Building Intelligent Probabilistic Systems

In the HIPS group, we are interested in building intelligent algorithms. What makes a system intelligent? Our philosophy is that "intelligence" means making decisions under uncertainty, adapting to experience, and discovering structure in high-dimensional noisy data. The unifying theme for research in these areas is developing new approaches to statistical inference: uncovering the coherent structure that we cannot directly observe and using it for exploration and to make decisions or predictions. We develop new models for data, new tools for performing inference, and new computational structures for representing knowledge and uncertainty.

Recent News

Netflix Using Spearmint for Bayesian Optimization


As reported by Wired magazine and on the Netflix tech blog. Netflix has been experimenting with deep learning tools for making recommendations. Moreover, they've been using our software Spearmint to set the hyperparameters with a cluster of machines on Amazon EC2.

Collaboration on the Cover of Science Translational Medicine


A collaboration between UCLA bioengineers and HIPS group members Yo Sup "Joseph" Moon and Ryan Adams has just published new results in the AAAS journal Science Translational Medicine. The paper, "Quantitative Diagnosis of Malignant Pleural Effusions by Single-Cell Mechanophenotyping" develops a new way to identify whether cells are cancerous or not, using a combination of microfluidic measurement and machine learning techniques. This work appeared on the cover of STM.

Finale Doshi-Velez named among "AI's 10 to Watch"


Postdoctoral researcher at SEAS and HMS recognized for impressive research contributions to artificial intelligence

Cambridge, Mass. - July 29, 2013 - Our very own Finale Doshi-Velez, a postdoctoral research associate at the Harvard School of Engineering and Applied Sciences (SEAS) and at Harvard Medical School (HMS), has been named among "AI's 10 to Watch" by IEEE Intelligent Systems.

Announcing a Group Blog


We're excited to announce a new collaborative blog, written by members of the HIPS group. The idea is to use this as a venue to discuss interesting ideas and results — new and old — about probabilistic modeling, inference, artificial intelligence, theoretical neuroscience, or anything else research-related that strikes our fancy. There will be posts from folks at both Harvard and MIT, in computer science, mathematics, biophysics, and BCS departments, so expect a wide variety of interests.

Recent Publications

Message Passing Inference with Chemical Reaction Networks. Napp, N, Adams RP.  Advances in Neural Information Processing Systems 26. 2013.   { PDF }
Multi-Task Bayesian Optimization. Swersky, K, Snoek J, Adams RP.  Advances in Neural Information Processing Systems 26. 2013.   { PDF }
Contrastive Learning Using Spectral Methods. Zou, J, Hsu D, Parkes D, Adams RP.  Advances in Neural Information Processing Systems 26. 2013.   { PDF }
Computational Caches. Waterland, A, Angelino E, Cubuk ED, Kaxiras E, Adams RP, Appavoo J, Seltzer M.  International Systems and Storage Conference (SYSTOR). 2013.   { PDF }