Alexander Broad

I work at Boston Dynamics where I explore how machine learning can be used to improve the capabilities of complex, highly dynamic robots. I am particularly interested in ideas that fall at the intersection of perception and control.

Bio

I have a PhD in Computer Science from Northwestern University where I studied how machine learning can be used to improve assistive robots in the context of home assistance and physical rehabilitation. At Northwestern, I worked with Professors Brenna Argall and Todd Murphey. During my graduate studies I interned at MERL where I researched video object detection with Dr. Michael Jones. Prior to starting my doctoral studies I was an Associate Technical Staff member at MIT Lincoln Laboratory where I worked in an applied machine learning group. I received a B.A. in Mathematics and P-N-P (Philosophy-Neuroscience-Psychology) from Washington University in St. Louis. I also graduated with a M.S. in Computer Science from the same university, where I worked with Professor Bill Smart and Dr. Tom Erez on biologically motivated learning algorithms for robotic control.

Experience

Machine Learning Researcher Boston Dynamics 2019-Present: Exploring how machine learning can be used to improve the capabilities of complex and highly dynamic robots, with a particular focus on perception.
Machine Learning Research Intern MERL 2017-2017: Worked with Dr. Michael Jones to explore how motion features from video data can be used to improve video object detection.
PhD Student Northwestern University 2014-2019: Worked with Professors Brenna Argall and Todd Murphey to explore how machine learning can be used to improve a human operator's control of complex, dynamic machines and restore lost functionality to people in need.
Associate Technical Staff MIT Lincoln Laboratory 2011-2014: Worked with a small team to apply machine learning techniques to large datasets with applications in natural language processing and computer vision.

Publications

Highly Parallelized Data-driven MPC for Minimal Intervention Shared Control Broad, A., Murhpey, T., Argall, B. Robotics: Science and Systems. 2019 Supplementary Material GitHub Video Bibtex
Generalizable Data-Driven Models for Personalized Shared Control of Human-Machine Systems Broad, A. PhD Thesis, Northwestern University. 2019. Bibtex
Operation and Imitation under Safety-Aware Shared Control Broad, A., Murhpey, T., Argall, B. Workshop on the Algorithmic Foundations of Robotics. 2018. Bibtex
Recurrent Multi-frame Single Shot Detector for Video Object Detection Broad, A., Lee, T-Y., Jones, M. British Machine Vision Conference. 2018. Supplementary Material Video Bibtex
Structured Neural Network Dynamics for Model-based Control Broad, A., Abraham, I., Murhpey, T., Argall, B. RSS Workshop on Learning and Inference in Robotics. 2018. Bibtex
Demonstration and Imitation of Novel Behaviors under Safety Aware Shared Control Broad, A., Murhpey, T., Argall, B. RSS Workshop on Causal Imitation in Robotics. 2018. Bibtex
Learning Models for Shared Control of Human-Machine Systems with Unknown Dynamics Broad, A., Murhpey, T., Argall, B. Robotics: Science and Systems. 2017 GitHub Bibtex

Real-Time Natural Language Corrections for Assistive Robotic Manipulators Broad, A., Arkin, J., Ratliff, N., Howard, T., Argall, B. International Journal of Robotics Research. 2017. Bibtex
Geometry-Based Region Proposals for Real-Time Robot Detection of Tabletop Objects Broad, A., Argall, B. arXiv CORR abs/1703.04665. 2017. ArXiv GitHub Bibtex
An Empirical Analysis of Methods for Learning Robot Kinematics from Demonstration Broad, A., Gopinath, D., Murphey, T., Argall, B. Midwest Robotics Workshop. 2017. Bibtex
Trust Adaptation Leads to Lower Control Effort in Shared Control of Crane Automation Broad, A., Schutlz, J., Derry, M., Murphey, T., Argall, B. Robotics and Automation Letters. 2016. Also presented at the Conference on Automation Science and Engineering. 2016. Bibtex
Towards Real-Time Natural Language Corrections for Assistive Robots Broad, A., Arkin, J., Ratliff, N., Howard, T., Argall, B. RSS Workshop on Model Learning for Human-Robot Communication. Oral. 2016. Bibtex
Geometry-Based Region Proposals for Accelerated Image-Based Detection of 3D Objects Broad, A., Schutlz, J., Derry, M., Murphey, T., Argall, B. RSS Workshop on Deep Learning. Oral. 2016. GitHub Bibtex
Inverted Trust Improves Shared Control of Complex Dynamic Systems Broad, A., Derry, M., Schutlz, J., Murphey, T., Argall, B. RSS Workshop on Social Trust in Autonomous Robots. Oral. 2016. Bibtex
Path Planning under Interface-Based Constraints for Assistive Robotics Broad, A., Argall, B. International Conference on Automated Planning and Scheduling. 2016. Bibtex
Probabilistic Models for Real-Time Natural Language Corrections to Assistive Robotic Manipulators Arkin, J., Broad, A., Howard, T., Argall, B. Midwest Robotics Workshop. 2016. Bibtex
Assistive Robotic Manipulation through Shared Autonomy and a Body-Machine Interface Jain, S., Farshchiansadegh, A., Broad, A., Abdollahi, F., Mussa-Ivaldi, F., Argall, B. International Conference on Rehabilitation Robotics. 2015. Bibtex
Feature-Rich Plagiarism Detection Using Structured Prediction Broad, A., King, D., Asarina, A. MIT Lincoln Laboratory Tech Report. 2014.
Generating Muscle Driven Arm Movements Using Reinforcement Learning Broad, A. Master’s Thesis, Washington University in St. Louis. 2011. Bibtex

Education

PhD in Computer Science Northwestern University 2014-2019: I studied machine learning for robotics, with a particular focus on assistive and rehabilitative technologies.
M.S. in Computer Science Washington University in St. Louis 2009-2011: I studied machine learning for robotics, with a particular focus on reinforcement learning and optimal control.
B.A. in Mathematics and Philosophy-Neuroscience-Psychology Washington University in St. Louis 2005-2009: I graduated with a double major in Mathematics and Philosophy-Neuroscience-Psychology (PNP).