A Principled Intelligent Occupational Training of Psychomotor Skills in Virtual Reality

Edward Kim, Zachary Pardos, Sanjit A. Seshia and Björn Hartmann

EECS Department
University of California, Berkeley
Technical Report No. UCB/EECS-2023-17
February 7, 2023

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2023/EECS-2023-17.pdf

The use of virtual reality (VR) for occupational training of psychomotor skills has been investigated for decades. Previous literature show that training in VR increases engagement, expedites learning, enhance safety, and improves skills in reality. Grounding on these progress, industry has begun to adopt VR to train their workers. However, we observe a disconnect between learning sciences and the current practice in VR training. Numerous literature across domains rely on self-assessment of the learners to predict whether they have reached a sufficient level, or mastery, of skill. However, it is well-established that self-assessment is inaccurate. Yet, there is no alternative for self-assessment to predict mastery of psychomotor skills. We propose to use bayesian knowledge tracing (BKT), a de facto standard in education to predict students' mastery of cognitive skills, for psychomotor skill mastery prediction. Using BKT, we design an intelligent occupational training system in VR. We conduct a between subjects study with 18 participants, with the control group relying on self-assessment to adjust curriculum and progression speed, while BKT is used instead for the experimental condition. Our results demonstrate the negative impact of self-assessment on psychomotor skill learning in VR, and shows the benefits of BKT as an alternative to self-assessment.


BibTeX citation:

@techreport{Kim:EECS-2023-17,
    Author = {Kim, Edward and Pardos, Zachary and Seshia, Sanjit A. and Hartmann, Björn},
    Title = {A Principled Intelligent Occupational Training of Psychomotor Skills in Virtual Reality},
    Institution = {EECS Department, University of California, Berkeley},
    Year = {2023},
    Month = {Feb},
    URL = {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2023/EECS-2023-17.html},
    Number = {UCB/EECS-2023-17},
    Abstract = {The use of virtual reality (VR) for occupational training of psychomotor skills has been investigated for decades. Previous literature show that training in VR increases engagement, expedites learning, enhance safety, and improves skills in reality. Grounding on these progress, industry has begun to adopt VR to train their workers. However, we observe a disconnect between learning sciences and the current practice in VR training. Numerous literature across domains rely on self-assessment of the learners to predict whether they have reached a sufficient level, or mastery, of skill. However, it is well-established that self-assessment is inaccurate. Yet, there is no alternative for self-assessment to predict mastery of psychomotor skills. We propose to use bayesian knowledge tracing (BKT), a de facto standard in education to predict students' mastery of cognitive skills, for psychomotor skill mastery prediction. Using BKT, we design an intelligent occupational training system in VR. We conduct a between subjects study with 18 participants, with the control group relying on self-assessment to adjust curriculum and progression speed, while BKT is used instead for the experimental condition. Our results demonstrate the negative impact of self-assessment on psychomotor skill learning in VR, and shows the benefits of BKT as an alternative to self-assessment.}
}

EndNote citation:

%0 Report
%A Kim, Edward
%A Pardos, Zachary
%A Seshia, Sanjit A.
%A Hartmann, Björn
%T A Principled Intelligent Occupational Training of Psychomotor Skills in Virtual Reality
%I EECS Department, University of California, Berkeley
%D 2023
%8 February 7
%@ UCB/EECS-2023-17
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2023/EECS-2023-17.html
%F Kim:EECS-2023-17