Embedded Files
Ongoing Funded Projects
Ongoing Funded Projects
Award number: 2405862
Total Grant:
$2,999,966.00
Start Date:
September 15, 2024
End Date:
August 31, 2028
i-SAIL: Investigation of students' learning, interest, and career aspirations in an integrated Science and Artificial Intelligence learning environment 🔗
To prepare a future-ready workforce, the AIAL lab is developing an innovative middle school curriculum that integrates fundamental AI concepts with core science disciplines. Supported by the NSF DRK-12 program, this initiative utilizes a user-friendly, block-based visual programming environment to allow students to solve real-world scientific problems without the barrier of complex syntax. Our research follows a design-based approach—collaborating directly with teachers to co-create curriculum units that foster inclusive learning environments for underrepresented and underserved students in STEM. By tracking the learning journeys of over 900 students, we are investigating how these integrated experiences influence scientific reasoning, AI literacy, and long-term career aspirations in technology and engineering.
Award number: 2426837
Total Grant:
$224,669.00
Start Date:
January 1, 2025
End Date:
December 31, 2027
NSF-AoF: Advanced Student Modeling and Tailored Large Language Models for Personalized Learning in Computer Science Education 🔗
General-purpose AI tools often fall short in educational settings, generating responses that can be inaccurate or mismatched to a student's actual knowledge level. Supported by the National Science Foundation, this project aims to close that gap by tailoring Large Language Models (LLMs) specifically for personalized computer science education. Our team is developing novel AI methods to continuously model students' problem-solving strategies and track their evolving competencies. By combining these fine-grained student models with advanced Reinforcement Learning with Human Feedback (RLHF) , we are training AI agents to generate highly relevant, pedagogical learning scaffolds—such as adaptive feedback, worked examples, and targeted practice problems. By integrating these techniques into existing intelligent learning environments, we aim to provide thousands of diverse learners with AI support that is not only technologically advanced but genuinely helpful for their learning journey.
Award number:Â 2418658
Total Grant:
$249,960.00
Start Date:
September 1, 2024
End Date:
August 31, 2027
C3PE: Comprehensive Personalized Programming Practice Environment 🔗
In collaboration with the University of Pittsburgh, Carnegie Mellon, and UMass, the AIAL lab is developing C-3PE, an AI-driven ecosystem designed to revolutionize introductory computer science education. This project utilizes a sophisticated "nested personalization" approach: an outer loop uses Large Language Model (LLM) powered adaptive testing to select the most informative next challenge for a student, while an inner loop employs preference optimization to provide real-time, personalized feedback tailored to the learner's specific knowledge state. By leveraging context-aware deep learning and knowledge tracing, C-3PE dynamically models student progress to determine whether a learner needs a worked example or a new problem to solve. Supported by the NSF RITEL program, our goal is to deploy these tools across diverse institutions and share our findings via open-source repositories to empower instructors worldwide.
Award number: 2236195
Total Grant:
$1,723,467.00
Start Date:
May 1, 2023
End Date:
April 30, 2027
INSIGHT: Transforming Introductory Computer Science with AI-Driven Assistance 🔗
Supported by the National Science Foundation, the INSIGHT project explores how artificial intelligence can transform introductory programming classes into highly active, engaging, and supportive environments for diverse learners. Our team is developing the INSIGHT AI-driven classroom assistant, a dual-purpose tool designed to provide students with real-time, adaptive feedback while giving instructors aggregate analytics on student coding behaviors. By equipping educators with these real-time insights, they can dynamically adapt their teaching to address students' conceptual roadblocks. In collaboration with a broad network of large and small public universities, as well as Historically Black Colleges and Universities (HBCUs), we are evaluating how students learn alongside AI assistants to establish effective, evidence-based principles for the future of STEM education.
Other Ongoing Projects
Other Ongoing Projects
Privacy-Preserving Federated Learning in Predictive Educational AI
This project is about two different applications of federated learning (FL). The first one focuses on link prediction in online social networks with an FL approach. For example, if there are multiple classroom/course datasets having the same features (horizontal FL platform), we want to train the global model based on the local model parameters. This means there are two steps: (i) Local Model Training - where each training unit (e.g., a local server at an institution) independently develops a local ML model, typically a neural network, using its own dataset; and (ii) Global Model Aggregation and Broadcast - where following local training, a central server periodically pulls and combines these local models into a comprehensive global model, which is then broadcast back to the units for further refinement. This step helps achieve a coherent global model that leverages learnings from all participating units. FL is adopted to ensure the training and fine-tuning of the models in a way that prevents data transfers across the network. Our goal is to find future links among students, and for this, we need temporal/time series data. With temporal data, we can find information about the linkage information of students across different courses/classrooms. Â
The second project is about integrating FL and code-aware knowledge tracing methods. Based on student code submissions, this framework offers better forecasting performance while maintaining student data privacy.
Concluded Projects
Concluded Projects
Award number: 2021330
Total Grant:
$174,938.00
Start Date:
July 1, 2020
End Date:
June 30, 2023
Analysis of a Simple, Low-cost Intervention's Impact on Retention of Women in Computer Science 🔗
Supported by the National Science Foundation, this project aims to increase the persistence of women in computer science by addressing critical gender disparities in the self-assessment of STEM abilities. Our research investigates a highly scalable, low-cost intervention: carefully worded email messages delivered to students in introductory courses that provide contextual performance feedback and specific encouragement. Building on pilot data showing that a single, targeted email can increase women's intentions to stay in the field by 18%, our team is conducting large-scale field and online experiments to maximize the efficacy of this approach. By combining quantitative performance data with qualitative student interviews, we aim to uncover the social-psychological processes driving career choices and equip educational institutions with a proven, easily implemented tool to promote gender equity across STEM disciplines.
Page updated
Report abuse