论文阅读及文献综述 AI辅助网站

Web1: 总结文献要点

https://www.paper-digest.com/

举例：Robot education peers in a situated primary school study: Personalisation promotes child learning
Baxter, Ashurst, Read, Kennedy, Belpaeme PLOS ONE 2017
Full text link: https://doi.org/10.1371/journal.pone.0178126

分以下模块：

1.1. What this paper is about？

Introduction and Objective

    While the examination of learning outcomes for the children was not the focus of the study, with the development of relationships between the children and robot the primary aim, it was shown that children who maintained peer-like interactions with the robot maintained interactions over the extensive experimental period.
    A number of other studies have recently followed from these seminal works to further explore the specific potential role that such social robots can play in helping children to learn, although typically these have taken place outside of school classrooms or over isolated interaction sessions.
    However, these studies are ambiguous regarding the actual impact of social behaviour on child learning: the presence of robots appears to facilitate increased learning, but the role of social behaviour to extend this effect remains unclear, in contrast to the human-centred theory.

1
2
3
4
5
6

1.2. What you can learn？

Results

    The two classes used in this study were not divided on the basis of ability, although they were of the same age.
    The analysis shows that this is the case for the novel MPNP = 0.629, 95% CI = 0.557, 0.589 and the incidental MPNP = 0.682, 95% CI = 0.588, 0.589 learning tasks, indicating positive learning effects in these learning tasks. This is not observed in the familiar learning task MPNP = 0.059, 95% CI = 0.174, 0.175.106 replications on the difference between the conditions PNP, compared to observed difference.
    Conversely, however, the response in interaction one is strongly and positively correlated with the interest/enjoyment post-experiment questionnaire response in the P condition r26 = 0.743, p<.001, but not in the NP condition r24 = 0.032, p =.877. This appears to suggest that the levels of enjoyment experienced in the first interaction are maintained throughout the experiment in the P condition, but not necessarily in the NP condition.

Discussion and Conclusions

    It is thus possible that the difference in robot personalisation between conditions leads to a difference in perception of friendliness, which in turn could have an effect on the learning outcomes.
    However, the outcome of the post-study questionnaires indicates that that this is not the case.
    Secondly, it is possible that the novelty factor of having robots in the classroom increased overall motivation and hence performance in the tasks. This is unlikely for two reasons: a given the same hardware setup in both classrooms, there is nevertheless an increased performance in the novel task for the personalised condition but not the non-personalised condition, indicating the influence of condition differences over a novelty factor; and b the qualitative results indicate that the novelty factor decreased in the second week also see point below.
    For example, the teachers were not informed of the specific hypotheses nor conditions of the study, and were involved only in the learning task content and procedural issues to ensure that similar methods would be used by the teachers when interacting with and referring to the robot in their classroom.
    However, they noted that in both classrooms, this distracting effect dissipated in the second week, although they reported still being able to use the robots as a motivator for the children. This is supported by the high levels of interest/enjoyment in the activity at the end of the study non-significantly higher for the personalised condition. This maintenance of motivation speaks to the wider role of technology, including social robotics, in the classroom and how it is handled orchestrated by the teachers.
    While acceptance was high in the present study, this may be a self-selection bias i.e. the school and teachers were enthusiastic about the study prior to implementation, and further examination of the effort required on the part of the teachers and the school versus the learning benefits afforded by the type of personalised social robot systems we have demonstrated here is necessary, particularly in embedded applications i.e. inside the classroom itself, as we have achieved in the present study.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

Web2: 文献综述

https://www.paperdigest.org/

2.1. Related Work

2.2. Summary of the Related Work