daniel.grollman AT epfl.ch

Home Publications Videos Code Media CV/Resume PGP

These pages are out of date, I have moved to Vecna Robotics



Short Bio:

Currently I am a postdoc at for Aude Billard at EPFL in the Learning Algorithms and Systems Laboratory (LASA). I completed my PhD in 2009 in the Department of Computer Science at Brown University. There I worked with Odest Chadwicke Jenkins as a member of RLAB (Robotics, Learning and Autonomy at Brown). I also got my Masters of Science in Computer Science from Brown in 2005, after getting my Bachelors of Science in Electrical Engineering and Computer Science from Yale University in 2003. There I worked with Brian Scassellati, as part of the Social Robotics Lab. During the summers I have worked variously with Gillian Hayes at the Mobile Robotics Group of the University of Edinburgh School of Informatics, Brian Yamauchi at iRobot Corporation, and Carsten Magerkurth at the Fraunhofer Institute (AMBIENTE).

Research Summary:

In my research I consider Robot Learning from Demonstration (RLfD) as a means of performing Human-Robot Policy Transfer. The goal is to allow a human user to instantiate, on a robot, an autonomous control policy to perform a desired task without performing a detailed analysis of the task itself or explicit coding. My work focuses on using statistical machine learning approaches to infer these controllers from noisy, suboptimal, and ambiguous human demonstrations.

In my dissertation, I developed an approach to extract an unknown number of overlapping finite-state-machine states from unsegmented demonstration. By assigning each of multiple 'correct' actions to different subtasks, the robot can then execute only one instead of averaging them, a potentially dangerous solution.

Currently, I am looking at learning from failed demonstrations by modeling the demonstrations, and then deliberately avoiding them during task reproduction. I am also investigating an alternate parameterization of trajectories (as single high-dimensional points) to take greater advantage of covariance information.

Much of my work deals with Gaussians, particularly Gaussian Mixture Models and Gaussian Processes. You can read my most recent work-related thoughts on my blog.

Recent Activities:

AAAI 2011 learning from demonstration challenge.
AAAI 2011 Robot Challenge - LfD, San Francisco, CA, July 2011.
bib | Video | Url ]
Getting robots to do what you want (even if you don't know what that is, exactly).
Willow Garage, Menlo Park, CA, June 2011.
bib | Slides ]
Imitation and reinforcement learning from failed demonstrations.
Daniel H Grollman and Aude G Billard.
In ICML workshop on New Developments in Imitation Learning, Bellevue, WA, June 2011.
bib | Text | Poster ]



Contact info:
Daniel Grollman
EPFL, Ecole Polytechnique Federale de Lausanne
EPFL-STI-I2S-LASA, Station 9
CH 1015, Lausanne, Switzerland


Tel: (+41) 21 693 5309
Fax: (+41) 21 693 78 50