Brief background: I've been following developments in intelligent tutoring systems for a while, and find it interesting to see how researchers are combining artificial intelligence, learning theory, affective computing, and sensor networks to create applications that might prove to be useful and effective.
The advantage of using intelligent tutoring applications in some cases is that it provides students with additional support and feedback the moment it is needed, something that is difficult for teachers to provide to students in large classrooms. With the increase in use of smartphones and other mobile devices such as the iPad, there is a good chance that this sort of technology will be used to support learning anywhere, anytime.
Although most intelligent tutoring systems are geared for 1-1 computing, I think there are some components that could be tweaked and then transfered to create intelligent "tutoring" systems for collaborative learning. Students like game-based learning, and what could be more fun than playing AND learning with a partner or group of peers? (I plan to revisit the research in this area in an upcoming post.)
The advantage of using intelligent tutoring applications in some cases is that it provides students with additional support and feedback the moment it is needed, something that is difficult for teachers to provide to students in large classrooms. With the increase in use of smartphones and other mobile devices such as the iPad, there is a good chance that this sort of technology will be used to support learning anywhere, anytime.
Although most intelligent tutoring systems are geared for 1-1 computing, I think there are some components that could be tweaked and then transfered to create intelligent "tutoring" systems for collaborative learning. Students like game-based learning, and what could be more fun than playing AND learning with a partner or group of peers? (I plan to revisit the research in this area in an upcoming post.)
Some thoughts:
I envision a system could support learning as well as important skills useful to students in life beyond the school walls, such as positive social interaction, teamwork, and problem-solving skills. The path of least resistance? Most likely applications that support the learning of pairs or small groups of students working at one display. However, in this era of the "21st Century Learner", there is a growing need for applications that can support small groups of students for collaborative groups and project-based learning activities.
There are a few applications developed for collaborative learning activities around a multi-touch table, such as the SMARTTable or the Surface, and more are needed. Also needed are intelligent systems that can support video conferencing and collaborative learning between students who are not physically co-located.
There are some problems that have yet to be solved. For example, the use of multiple sensors for an application designed for young people might be too intrusive. There are serious issues related to privacy/security. Who would have access to data regarding a student's emotional or physiological state? How would this data be utilized? How would this information be protected? Many school districts have security vulnerabilities, so it is possible that this information could be misused, if in the wrong hands.
ClassmateAssist is an application developed by Intel's Everyday Sensing and Perception team. Here is the description of the application from Intel Research:
In the following video, Richard Beckwith, a developmental psychologist at Intel, demonstrates a prototype of an application that uses video-sensing to track student's hand movements during a coin sorting lesson. The application provides feedback to the student, and also tracks data about the student's progress that can be transformed into a report for the teacher. The system can also monitor student's facial expression, note attention levels, and provide feedback.
Intel Research: Everyday Sensing and Perception
Socially and Physically Aware Interaction Systems (SPAIS)
Intel Research Press kit
SPAIS Publications:
Theocharous, G., Beckwith, R., Butko, N., Philipose, M. Tractable POMDP Planning Algorithms for Optimal Teaching in "SPAIS". International Joint Conferene on Artificial Intelligence (IJCAI) workshop on Plan Activity, and Intent Recognition (PAIR), Pasadena, California, July 2009.
Theocharous, G., Butko, N., Philipose, M. Designing a Mathematical Manipulatives Tutoring System using POMDPS. (pdf). POMDP Practitioners Workshop: Solving Real-world POMDP Problems, International Conference on Automated Planning and Scheduling (ICAPS). Toronto, May 2010
Wayang Outpost: Web-based Interactive Math/Intelligent Tutoring System, with Sensors.
I've followed the work of Beverly P. Woolf and her colleagues for some time. Much of their research has centered around a web-based application, Wayang Outpost, an intelligent electronic tutoring system that incorporates multimedia and animated adventures while providing activities designed to prepare teens for standardized math tests, such as the SAT and state-mandated end-of-course exams.
In recent years, the team has been using non-invasive sensors in their research, including a camera that views facial expressions, a posture-sensing device located in the seat of the student's chair, and a pressure-sensitive mouse, and a wireless skin conductance wristband. Data collected through all of these sensors can provide useful information about student learning. The system can also note when students try to "game" the system.
Related Publications
Woolf, B.P., Arroyo, I., Muldner, K., Burleson, W., Cooper, D., Dolan, R., Christopherson, R.M (2010)The Effect of Motivational Learning Companions on Low Achieving Students and Students with Disabilties (pdf) International Conference on Intelligent Tutoring Systems, Pittsburgh.
Abstract "We report the results of a randomized controlled evaluation of the effectiveness of pedagogical agents as providers of affective feedback. These digital learning companion were embedded in an intelligent tutoring system for mathematics, and were used by approximately one hundred students in two public high schools. Students in the control group did not receive the learning companions. Results indicate that low-achieving students—one third of whom have learning disabilities—had higher affective needs than their higher achieving peers; they initially considered math problem-solving more frustrating, less exciting, and felt more anxious when solving math problems. However, after they interacted with affective pedagogical agents, low-achieving students improved their affective outcomes, e.g., reported reduced frustration and anxiety."
