Research

2025

1. IMOL Workshop

   Information-Theoretic Formulation and Combination of Intrinsic Rewards: Novelty, Surprise and Empowerment

   Tojoarisoa Rakotoaritina, Gaganpreet Jhajj, Chris Reinke, and 1 more author

   In Seventh International Workshop on Intrinsically Motivated Open-ended Learning, 2025

   Bib PDF

   @inproceedings{rakotoaritina2025informationtheoretic,
     title = {Information-Theoretic Formulation and Combination of Intrinsic Rewards: Novelty, Surprise and Empowerment},
     author = {Rakotoaritina, Tojoarisoa and Jhajj, Gaganpreet and Reinke, Chris and Doya, Kenji},
     booktitle = {Seventh International Workshop on Intrinsically Motivated Open-ended Learning},
     year = {2025},
     url = {https://openreview.net/forum?id=WN7ofwXNvv},
   }

2. RLC Workshop

   Decentralized Fire Seeking MARL UAVs

   Gaganpreet Jhajj, Tojoarisoa Rakotoaritina, and Fuhua Lin

   In Second Coordination and Cooperation in Multi-Agent Reinforcement Learning Workshop, 2025
3. IEEE SWC

   Representing and Tracing Students’ Cognitive Processes in Project-Based Learning Through the Function-Behavior-Structure Framework and Knowledge Graphs

   Jerry Ryan David Gustafson, Xiaokun Zhang, Gaganpreet Jhajj, and 1 more author

   In IEEE Smart World Congress 2025 (IEEE SWC’25), Aug 2025

   Abs

   This research proposes a framework to represent and track students’ cognitive processes during project-based software development using the Function-Behavior-Structure (FBS) model, Knowledge Graphs (KGs), and AI-assisted analysis. The goal is to make student thinking more visible, recursive, and actionable for both learners and instructors. The system supports real-time feedback and adaptive learning interventions by mapping student reflections, decisions, and misconceptions to evolving KGs. A prototype case study from a Computer Science Training Through Projects course illustrates how AI can automatically extract and update cognitive elements from student input, compare them to expert models, and provide targeted guidance. This approach builds on dynamic semantic network theory and leverages large language models to reduce instructor workload while improving transparency and formative assessment. Future work includes piloting the system with students to evaluate usability, precision, and impact on metacognitive development.

4. IEEE SWC

   Augmenting Japanese Language Acquisition via LLMs and ASR

   Gaganpreet Jhajj and Fuhua Lin

   In IEEE Smart World Congress 2025 (IEEE SWC’25), Aug 2025

   Abs

   Duolingo has recently expanded its speech recognition capabilities to include languages like Japanese. However, opportunities remain to enhance the speaking and listening practice experience, especially for early-stage learners. We introduce a prototype adaptive learning approach that leverages OpenAI’s Whisper and GPT-3.5 Turbo to generate targeted Anki flashcards and role-play dialogue.

5. IEEE SWC

   Linguistic Analysis of Japanese Text Simplification and Implications for AI-Driven Educational Tools

   Gaganpreet Jhajj and Fuhua Lin

   In IEEE Smart World Congress 2025 (IEEE SWC’25), Aug 2025

   Abs

   This study examines the SNOW Simplified Japanese Corpora (T15 and T23) to analyze how text simplification transforms linguistic features across Japanese and English, intending to inform AI-driven educational tools. Applying POS tagging, dependency parsing, and named-entity recognition to parallel texts, we identify distinctive patterns including increased noun usage and reduced determiner frequency in simplified Japanese. This analysis reveals that simplified Japanese texts demonstrate higher noun density, fewer determiners, and selective preservation of named entities compared to their English counterparts. These findings suggest that simplification reduces language learners’ cognitive load by anchoring discourse in more concrete concepts. The linguistic patterns identified can inform the development of adaptive AI-powered learning systems that effectively balance accessibility with authentic language exposure and nuance.

6. IEEE SWC

   Exploring Neuromorphic Computing for UAV Navigation

   Gaganpreet Jhajj and Fuhua Lin

   In IEEE Smart World Congress 2025 (IEEE SWC’25), Aug 2025

   Abs

   Energy-constrained UAV platforms operating at the network edge face significant computational challenges when implementing traditional artificial intelligence approaches for autonomous navigation. Neuromorphic computing, inspired by the energy-efficient processing principles of biological neural systems, offers a promising alternative through spiking neural networks (SNNs) that process information via discrete spike events rather than continuous activations. This paper presents a preliminary exploration of neuromorphic computing for UAV navigation through comparative analysis with traditional reinforcement learning methods. We conduct an exploratory study comparing a bio-inspired SNN approach against Q-learning in a controlled 20×20 grid world environment with clustered obstacles, training both agents for 500 episodes under identical conditions. Our software-based simulation investigation reveals essential insights into task-algorithm compatibility and highlights critical distinctions between theoretical neuromorphic advantages and practical implementation realities. While this work represents a limited simulation-based study that cannot fully capture the energy efficiency and processing advantages of dedicated neuromorphic hardware, it establishes essential baseline comparisons for future neuromorphic UAV research. It emphasizes the need for evaluation in continuous, dynamic environments where bio-inspired computation principles are expected to excel.

