An official website of the United States government
The paper develops datasets and methods to assess student participation in real-life collaborative learning environments. In collaborative learning environments, students are organized into small groups where they are free to interact within their group. Thus, students can move around freely causing issues with strong pose variation, move out and re-enter the camera scene, or face away from the camera. We formulate the problem of assessing student participation into two subproblems: (i) student group detection against strong background interference from other groups, and (ii) dynamic participant tracking within the group. A massive independent testing dataset of 12,518,250 student label instances, of total duration of 21 hours and 22 minutes of real-life videos, is used for evaluating the performance of our proposed method for student group detection. The proposed method of using multiple image representations is shown to perform equally or better than YOLO on all video instances. Over the entire dataset, the proposed method achieved an F1 score of 0.85 compared to 0.80 for YOLO. Following student group detection, the paper presents the development of a dynamic participant tracking system for assessing student group participation through long video sessions. The proposed dynamic participant tracking system is shown to perform exceptionally well, missing a student in just one out of 35 testing videos. In comparison, a stateof- the-art method fails to track students in 14 out of the 35 testing videos. The proposed method achieves 82.3% accuracy on an independent set of long, real-life collaborative videos.
more »
« less
Dynamic Group Interactions in Collaborative Learning Videos
We introduce a new method to detect student group interactions in collaborative learning videos. We consider the following video activities: (i) human to human, (ii) human to others, and (iii) lack of any interaction. The system uses multidimensional AM-FM methods to detect student faces, hair, and then use the results to detect possible interactions. We use dynamic graphs to represent group interactions within each video.We tested our methods with 15 videos and achieved an 84% accuracy for students facing the camera and 76% for students facing both towards and away from the camera.
more »
« less
- PAR ID:
- 10110859
- Date Published:
- Journal Name:
- 52nd Asilomar Conference on Signals, Systems, and Computers
- Page Range / eLocation ID:
- 1528 to 1531
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Many large introductory classes are taught in stadium-style classrooms, which makes group work more difficult due to the room layout and immobile seating. These classrooms may create challenges for an instructor who wants to monitor student engagement because the layouts make it difficult to interact with the students as they work. Student nonverbal actions, such as eyes on the paper or an unsettled gaze, can be used to determine when students are actively engaged during group work. While other methods have been implemented to determine student actions during a class period, in larger settings these protocols require time-consuming data collection and cannot give in-the-moment feedback. In this study, student verbal and nonverbal interactions were analyzed and compared to determine the types of nonverbal interactions students take when collaboratively engaging in group work during lectures. It was found that a larger variety of nonverbal interactions, such as gesturing and leaning, were used when students were collaboratively working within their groups. Instructors of large enrollment classrooms can use the results of this work to aid in their facilitation of group work within stadium-style classrooms.more » « less
-
Problem solving is a signature skill of engineers. Incorporating videos in engineering education has potential to stimulate multi-senses and further open new ways of learning and thinking. Here, problem solving was examined on problems written by previous students that applied course concepts by reverse engineering the actions in videos. Since the videos usually come from YouTube, the student-written problems are designated YouTube problems. This research focused on examining the rigor of YouTube problems as well as students’ problem-solving skills when solving YouTube problems compared to Textbook problems. A quasi-experimental, treatment/control group design was employed, and data collected was evaluated using multiple instruments. NASA Task Load Index survey was used to collect 1200 ratings that assessed rigor of homework problems. Problem-solving ability was assessed using a previously-developed rubric with over 2600 student solutions scored. In the treatment group where students were assigned ten Textbook and nine YouTube problems, students reported an overall similarity in rigor for both YouTube and Textbook problems. Students in the treatment group displayed 6% better problem solving when completing YouTube problems compared to Textbook problems. Although higher perceptions of problem difficulty correlated with lower problem-solving ability across both groups and problem types, students in the treatment group exhibited smaller decreases in problem-solving ability as a result of increasing difficulty in the Textbook problems. Overall, student-written problems inspired by YouTube videos can easily be adapted as homework practice and possess potential benefits in enhancing students’ learning experience. Link: https://www.ijee.ie/contents/c370521.htmlmore » « less
-
We introduce the problem of detecting a group of students from classroom videos. The problem requires the detection of students from different angles and the separation of the group from other groups in long videos (one to one and a half hours). We use multiple image representations to solve the problem. We use FM components to separate each group from background groups, AM-FM components for detecting the back-of-the-head, and YOLO for face detection. We use classroom videos from four different groups to validate our approach. Our use of multiple representations is shown to be significantly more accurate than the use of YOLO alone.more » « less
-
Abstract As the use of computers in education increases, adaptive learning platforms are becoming more common. However, these adaptive systems are typically designed to support acquisition of declarative knowledge and/or procedural fluency but rarely address conceptual learning. In this work, we developed the Crystallography Adaptive Learning Module (CALM) for materials science to provide students a tool for individualized conceptual learning. We used a randomized quasi‐experimental design comparing two instructional designs with different levels of computer‐provided direction and student agency. Undergraduate students were randomly assigned to one of two different instructional designs; one design had students complete an individualized, adaptive path using the CALM (N = 80), and the other gave students the freedom to explore CALM's learning resources but with limited guidance (N = 85). Within these two designs, we also investigated students among different cumulative grade point average (GPA) groups. While there was no statistically significant difference in the measure of conceptual understanding between instructional designs or among the groups with the same GPA, there is evidence to suggest the CALM improves conceptual understanding of students in the middle GPA group. Students using CALM also showed increased participation with the interactive learning videos compared to the other design. The number of videos watched in each instructional condition aligns with overall academic performance as the low GPA group received the most assigned supplements but watched the least videos by choice. This study provides insight for technology developers on how to develop educational adaptive technology systems that provide a proper level of student agency to promote conceptual understanding in challenging STEM topics.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/ACSSC.2018.8645132
