The Speakers

Title: Using Insights about Human Cognition to Design Interactive Robots

Abstract: It is expected that general-purpose, autonomous machines will become pervasive in domestic, public, and industrial spaces within the next decade. They will assist humans in a variety of activities including doing household tasks and collaborating on the assembly line. Personal robots, along with other intelligent agents such as smart homes and cars, will add tremendously to the quality of human life. They will offer persons with impairments more independence, help older adults with their daily chores, transport people and goods, and perform search and rescue in environments that are too dangerous for humans. Several challenges have to be addressed to make progress towards this vision. Each home, office, or assembly line is organized differently. Users will want the agents to perform a variety of tasks and will have different preferences. Customizing every agent for its deployment environment and user preferences is resource-intensive and costly. One approach to this challenge is designing a generally intelligent robot that can adapt to the user requirement on its own instead of relying on dedicated programming. With my colleagues, I study how human feedback, guidance, and structure can be used to reduce the complexity of robot learning. Our goal is to develop agents whose behavior can be easily extended by human users using natural interactions. This research agenda is a fascinating and gratifying mix of insights from various scientific disciplines including cognitive science, human-computer interaction, machine learning, reasoning & inference, and robotics. In this talk, I will introduce the research agenda - Interactive Task Learning - and present the latest advances. I will specifically discuss our recent work on using analogical reasoning and generalization as a basis for interactive robot learning. Additionally, I will summarize insights from our recent human participant studies that studies the space of variability in human robot teaching.

Bio: Dr. Shiwali Mohan is a senior member of research staff at Xerox PARC. She studies the design and analysis of collaborative human-AI systems. Her research brings together methods from artificial intelligence (AI) and machine learning (ML) with insights from human-centered sciences to design systems that can collaborate with humans effectively. Her research has had an inter-disciplinary impact and has been published at venues for research on AI, human cognition, cognitive systems, human-computer interaction (HCI), medical informatics, AI for social good, and robotics. She is leading research on two DARPA programs - SAIL-ON and GAILA - which explore how complex, collaborative systems can be built by combining methods from statistical and cognitive streams of AI. Previously, she has been a key personnel on grants from ARPA-E, NSF, ONR among other government agencies.  Dr. Mohan received her B.E. in Instrumentation and Control Engineering from Netaji Subhas Institute of Technology, Delhi University. She received her M.S. and Ph.D. degrees in computer science from the University of Michigan, Ann Arbor with a focus on artificial intelligence. Her doctoral research on collaborative robots paved the way for a new challenge problem for AI systems research - Interactive Task Learning. With John Laird, she won the Blue Sky Award at AAAI 2018 for proposing a novel intelligent systems framework. She is actively involved in the AI and HCI scientific communities. Shiwali loves to dance and was a principal in a prominent Bay area dance company that has its roots in Indian forms. She is also known (in a relatively small circle) for her phenomenal butter chicken. In her recent trip to India, she was delighted to discover Dastakar (https://www.dastkar.org/) an organization that supports handmade and natural products developed by marginalized communities left out from the economic mainstream. She can be found at www.shiwali.me and @shiwalimohan.

Title: Adapting Robot Learning for Human-Centered Environments

Abstract: While there have been some notable steps forward in robot learning (and machine learning more generally), these results often deal with robots in human-free environments where safety, time, and coordination are not critical issues.  In this talk I will focus on three recent results from our group that highlight how to bring robot-based learning into human-friendly spaces: (1) a method for learning to use tools in novel ways that allows skill transfer from tool to tool and from robot to robot, (2) a method for integrating abstract symbolic rules ("don't touch the stove") with experience-driven statistical learning methods, and (3) an affordance-based addition to a task and motion planning pipeline that can rapidly shrink computation time in constrained environments.

Bio: Brian Scassellati is the A. Bartlett Giamatti Professor of Computer Science, Cognitive Science, and Mechanical Engineering at Yale University.  His research focuses on human-robot interaction, especially on how robots can help support social and cognitive skills in people.