AI & Its Alternatives in Assistive & Collaborative Robotics: Decoding Intent

RSS 2020 Workshop

July 14, 2020: Thank you all for taking time out of your day and attending our workshop! We hope you have gotten as much out of it as we did and had fun along the way. In case you weren't able to join us or want to re-watch some of the talks, check out the Live Recordings page. Also, please be on the lookout for a feedback form as we would love to hear what you enjoyed this year as well as to get your input on how we can improve for the next year. Happy RSS 2020!

Human-robot teaming (HRT) is an interdisciplinary domain, largely because it is a bidirectional challenge — the robot must infer human intent to provide assistance, and the human must understand the robot’s intent to provide commands or adjust behavior. For one, the autonomous agent is constrained by safety, calling for the application of formal methods and explainable AI. What’s more, the exchange of information is limited by partial observability and often underdefined — it can take place via passive observation, an established language of communication or through active interaction with a collaborating partner — lending itself to insights on how people learn to collaborate with each other. Finally, context plays an important role, making the problem high-dimensional with often ill-specified reward structures. Our workshop will facilitate a discussion between researchers that explore the human side of these interactions and those that propose new algorithmic solutions, spanning cognitive science, artificial intelligence, control theory, and more. We will aim to define and address challenges in human-robot teaming by asking questions like — what are effective approaches to inferring intent? How do we incorporate formal notions of safety with data-driven methods? Can we take learnings from autonomous (or assisted) driving and generalize them to other HRT scenarios? Can we design interpretable AI to better facilitate HRT? And finally, how do we use insights from psychology, neuroscience, and theory of mind to create more intuitive robotic interfaces? This workshop will foster multidisciplinary discussion and friendly debate as well as consolidate perspectives, methodologies, and assessment tools to grow research efforts in human-centered robotics.



Confirmed Speakers

Dorsa Sadigh

Stanford University

Guy Hoffman

Cornell University

Sangeet Khemlani

Naval Research Laboratory

Chris Baker

iSee AI

Brenna D. Argall

Northwestern University

Guy Rosman

Toyota Research Institute

Important Dates

Submission deadline (AoE time) June 20th

Notification of acceptance June 27th

Camera-ready deadline July 7th

Workshop July 13th

Call for Abstracts

We solicit extended abstracts for peer review. Abstracts should conform to RSS style guidelines and should be a maximum of 4 pages (excluding references). Submissions can include archived or previously accepted work (please make a note of this in the submission). Specifically, we invite participants to contribute with papers for two kinds of submissions:

1. Traditional workshop submissions presenting preliminary results of ongoing research.

2. Submissions linked to already published papers coming from:

a. Journal papers, which have not had the chance to be discussed in a conference.

b. Conference papers, which come from non-robotics conferences (e.g. cognitive science, machine learning). Although these papers already had exposure in a conference, the workshop is a great opportunity to give visibility to the these contributions within the robotics community.

Submission link: https://easychair.org/cfp/RSS2020_AI-ACR-Workshop

Topics of interest include, but are not limited to:

  • Shared autonomy / Human-in-the-loop systems

  • Cognitive science for Human-Robot Interaction (HRI)

  • Theory of mind for HRI

  • Data-driven models of the human and/or autonomous partner for HRI

  • Inferring intent with limited sensory input Shared autonomy / Human-in-the-loop systems

  • Interpretable AI for HRI

  • Formal notions of safety in HRI

  • Assistive robotics

  • Collaborative robotics

Reviewing will be single blind. All accepted contributions will be presented in an interactive poster session and short talks during the workshop.