OKO, a Multimodal Intelligent Assistant for Small Group Collaborative Learning in Elementary Math

During Phase I in 2023, the project team developed a prototype platform for students to collaborate in solving math puzzles. The prototype employed vision and natural language processing (NLP) technologies to personalize learning by providing prompts to students as they collaborate on solving math puzzles, with an interface for educators to watch replays of multiple groups engaging in the activities simultaneously. At the end of Phase I, a pilot study included 2 educators and 40 students in grade 4 and 5. Results demonstrated that the prototype functioned as intended, educators were able to implement the intervention as a part of regular classroom practice, and a majority of students agreed that they enjoyed playing the game and agreed that the instructions were easy to understand. Most students were observed to be engaged and on-task; engaged while listening to one another, and eager to explain their answers to the members of their group.

In Phase II, the project team will fully develop the product, including: adding more math puzzles and collaborative games across grades 3 to 5; fully populating, testing, and finalizing the vision and NLP processing models that provide the groups real time prompts; adding a new generative AI feature to nudge and encourage participation of individual students in real time; and improving the educator dashboard, with data and reporting and summaries of group sessions. After development concludes, the project team will conduct a pilot study to l test the feasibility and usability, fidelity of implementation, and the promise of the product for improving math learning. The team will collect data from 20 grade 4 and 5 with 25 students per class for a total of approximately 500 students. Half of the classes will be randomly assigned to use the product and the other half to use business-as-usual activities. Researchers will compare pre-and-post scores using the iReady Diagnostic Assessment in Math to assess student growth on Common Core aligned math standards, as well as through researcher developed measures. Researchers will gather cost information using the "ingredients method" and will include all expenditures on things such as personnel, facilities, equipment, materials, and training.