- History of JMLR
- Probabilistic Machine Learning: An Introduction
- Probabilistic Machine Learning: Advanced Topics
- Some recent talks by Kevin Murphy
- Kevin Murphy | Computer Science at UBC
- Kevin P. Murphy | Google Research