Reward shaping for reinforcement learning by emotion expressions

Hwang, Kao-Shing; Ling, Jiajing; Chen, Yuying; Wang, Wei-Han

doi:10.1109/smc.2014.6974092

Cited by 3 publications

(3 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Mirnig et al [98] carried a video analysis in a large corpus of humans interacting with failing robots, and characterized human reactions, including reaction times and types of behavior displayed (verbal utterances and body motion, mostly). In sum, human reactions to robots failing are complex and include a great variety of behaviors, such as verbalizations [71], body motion [48,71,139], gaze [5,71] and facial expressions [58,71,130]. We expand more on robots can use human social signals upon robot failures in Section 4 (Theme 2).…”

Section: How Do People Behave When They Detect Another Person's Mistake?mentioning

confidence: 99%

“…They found that most participants display nonverbal behavior that is congruent with the robot actions -smiling and nodding when the robot performed well or corrected an action, and head shakes, frowning, averted gaze or closing of eyes for a long period of time in response to a robot error. Hwang et al [58] made use of the facial expressions of human observers, captured by a webcam, as input to a reinforcement learning system. In both tasks with a clear right or wrong solution and more open-ended tasks, robots used the observer's facial expression as feedback input to learn the tasks.…”

Section: How Can Robots Harness Nonverbal Social Feedback From Humans?mentioning

confidence: 99%

“…Humans are responsive to robots' social cues [17]; robots, in turn, should also be equipped with strategies that allow them to leverage users' social cues in order to achieve better task performance. Interesting work has used these types of inputs in reinforcement learning systems [58]. Future work should further explore the potential of human nonverbal inputs to achieve continuous robot learning and improvements in task completion.…”

Section: How Can Robots Harness Nonverbal Social Feedback From Humans?mentioning

confidence: 99%

See 2 more Smart Citations

Using Social Cues to Recognize Task Failures for HRI: A Review of Current Research and Future Directions

Bremers¹,

Pabst²,

Parreira³

et al. 2023

Preprint

View full text Add to dashboard Cite

Robots that carry out tasks and interact in complex environments will inevitably commit errors. Error detection is thus an important ability for robots to master, to work in an efficient and productive way. People leverage social cues from others around them to recognize and repair their own mistakes. With advances in computing and AI, it is increasingly possible for robots to achieve a similar error detection capability. In this work, we review current literature around the topic of how social cues can be used to recognize task failures for human-robot interaction (HRI). This literature review unites insights from behavioral science, human-robot interaction, and machine learning, to focus on three areas: 1) social cues for error detection (from behavioral science), 2) recognizing task failures in robots (from HRI), and 3) approaches for autonomous detection of HRI task failures based on social cues (from machine learning). We propose a taxonomy of error detection based on self-awareness and social feedback. Finally, we leave recommendations for HRI researchers and practitioners interested in developing robots that detect (physical) task errors using social cues from bystanders.

show abstract