Picture of Brian Ziebart

Brian Ziebart

I conduct machine learning research on predictive models of purposeful adaptive behavior. I am a PhD candidate in the Machine Learning Department at Carnegie Mellon University under the advisement of Anind Dey and Drew Bagnell. Here you will find information about myself and my research.

: Research :: Projects :: Publications :: Miscellaneous :


Research Summary


Computer systems that can understand (and predict) human behavior will help to improve our interactions with the technologies that touch our everyday lives.

My research explores a maximum entropy/decision theoretic approach for probabilistically modeling human behavior, and applies this approach to problems in vehicle navigation, pedestrian/robot movement, and activity forecasting.

Current Projects


Pedestrian Trajectory Forecasting

Using the recent position track of a person, we can predict their intended destination and efficiently forecast their future position over time.

We use these forecasts to plan robot trajectories that do not hinder people.

Assistive Navigation Systems

Existing personal navigation devices (PNDs) are largely oblivious - they "ride along" with drivers, but ignore everything about how the drivers actually drive.
We develop navigation technology that reasons about observed driving to learn the driver's preferences and predict their future paths (to warn about unanticipated hazards).

Previous Projects


Bayes Net Structure Learning

In the structure learning setting, the relationships between variables (of a Bayesian Network) are learned from data.
In this work, polynomial inference algorithms (MAP structure estimation and Bayesian model averaging) for selectively conditioned forests (c) -- the combination of tree structures (a) and ordered variable sets (b) -- are presented, enabling generalized naïve Bayes classifiers.

Pervasive Computing Environments


A long time ago, I developed distributed applications for a middleware-based meta-operating system, and became interested in how machine learning could be used to automatically configure such systems.

Publications

(reverse chronological)


Planning-based Prediction for Pedestrians
B. D. Ziebart, N. Ratliff, G. Gallagher, C. Mertz, K. Peterson, J. A. Bagnell, M. Hebert, A K. Dey, S. Srinivasa
International Conference on Intelligent Robots and Systems (IROS 2009). [pdf]

Inverse Optimal Heuristic Control for Imitation Learning
N. Ratliff, B. D. Ziebart, K. Peterson, J. A. Bagnell, M. Hebert, A K. Dey, S. Srinivasa
Artificial Intelligence and Statistics (AISTATS 2009). [pdf]

Human Behavior Modeling with Maximum Entropy Inverse Optimal Control
B. D. Ziebart, A. Maas, J. A. Bagnell, A K. Dey
AAAI Spring Symposium on Human Behavior Modeling. 2009. [pdf]

Navigate Like a Cabbie: Probabilistic Reasoning from Observed Context-Aware Behavior
B. D. Ziebart, A. Maas, A. K. Dey, and J. A. Bagnell.
International Conference on Ubiquitous Computing (Ubicomp 2008). [pdf]

Fast Planning for Dynamic Preferences
B. D. Ziebart, A. K. Dey, and J. A. Bagnell.
International Conference on Automated Planning and Scheduling (ICAPS 2008). [pdf]

Maximum Entropy Inverse Reinforcement Learning
B. D. Ziebart, A. Maas, J. A. Bagnell, and A. K. Dey.
AAAI Conference on Artificial Intelligence (AAAI 2008). [pdf]

Learning Selectively Conditioned Forest Structures with Applications to DBNs and Classification
B. D. Ziebart, A. K. Dey, and J. A. Bagnell.
Uncertainty in Artificial Intelligence (UAI 2007). [pdf]

Learning Automation Policies for Pervasive Computing Environments
B. D. Ziebart, D. Roth, R. H. Campbell, and A. K. Dey.
IEEE International Conference on Autonomic Computing (ICAC 2005).

Towards a Pervasive Computing Benchmark
A. Ranganathan, J. Al-Muhtadi, J. Biehl, B. Ziebart, R. H. Campbell, and B. Bailey.
PerWare '05 Workshop on Support for Pervasive Computing at PerCom 2005.

Learning Context-Dependent User Preferences in a Ubiquitous Computing Environment
B. Ziebart and D. Roth.
Undergraduate Thesis. Department of Electrical and Computer Engineering. University of Illinois at Urbana-Champaign. May 2004.

System Support for Rapid Ubiquitous Computing Application Development and Evaluation
M. Roman, J. Al-Muhtadi, B. Ziebart, and R. H. Campbell.
Systems Support for Ubiquitous Computing Workshop, at UbiComp 2003. [pdf]

Dynamic Application Composition: Customizing the Behavior of an Active Space
M. Roman, B. Ziebart, and R. H. Campbell.
IEEE International Conference on Pervasive Computing and Communications (PerCom 2003). [pdf]

(Note: Some of this material is Copyright IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE)


Miscellany
I am a co-founder of NavPrescience, a CMU spin-off company developing personalization and prediction technologies for next generation GPS devices.

Gates-Hillman Predictive Market: I finished in 1st place (of 210) with 15.09% of the tickets by maximizing the (relative) entropy (i.e., buying low and selling high).

Bid-Tac-Toe: Exponentiated stochastic gradient ascent + logistic regression over pools of (approximate) equilibria strategies = $2500. However, winning a hedge fund's recruitment programming contest + market meltdown &ne lucrative quant career (yet..?).