Augmentative and Alternative Communication
Speech generating alternative and augmentative communication (AAC) devices may be used to express thoughts, needs, wants, and ideas when someone cannot rely on their own speech to communicate. Some examples of these devices include specialized keyboards along with adapted controllers and text-to-speech interfaces. We study conversational agency in AAC, how people who use AAC devices augmented communicators, advance their goals in conversation under social and AAC constraints. We study how people use AAC devices with different types of conversation partners to inform new designs that can reduce user burden and favor the expression of conversational agency of augmented communicators.
For more information, please contact Stephanie.
Check out our CHI 2020 video presentation to learn more.
Co-designing Socially Assistive Sidekicks for Motion-based AAC.
Stephanie Valencia, Michal Luria, Amy Pavel, Jeffrey P. Bigham, Henny Admoni. Proceedings of the ACM/IEEE International Conference on Human-Robot Interaction (HRI). 2021. pdf
Eye-gaze for Intelligent Driving Assistance
Using Programmable Light Curtains - an active sensor, we can obtain 3D information about the world in a more precise, dense, and frequent manner than with LiDAR devices. We study the use of driver eye gaze to inform policies of placing and shaping the light curtain for autonomous and assisted driving.
Evaluation tools for Social Robot Navigation
We are actively maintaining SocNavBench, a social navigation benchmark for evaluating social navigation algorithms against each other in a realistic, consistent, and scalable way. Checkout the github page!
SocNavBench: A Grounded Simulation Testing Framework for Social Navigation.
Abhijat Biswas, Allan Wang, Gustavo Silvera, Aaron Steinfeld, Henny Admoni. ACM Transactions on Human-Robot Interaction (THRI). 2021. pdf
Assistive Manipulation Through Intent Recognition
An upper body mobility limitation can severely impact a person's quality of life. Such limitations can prevent people from performing everyday tasks such as picking up a cup or opening a door. The U.S. Census Bureau has indicated that more than 8.2% of the U.S. population, or 19.9 million Americans, suffer from upper body limitations. Assistive robots offer a way for people with severe mobility impairment to complete daily tasks. However, current assistive robots primarily operate through teleoperation, which requires significant cognitive and physical effort from the user. We explore how these assistive robots can be improved with artificial intelligence to take an active role in helping their users. Drawing from our understanding of human verbal and nonverbal behaviors (like speech and eye gaze) during robot teleoperation, we study how intelligent robots can predict human intent during a task and assist toward task completion. We aim to develop technology to decrease operator fatigue and task duration when using assistive robots by employing human-sensitive shared autonomy.
HARMONIC: A Multimodal Dataset of Assistive Human–Robot Collaboration.
Benjamin A. Newman *, Reuben M. Aronson *, Siddhartha S. Srinivasa, Kris Kitani, Henny Admoni. The International Journal of Robotics Research (IJRR). 2021. pdf
Inferring Goals with Gaze during Teleoperated Manipulation.
Reuben M. Aronson, Nadia AlMutlak, and Henny Admoni. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). 2021. pdf
Eye Gaze for Assistive Manipulation.
Reuben M. Aronson, Henny Admoni. HRI Pioneers workshop. 2020. pdf
Semantic Gaze Labeling for Human-Robot Shared Manipulation.
Reuben M. Aronson and Henny Admoni. Proceedings of the ACM Symposium on Eye Tracking Research and Applications (ETRA). 2019. pdf
Gaze for Error Detection During Human-Robot Shared Manipulation.
Reuben M. Aronson and Henny Admoni. Towards a Framework for Joint Action Workshop at RSS. 2018. pdf
Eye-Hand Behavior in Human-Robot Shared Manipulation.
Reuben M. Aronson, Thiago Santini, Thomas C. Kübler, Enkelejda Kasneci, Siddhartha Srinivasa, and Henny Admoni. Proceedings of the ACM/IEEE International Conference on Human-Robot Interaction (HRI). 2018. pdf
Robot Self-Assesment (MURI)
Autonomous agents need to learn increasingly competent and complex behaviors. One way of effectively learning these behaviors is to include people in the learning process. Therefore, we are investigating human-in-the-loop strategies that are both more user friendly and lead to efficient learning. Please direct any questions to Pallavi Koppol.
Even after autonomous agents have learned complex behaviors, we likely won’t rely on them until we can reliably predict their behavior in new situations. Thus we are researching how agents can teach the nuances of their learned behaviors to humans through well-selected demonstrations. We first leverage inverse reinforcement learning and human learning strategies (e.g. scaffolding) to select demonstrations that are both informative and easily understood by humans. We then ask humans to predict agent behavior in unseen environments to test their understanding and inform the next demonstrations to be shown in a closed-loop teaching process. Please contact Michael Lee for more information.
INQUIRE: INteractive Querying for User-aware Informative REasoning.
Tesca Fitzgerald, Pallavi Koppol, Patrick Callaghan, Russell Quinlan Jun Hei Wong, Reid Simmons, Oliver Kroemer, Henny Admoni.. Conference on Robot Learning (CoRL). 2022.
Reasoning about Counterfactuals to Improve Human Inverse Reinfocement Learning.
Michael S. Lee, Henny Admoni, Reid Simmons. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). 2022. pdf
Metrics for Robot Proficiency Self-Assessment and Communication of Proficiency in Human-Robot Teams.
