My name is Brad Hayes.
I'm a Postdoctoral Associate in the MIT Interactive Robotics Group working on
Human-Robot Teaming
to improve the safety and capabilities of collaborative robots.
Research
My research contributes toward the creation of autonomous robots that can safely and productively learn from and work with humans. I am particularly interested in developing algorithms that enable and facilitate collaborative artificial intelligence, allowing robots to make human workers safer, more effective, and more efficient at their jobs. As such, my research spans many fields, including Learning from Demonstration, Hierarchical Learning, Multi-agent Planning, User Modeling, Cognitive Science, and Human-Robot Interaction.
 
Active Research Areas
Learning from Demonstration and Explanation
Motivation: How can we make collaborative robots that flexibly learn from observation and natural interactions, such that they increase capability with world experience?
My work involves robots learning and using motor skills in complex and interactive ways. Primitive skill acquisition should be driven by peer demonstration and human instruction, allowing collaborators to transparently impart skill knowledge to a robot without having to directly program behaviors. This involves learning from demonstration to acquire motor primitives, modeling teammate preferences, and learning task structure.
Interpretable Machine Learning: Shared Mental Models of Actions and Tasks
Motivation: How can a robot learn models of complex, multi-step tasks and communiate effectively with its human teammates?
Humans are very adept at discerning the goals and ordering constraints of a task solely through observation or exploring the task-space, but cannot interpret most machine learning models. My work builds and uses goal-directed hierarchical task networks that inform team planning, individual contingency planning, and behavioral explanation. My research in this area spans the development of novel task-space exploration functions, active learning strategies, and task representations to produce task and communication models for use in collaboration.
Planning and User Modeling for Human-Robot Collaboration
Motivation: What features should a robot be sensitive to when building models of its collaborators? How can a robot operationalize this information to improve team fluency and effectiveness?
My work in the planning, user modeling, and teamwork domain centers on constructing adaptive models of a one's collaborators during live operation. This involves modeling preferences and relevant personal attributes, integrating consideration for these features into a real-time, far horizon collaborative planner. By incorporating co-worker models into the planning systems of robots in team execution contexts, we can dramatically increase team fluency and effectiveness.
Highly Refereed Publications
2016
  • Matthew Gombolay, Jessie Yang, Bradley Hayes, Nicole Seo, Samir Wadhwania, Zixi Liu, Tania Yu, Neel Shah, Toni Golen, and Julie Shah. (2016). Robotic Assistance in Coordination of Patient Care. To appear: Proceedings of Robotics: Science and Systems (RSS 2016), Michigan, USA, June 18-22.
  • Bradley Hayes and Brian Scassellati. (2016). Autonomously Constructing Hierarchical Task Networks for Planning and Human-Robot Collaboration. In Proceedings of 2016 IEEE International Conference on Robotics and Automation (ICRA 2016), Stockholm, Sweden, May 16 - 21.
  • Henny Admoni, Thomas Weng, Bradley Hayes, and Brian Scassellati. (2016). Robot Nonverbal Behavior Improves Task Performance In Difficult Collaborations. In Proceedings of the 11th ACM/IEEE International Conference on Human-Robot Interaction (HRI 2016). Christchurch, New Zealand, March 7-10.
2015
  • Bradley Hayes and Brian Scassellati. (2015). Effective Robot Teammate Behaviors for Supporting Sequential Manipulation Tasks. In Proceedings of the 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2015). Hamburg, Germany, September 27 - October 3.
  • Benjamin Rosman, Bradley Hayes, and Brian Scassellati. (2015). Enhancing Agent Safety through Autonomous Environment Adaptation. In Proceedings of the 5th joint IEEE International Conference on Development and Learning and on Epigenetic Robotics (ICDL 2015). Providence, Rhode Island, USA, August 13-16.
  • Bradley Hayes. (2015). Social Hierarchical Learning (Extended Abstract). In Proceedings of the 20th AAAI/SIGAI Doctoral Consortium. Austin, Texas, USA, January 26-27.
2014
  • Bradley Hayes, Elena Corina Grigore, Alexandru Litoiu, Aditi Ramachandran, Brian Scassellati. (2014). A Developmentally Inspired Transfer Learning Approach for Predicting Skill Durations. In Proceedings of the 4th joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL 2014). Genoa, Italy, October 13-16.
  • Bradley Hayes and Brian Scassellati. (2014). Discovering Task Constraints Through Observation and Active Learning. In Proceedings of the 2014 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2014). Chicago, USA, September 14-18.
  • Bradley Hayes, Daniel Ullman, Emma Alexander, Caroline Bank, and Brian Scassellati. (2014). People Help Robots Who Help Others, Not Robots Who Help Themselves. In Proceedings of the 23rd IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN 2014). Edinburgh, Scotland, August 25-29. (Finalist, RSJ/KROS Distinguished Interdisciplinary Research Award)
  • Emma Alexander, Caroline Bank, Jie Jessica Yang, Bradley Hayes, and Brian Scassellati. (2014). Asking for Help from a Gendered Robot. In Proceedings of the 36th Annual Conference of the Cognitive Science Society (CogSci 2014). Quebec City, Canada, July 23-26.
2013
  • Bradley Hayes and Brian Scassellati. (2013). Improving Implicit Communication In Mixed Human-Robot Teams With Social Force Detection. In Proceedings of the 3rd joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL 2013). Osaka, Japan, August 18-22.