Arroyo, I., Cooper, D.G., Burleson, W., Woolf, B.P., Muldner, K., Christopherson, R. (2009)
Emotion Sensors Go To School. AIED 2009. Pp. 17-24. IOS Press.
Low-tech description of Wayang Outpost, the math application used in the above publication:
Paul Franz, Recoder.Com 5/16/09
Cooper, D.G., Arroyo, I., Woolf, B.P., Muldner, K., Burleson, W., Christoperson, R. Sensors Model Student Self-Concept in the Classroom (pdf) UMass Amherst, June 22, 2009/UMAP 2009
There are a few applications developed for collaborative learning activities around a multi-touch table, such as the SMARTTable or the Surface, and more are needed. Also needed are intelligent systems that can support video conferencing and collaborative learning between students who are not physically co-located.
There are some problems that have yet to be solved. For example, the use of multiple sensors for an application designed for young people might be too intrusive. There are serious issues related to privacy/security. Who would have access to data regarding a student's emotional or physiological state? How would this data be utilized? How would this information be protected? Many school districts have security vulnerabilities, so it is possible that this information could be misused, if in the wrong hands.
Below I've highlighted two "intelligent" tutoring systems that incorporate the use of sensors in one form or another to generate information about student learning in a way that simulates what good teachers do every day. The ClassroomAssist application was developed by researchers at Intel, in collaboration with several universities. The Wayang Outpost application was developed by researchers at UMASS, and is aligned with the principles of Universal Design for Learning.
CLASSMATE ASSIST
ClassmateAssist is an application developed by Intel's Everyday Sensing and Perception team. Here is the description of the application from Intel Research:
"The advent of 1:1 computing in the classroom opens the door for teachers to set up individualized learning for their students who have a wide spectrum of interests and skills. ClasmateAssist technology uses computer vision and image projection to assist and guide students in a 1:1 learning environment, helping them to independently accomplish tasks at their own pace, while at the same time allowing teachers to be apprised of student progress."
Intel Research: Everyday Sensing and Perception
Socially and Physically Aware Interaction Systems (SPAIS)
Intel Research Press kit
SPAIS Publications:
Theocharous, G., Beckwith, R., Butko, N., Philipose, M. Tractable POMDP Planning Algorithms for Optimal Teaching in "SPAIS". International Joint Conferene on Artificial Intelligence (IJCAI) workshop on Plan Activity, and Intent Recognition (PAIR), Pasadena, California, July 2009.
Theocharous, G., Butko, N., Philipose, M. Designing a Mathematical Manipulatives Tutoring System using POMDPS. (pdf). POMDP Practitioners Workshop: Solving Real-world POMDP Problems, International Conference on Automated Planning and Scheduling (ICAPS). Toronto, May 2010
Wayang Outpost: Web-based Interactive Math/Intelligent Tutoring System, with Sensors.
I've followed the work of Beverly P. Woolf and her colleagues for some time. Much of their research has centered around a web-based application, Wayang Outpost, an intelligent electronic tutoring system that incorporates multimedia and animated adventures while providing activities designed to prepare teens for standardized math tests, such as the SAT and state-mandated end-of-course exams.
In recent years, the team has been using non-invasive sensors in their research, including a camera that views facial expressions, a posture-sensing device located in the seat of the student's chair, and a pressure-sensitive mouse, and a wireless skin conductance wristband. Data collected through all of these sensors can provide useful information about student learning. The system can also note when students try to "game" the system.
Related Publications
Woolf, B.P., Arroyo, I., Muldner, K., Burleson, W., Cooper, D., Dolan, R., Christopherson, R.M (2010)The Effect of Motivational Learning Companions on Low Achieving Students and Students with Disabilties (pdf) International Conference on Intelligent Tutoring Systems, Pittsburgh.
Abstract "We report the results of a randomized controlled evaluation of the effectiveness of pedagogical agents as providers of affective feedback. These digital learning companion were embedded in an intelligent tutoring system for mathematics, and were used by approximately one hundred students in two public high schools. Students in the control group did not receive the learning companions. Results indicate that low-achieving students—one third of whom have learning disabilities—had higher affective needs than their higher achieving peers; they initially considered math problem-solving more frustrating, less exciting, and felt more anxious when solving math problems. However, after they interacted with affective pedagogical agents, low-achieving students improved their affective outcomes, e.g., reported reduced frustration and anxiety."
Arroyo, I., Cooper, D.G., Burleson, W., Woolf, B.P., Muldner, K., Christopherson, R. (2009)
Emotion Sensors Go To School. AIED 2009. Pp. 17-24. IOS Press.
Low-tech description of Wayang Outpost, the math application used in the above publication:
Paul Franz, Recoder.Com 5/16/09
Cooper, D.G., Arroyo, I., Woolf, B.P., Muldner, K., Burleson, W., Christoperson, R. Sensors Model Student Self-Concept in the Classroom (pdf) UMass Amherst, June 22, 2009/UMAP 2009
