Robots Show Us How to Teach Them: Feedback from Robots Shapes Tutoring Behavior during Action Learning

Vollmer, Anna-Lisa; Mühlig, Manuel; Steil, Jochen J.; Pitsch, Karola; Fritsch, Jannik; Rohlfing, Katharina J.; Wrede, Britta

doi:10.1371/journal.pone.0091349

Cited by 40 publications

(26 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Within this paradigm, a robot can be trained in a similar fashion to Skinner's Operant Conditioning [6], using a Reinforcement Learning process with positive and negative feedback [7]. This, combined with studies in Human-Robot Interaction, lead to the concept of Transparency in Learning, introduced by Thomaz and Breazeal, for Social Robotics [8][9][10][11][12][13].…”

Section: State Of the Artmentioning

confidence: 99%

“…During the training process between humans and robots, if the latter shows a clear feedback of its inner state, it is said to have a Transparent Behavior [8]. This transparency in learning improves how people give instructions to the robot [9] allowing them to adapt their teaching methods by observing the agent's feedback [10][11][12].…”

Section: Scaffolding and Transparency In Learningmentioning

confidence: 99%

See 1 more Smart Citation

Design and Evaluation of a Spherical Robot with Emotion-Like Feedback during Human-Robot Training

Onchi

Lee

2020

Transactions of Japan Society of Kansei Engineering

View full text Add to dashboard Cite

While robotic assistants capable of automating many a task are becoming more ubiquitous, the complexity and unfamiliarity of such machines may discourage users from interacting and personalizing their robots. For that reason, the goal of this research is to study how a robot that expresses emotion-like feedback using movement affects the training interaction experience of non-expert users.To that end, a single-eyed spherical robot was developed using a Kansei design approach. A within-subjects experiment, between the robot expressing feedback and a control version that did not, was conducted with participants who had little to no experience interacting with robots. The results showed an increase in arousal when the robot expressed feedback and the overall interaction was considered engaging. Also, a positive correlation between bonding scores and relevance of the task was found. The contributions of this research are toward the development of human-trainable robots with transparent learning behavior.

show abstract

Section: State Of the Artmentioning

confidence: 99%

Section: Scaffolding and Transparency In Learningmentioning

confidence: 99%

Design and Evaluation of a Spherical Robot with Emotion-Like Feedback during Human-Robot Training

Onchi

Lee

2020

Transactions of Japan Society of Kansei Engineering

View full text Add to dashboard Cite

show abstract

“…Previous HRI studies have investigated the naive demonstration of continuous robot movement skills, focusing on the usability of kinesthetic teaching Weiss et al (2009), or not applying machine learning algorithms but studying the influence of designed robot behavior, for example incorporating findings from adult-infant interactions (Vollmer et al, 2009(Vollmer et al, , 2010(Vollmer et al, , 2014. Weiss et al (2009) have shown that naive users are able to teach a robot new skills via kinesthetic teaching.…”

Section: Human-robot Interaction With Machine Learnersmentioning

confidence: 99%

A User Study on Robot Skill Learning Without a Cost Function: Optimization of Dynamic Movement Primitives via Naive User Feedback

Vollmer

Hemion

2018

Front. Robot. AI

Self Cite

View full text Add to dashboard Cite

Enabling users to teach their robots new tasks at home is a major challenge for research in personal robotics. This work presents a user study in which participants were asked to teach the robot Pepper a game of skill. The robot was equipped with a state-of-the-art skill learning method, based on dynamic movement primitives (DMPs). The only feedback participants could give was a discrete rating after each of Pepper's movement executions ("very good," "good," "average," "not so good," "not good at all"). We compare the learning performance of the robot when applying user-provided feedback with a version of the learning where an objectively determined cost via hand-coded cost function and external tracking system is applied. Our findings suggest that (a) an intuitive graphical user interface for providing discrete feedback can be used for robot learning of complex movement skills when using DMP-based optimization, making the tedious definition of a cost function obsolete; and (b) un-experienced users with no knowledge about the learning algorithm naturally tend to apply a working rating strategy, leading to similar learning performance as when using the objectively determined cost. We discuss insights about difficulties when learning from user provided feedback, and make suggestions how learning continuous movement skills from non-expert humans could be improved.

show abstract

“…Similar trajectories have been used in another HRI study which investigated how a robot's imitative motions might affect a human teacher [52]. In this study, participants were asked to perform eight "everyday" motions described as follows: how an airplane flies to an airport, how a phone is hung up, how a dog walks to a bowl, how an elevator descends, how an airplane does a loop, how to clean a window with a sponge, how a frog jumps, and how a feather falls.…”

Section: Playful Reachingmentioning

confidence: 99%

Avoiding Playfulness Gone Wrong: Exploring Multi-objective Reaching Motion Generation in a Social Robot

Cooney

Sant'Anna

2017

Int J of Soc Robotics

View full text Add to dashboard Cite

Companion robots will be able to perform useful tasks in homes and public places, while also providing entertainment through playful interactions. "Playful" here means fun, happy, and humorous. A challenge is that generating playful motions requires a non-trivial understanding of how people attribute meaning and intentions. The literature suggests that playfulness can lead to some undesired impressions such as that a robot is obnoxious, untrustworthy, unsafe, moving in a meaningless fashion, or boring. To generate playfulness while avoiding such typical failures, we proposed a model for the scenario of a robot arm reaching for an object: some simplified movement patterns such as sinusoids are structured toward appearing helpful, clear about goals, safe, and combining a degree of structure and anomaly. We integrated our model into a mathematical framework (CHOMP) and built a new robot, Kakapo, to perform dynamically generated motions. The results of an exploratory user experiment were positive, suggesting that: Our proposed system was perceived as playful over the course of several minutes. Also a better impression resulted compared with an alternative playful system which did not use our proposed heuristics; furthermore a negative effect was observed for several minutes after showing the alternative motions, suggesting that failures are important to avoid. And, an inverted u-shaped correlation was observed between motion length and degree of perceived playfulness, suggesting that motions should neither be too short or too long and that length is also a factor which can be considered when generating playful motions. A short follow-up study provided some additional support for the idea that playful motions which seek to avoid failures can B Martin Cooney martin.daniel.cooney@gmail.com 1 School of Information Technology, Halmstad University, Halmstad, Halland, Sweden be perceived positively. Our intent is that these exploratory results will provide some insight for designing various playful robot motions, toward achieving some good interactions.

show abstract

Robots Show Us How to Teach Them: Feedback from Robots Shapes Tutoring Behavior during Action Learning

Cited by 40 publications

References 23 publications

Design and Evaluation of a Spherical Robot with Emotion-Like Feedback during Human-Robot Training

Design and Evaluation of a Spherical Robot with Emotion-Like Feedback during Human-Robot Training

A User Study on Robot Skill Learning Without a Cost Function: Optimization of Dynamic Movement Primitives via Naive User Feedback

Avoiding Playfulness Gone Wrong: Exploring Multi-objective Reaching Motion Generation in a Social Robot

Contact Info

Product

Resources

About