  • Henny Admoni, Bradley Hayes, David Feil-Seifer, Daniel Ullman, and Brian Scassellati. (2013). Dancing With Myself: The effect of majority group size on perceptions of majority and minority robot group members. In: Proceedings of the 35th Annual Conference of the Cognitive Science Society. Austin, TX: Cognitive Science Society (CogSci 2013). Berlin, Germany, July 31-August 3.
  • Henny Admoni, Bradley Hayes, David Feil-Seifer, Daniel Ullman, and Brian Scassellati. (2013). Are You Looking At Me? Perception of Robot Attention is Mediated by Gaze Type and Group Size. In Proceedings of the 8th ACM/IEEE International Conference on Human-Robot Interaction (HRI 2013). Tokyo, Japan, March 3-6.
Lightly Refereed Publications
  • Bradley Hayes, Matthew C Gombolay, Malte F Jung, Koen Hindriks, Joachim de Greeff, Catholijn Jonker, Mark Neerincx, Jeffrey M Bradshaw, Matthew Johnson, Ivana Kruijff-Korbayova, Maarten Sierhuis, Julie A Shah, Brian Scassellati. (2015). HRI Workshop on Human-Robot Teaming. In Proceedings of the Tenth Annual ACM/IEEE International Conference on Human-Robot Interaction Extended Abstracts. Portland, Oregon, March 2-5.
  • Brian Scassellati and Bradley Hayes. (2014). Human-robot collaboration. AI Matters 1, 2 (December 2014), 22-23.
  • Bradley Hayes and Brian Scassellati. (2014). Developing Effective Robot Teammates for Human-Robot Collaboration. In Proceedings of the "Artificial Intelligence and Human-Robot Interaction" (AI-HRI) Fall Symposium. Arlington, Virginia USA, November 13-15.
  • Bradley Hayes and Brian Scassellati. (2014). Online Development of Assistive Robot Behaviors for Collaborative Manipulation and Human-Robot Teamwork. In: Proceedings of the "Machine Learning for Interactive Systems" (MLIS) Workshop at AAAI 2014. Quebec City, Canada, July 27.
  • Bradley Hayes and Brian Scassellati. (2013). Challenges in Shared-Environment Human-Robot Collaboration. In: Proceedings of the "Collaborative Manipulation" Workshop at HRI 2013. Tokyo, Japan, March 3.
  • Bradley Hayes and Brian Scassellati. (2013). Social Hierarchical Learning. In: Proceedings of the "HRI Pioneers" Workshop at HRI 2013. Tokyo, Japan, March 3.
Posters
  • A Developmentally Inspired Transfer Learning Approach for Predicting Skill Durations. At the 4th joint IEEE International Conference on Development and Learning and Epigenetic Robotics. Genoa, Italy, October 13-16.
  • Social Hierarchical Learning: Enabling Human-Robot Teamwork. "Robotics Exhibition" at CogSci/AAAI 2014. Quebec City, Canada, July 26 - 30.
  • Online Development of Assistive Robot Behaviors for Collaborative Manipulation and Human-Robot Teamwork. Machine Learning for Interactive Systems Workshop at AAAI 2014. Quebec City, Canada, July 28.
  • Asking for Help from a Gendered Robot. CogSci 2014. Quebec City, Canada, July 23-26.
  • Challenges in Shared-Environment Human-Robot Collaboration. "Collaborative Manipulation" Workshop at HRI 2013. Tokyo, Japan, March 3.
  • Social Hierarchical Learning. "HRI Pioneers" Workshop at HRI 2013. Tokyo, Japan, March 3.
Workshop Committees
Research Talks
E-mail: <last name>bh at mit dot edu
Professional Experience
Software Engineer, BAE Systems/Alphatech (Burlington, MA)6/2008 - 8/2009
  • Performed computer vision and artificial intelligence algorithm analysis and development on DARPA projects
  • Worked on DARPA programs: VIRAT, NETTRACK, URGENT, and PANDA
Software Development Engineer Intern, Microsoft (Redmond, WA)6/2007 - 8/2007
  • Designed and implemented two scripting languages and wrote high performance generic malware detection algorithms.
Extreme Blue Technical Intern, IBM (Austin, TX)6/2006 - 8/2006
  • Architected and implemented project 'Sentinel': Policy driven information security compliance
  • Received extensive public speaking training, culminating in a project presentation to IBM's top executives
Co-op Preprofessional Programmer, IBM (Cambridge, MA)6/2005 - 8/2005
  • Authored "Adding accessibility to drag-and-drop web content", US Patent #7877700
Education
Yale University10/2015
Doctor of Philosophy, Computer Science
Yale University5/2012
Master of Philosophy, Computer Science
Master of Science, Computer Science
Boston College5/2008
Bachelor of Science, Computer Science
Bachelor of Arts, Mathematics
Skills
Android Arduino Baxter C++ CMake Create CSS3 Eclipse Espresso GIMP GIT HTML5 Java Javascript jQuery Keepon KUKA YouBot Matlab SQL Nao OpenCV OpenGL PHP Pixelmator Pleo Python ROS Scipy Subversion Visual Studio
Programming Languages Software/IDEs Robot Platforms
Personal
Blue Belt, Brazilian Jiu-Jitsu4/2013 - Present
  • 6/2013: 2nd Place - White Belt Adult Heavyweight CTBJJF Spring 2013 Championship
  • 8/2014: Promoted to Blue Belt
Captain, Yale Graduate Ice Hockey Team6/2010 - 9/2015