Adam Norton, Henny Admoni, Jacob Crandall, Tesca Fitzgerald, Alvika Gautam, Michael Goodrich, Amy Saretsky, Matthias Scheutz, Reid Simmons, Aaron Steinfeld, Holly Yanco. ACM Transactions on Human-Robot Interaction (THRI), 11(3). 2022. pdf
Machine Teaching for Human Inverse Reinforcement Learning.
Michael S. Lee, Henny Admoni, Reid Simmons. Frontiers in Robotics and AI. 2021. pdf
Interaction Considerations in Learning from Humans.
Pallavi Koppol, Henny Admoni, Reid Simmons. Proceedings of the 30th International Joint Conference on Artificial Intelligence (IJCAI). 2021. pdf
Understanding the Relationship between Interactions and Outcomes in Human-in-the-Loop Machine Learning.
Yuchen Cui *, Pallavi Koppol *, Henny Admoni, Scott Niekum, Reid Simmons, Aaron Steinfeld, Tesca Fitzgerald. Proceedings of the 30th International Joint Conference on Artificial Intelligence (IJCAI). 2021. pdf
Iterative Interactive Reward Learning.
Pallavi Koppol, Henny Admoni, Reid Simmons. Participatory Approaches to Machine Learning Workshop at ICML. 2020. pdf
Robots often need to communicate their goals to humans when navigating in a shared space to assist observers in anticipating the robot’s future actions. These human observers are often scattered throughout the environment, and each observer only has a partial view of the robot and its movements. A path that non-verbally communicates with multiple observers will need to be sufficiently understood by all of them. We aim to create an algorithm for intent-expressive, legible motion that takes into account the perspectives of multiple observers with limited fields of view in order to balance communicating with multiple observers effectively. Prior work in legible motion does not account for the limited field of view of observers, which can lead to wasted communication efforts that are unobserved by the intended audience, which we have improved upon with observer-aware legibility. Our user studies have shown that audience-aware legibility will require accounting for even more nuanced tradeoffs to enable better performance across multiple observers with different constraints.
Questions can be directed to Ada.
Observer-Aware Legibility for Social Navigation.
Ada V. Taylor, Ellie Mamantov, Henny Admoni. Proceedings of IEEE International Conference on Robot & Human Interactive Communication (RO-MAN). 2022. pdf
Wait Wait, Nonverbally Tell Me: Legibility for Use in Restaurant Navigation.
Ada V. Taylor, Ellie Mamantov, and Henny Admoni. Workshop on Social Robot Navigation at RSS 2021. 2021. pdf
Now You See It: The Effect of Multiple Audience Perspectives on Path Legibility.
Ada V. Taylor, Henny Admoni. Workshop on AIxFood at IJCAI-PRICAI 2020. 2021.
Mutual Adaptation in Human-Robot Collaboration
Effective team collaboration involves many factors, including understanding capabilities, coordination, and communication. In the context of collaborative tasks, people have different goals, as well as different preferred strategies for accomplishing them. While robots adapt to human partners, humans are simultaneously adapting to the robot. Mutual adaptation occurs when both partners can infer each other’s preferences and adapt their own behavior as necessary. How can robots reason about how its actions will affect a human partner? How can actions be strategically selected to elicit specific behavior from a human partner? We aim to provide insight on team coordination and extend existing frameworks for human-robot collaboration by exploring the effects of communication, collaboration, and mutual coordination on team fluency and performance.
Questions can be directed to Michelle.
Coordination with Humans via Strategy Matching.
Michelle Zhao, Reid Simmons, Henny Admoni. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). 2022. pdf
The Role of Adaptation in Collective Human-AI Teaming.
Michelle Zhao, Fade Eadeh, Thuy-Ngoc Nguyen, Pranav Gupta, Henny Admoni, Cleotide Gonzalez, Anita Williams Woolley. Topics in Cognitive Science. 2022. pdf
Teaching agents to understand teamwork: Evaluating and predicting collective intelligence as a latent variable via Hidden Markov Models.
Michelle Zhao, Fade Eadeh, Thuy-Ngoc Nguyen, Pranav Gupta, Henny Admoni, Cleotide Gonzalez, Anita Williams Woolley. Computers in Human Behavior. 2022. pdf
Adapting Language Complexity for AI-Based Assistance.
Michelle Zhao, Reid Simmons, Henny Admoni. Workshop on Lifelong Learning and Personalization in Long-Term Human-Robot Interaction at HRI 2021. 2021. pdf
Learning from Various Forms of Human-Provided Feedback
Humans can convey useful information to one another through a host of seemingly-disparate modalities. For example, if I want to teach someone to set a breakfast table, I could demonstrate how to set the table, I could show the person two table settings and indicate which is my preferred choice (ie. indicate a preference), I could critique a person’s attempted table setting by telling them it was “good” or “bad,” or I could go a step further and correct a person’s attempt to show them where they erred. Each of these interactions is different in form, yet they each convey similar information that is useful to accomplish my teaching goal. When robots join us in daily living, we’ll expect them to interpret and learn from our various forms of communication as we do—a feat that is so effortless for us, but a feat which is still so challenging for them.
This line of research explores learning from different interaction types in the service of two primary questions: How might we ease the challenge of interpretation for the robot? And, how might we ease the teaching burden placed upon the human?