Limits...
Why Robots Should Be Social: Enhancing Machine Learning through Social Human-Robot Interaction.

de Greeff J, Belpaeme T - PLoS ONE (2015)

Bottom Line: We observe that conveying a learning preference through the use of social cues results in better and faster learning by the robot.In addition, the social learning shows a clear gender effect with female participants being responsive to the robot's bids, while male teachers appear to be less receptive.This work shows how additional social cues in social machine learning can result in people offering better quality learning input to artificial systems, resulting in improved learning performance.

View Article: PubMed Central - PubMed

Affiliation: Centre for Robotics and Neural Systems, Plymouth University, Plymouth, United Kingdom; Interactive Intelligence Group, Delft University of Technology, Delft, the Netherlands.

ABSTRACT
Social learning is a powerful method for cultural propagation of knowledge and skills relying on a complex interplay of learning strategies, social ecology and the human propensity for both learning and tutoring. Social learning has the potential to be an equally potent learning strategy for artificial systems and robots in specific. However, given the complexity and unstructured nature of social learning, implementing social machine learning proves to be a challenging problem. We study one particular aspect of social machine learning: that of offering social cues during the learning interaction. Specifically, we study whether people are sensitive to social cues offered by a learning robot, in a similar way to children's social bids for tutoring. We use a child-like social robot and a task in which the robot has to learn the meaning of words. For this a simple turn-based interaction is used, based on language games. Two conditions are tested: one in which the robot uses social means to invite a human teacher to provide information based on what the robot requires to fill gaps in its knowledge (i.e. expression of a learning preference); the other in which the robot does not provide social cues to communicate a learning preference. We observe that conveying a learning preference through the use of social cues results in better and faster learning by the robot. People also seem to form a "mental model" of the robot, tailoring the tutoring to the robot's performance as opposed to using simply random teaching. In addition, the social learning shows a clear gender effect with female participants being responsive to the robot's bids, while male teachers appear to be less receptive. This work shows how additional social cues in social machine learning can result in people offering better quality learning input to artificial systems, resulting in improved learning performance.

No MeSH data available.


Distribution of participants’ category choices in social and non-social condition, compared to the dataset distribution.Error bars indicate 95% confidence interval.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4589374&req=5

pone.0138061.g005: Distribution of participants’ category choices in social and non-social condition, compared to the dataset distribution.Error bars indicate 95% confidence interval.

Mentions: As illustrated in Fig 5, it is clear that participants diverge from the database distribution, both in the social and the non-social condition. In other words, they do not choose tutoring examples randomly, but follow a certain strategy (which may or may not be conscious). People tailor the learning input for the robot, even if it is not actively soliciting this, as is the case in the non-social condition. However, their diverging from the dataset distribution is more pronounced in the social condition; the difference in category use between social and non-social condition is significant for the fish, insect and mammal categories (two-sided t test with t(36) = 2.5385, p = 0.0156, t(36) = 2.3233, p = 0.0259 and t(36) = −2.1935, p = 0.0348 respectively), see Table 2. This illustrates that the additional social cues employed by the robot in the social condition can influence participants’ choice of topic.


Why Robots Should Be Social: Enhancing Machine Learning through Social Human-Robot Interaction.

de Greeff J, Belpaeme T - PLoS ONE (2015)

Distribution of participants’ category choices in social and non-social condition, compared to the dataset distribution.Error bars indicate 95% confidence interval.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4589374&req=5

pone.0138061.g005: Distribution of participants’ category choices in social and non-social condition, compared to the dataset distribution.Error bars indicate 95% confidence interval.
Mentions: As illustrated in Fig 5, it is clear that participants diverge from the database distribution, both in the social and the non-social condition. In other words, they do not choose tutoring examples randomly, but follow a certain strategy (which may or may not be conscious). People tailor the learning input for the robot, even if it is not actively soliciting this, as is the case in the non-social condition. However, their diverging from the dataset distribution is more pronounced in the social condition; the difference in category use between social and non-social condition is significant for the fish, insect and mammal categories (two-sided t test with t(36) = 2.5385, p = 0.0156, t(36) = 2.3233, p = 0.0259 and t(36) = −2.1935, p = 0.0348 respectively), see Table 2. This illustrates that the additional social cues employed by the robot in the social condition can influence participants’ choice of topic.

Bottom Line: We observe that conveying a learning preference through the use of social cues results in better and faster learning by the robot.In addition, the social learning shows a clear gender effect with female participants being responsive to the robot's bids, while male teachers appear to be less receptive.This work shows how additional social cues in social machine learning can result in people offering better quality learning input to artificial systems, resulting in improved learning performance.

View Article: PubMed Central - PubMed

Affiliation: Centre for Robotics and Neural Systems, Plymouth University, Plymouth, United Kingdom; Interactive Intelligence Group, Delft University of Technology, Delft, the Netherlands.

ABSTRACT
Social learning is a powerful method for cultural propagation of knowledge and skills relying on a complex interplay of learning strategies, social ecology and the human propensity for both learning and tutoring. Social learning has the potential to be an equally potent learning strategy for artificial systems and robots in specific. However, given the complexity and unstructured nature of social learning, implementing social machine learning proves to be a challenging problem. We study one particular aspect of social machine learning: that of offering social cues during the learning interaction. Specifically, we study whether people are sensitive to social cues offered by a learning robot, in a similar way to children's social bids for tutoring. We use a child-like social robot and a task in which the robot has to learn the meaning of words. For this a simple turn-based interaction is used, based on language games. Two conditions are tested: one in which the robot uses social means to invite a human teacher to provide information based on what the robot requires to fill gaps in its knowledge (i.e. expression of a learning preference); the other in which the robot does not provide social cues to communicate a learning preference. We observe that conveying a learning preference through the use of social cues results in better and faster learning by the robot. People also seem to form a "mental model" of the robot, tailoring the tutoring to the robot's performance as opposed to using simply random teaching. In addition, the social learning shows a clear gender effect with female participants being responsive to the robot's bids, while male teachers appear to be less receptive. This work shows how additional social cues in social machine learning can result in people offering better quality learning input to artificial systems, resulting in improved learning performance.

No MeSH data available.