7. IEEE SWC

   Can Question Validation Criteria Improve the Quality of LLM-Generated Multiple-Choice Questions?

   Raymond Morland, Gaganpreet Jhajj, Hongxin Yan, and 8 more authors

   In IEEE Smart World Congress 2025 (IEEE SWC’25), Aug 2025

   Abs

   Adaptive assessment has been extensively researched as a key strategy for providing formative assessments in education. Fundamental to adaptive assessment is the development and ongoing maintenance of a substantial and well-structured question bank. However, manual question development by human instructors is time-consuming and poses a significant barrier to adopting adaptive assessment. To address this problem, research has been conducted in the use of Large Language Models (LLMs) for automated question generation. However, LLMs tend to hallucinate, necessitating human-in-the-loop question validation. Furthermore, to ensure alignment with the instructional standards and learning outcomes of a course, it is essential that instructors verify that the generated questions satisfy a defined set of Question Validation Criteria (QVC). In this study, we evaluate the effectiveness of embedding QVC directly into LLM prompts for the automatic generation of multiple-choice questions (MCQ) with feedback for use in adaptive formative assessments. Using a novel web-based application for question generation, MCQs were produced for three undergraduate-level courses. Validation by subject matter experts revealed no statistically significant improvement in quality between questions generated with and without embedded QVC in the prompts. These findings contribute to the growing body of research indicating that prompt engineering alone may be insufficient to achieve improvements in the quality of generated output for educational purposes.

8. ITS

   Neuromorphic Knowledge Representation: SNN-Based Relational Inference and Explainability in Knowledge Graphs

   Gaganpreet Jhajj, Jerry Ryan David Gustafson, Raymond Morland, and 4 more authors

   In Generative Systems and Intelligent Tutoring Systems, Aug 2025

   Abs

   This paper explores a neuromorphic approach to semantic knowledge representation using spiking neural networks (SNNs) for relational inference within knowledge graphs (KGs). While traditional KG models often rely on dense embeddings that lack interpretability, SNNs offer a biologically inspired alternative by encoding relationships through discrete, event-driven spikes. We evaluate three architectures—LIF-Base, RecurrentSNN, and LiquidSNN—on a dataset of 54 computer science-related KG triplets. Performance is assessed using relationship classification accuracy, temporal and spatial stability, and spike variability. Results show that relationship types produce distinct spike activation patterns, with LiquidSNN achieving the highest accuracy and spatial coherence. These findings support the potential of SNNs for structured, interpretable KG reasoning, with future applications in adaptive learning systems and explainable AI.

9. Enhancing Project-Based Learning With a GenAI Tool Based on Retrieval: Augmented Generation and Knowledge Graphs

   Jerry Ryan David Gustafson, Gaganpreet Jhajj, Xiaokun Zhang, and 1 more author

   In AI Applications and Strategies in Teacher Education, Aug 2025

2024

1. ITS

   Educational knowledge graph creation and augmentation via LLMs

   Gaganpreet Jhajj, Xiaokun Zhang, Jerry Ryan Gustafson, and 2 more authors

   In International Conference on Intelligent Tutoring Systems, Aug 2024
2. ITS

   Preliminary Systematic review of open-source large language models in education

   Michael Pin-Chuan Lin, Daniel Chang, Sarah Hall, and 1 more author

   In International Conference on Intelligent Tutoring Systems, Aug 2024
3. EMNLP Workshop

   Jack and the BeansTALK: Towards Question Answering in Plant Biology

   Gaganpreet Jhajj and Yumeko Nomura

   In Eighth Widening NLP Workshop (WiNLP 2024) at EMNLP 2024, Aug 2024

2023

1. Unveiling uncertainty: Supporting learners through nlp-driven confusion identification

   Gaganpreet Jhajj, M Ali Akber Dewan, and Fuhua Lin

   In 2023 IEEE Intl Conf on Dependable, Autonomic and Secure Computing, Intl Conf on Pervasive Intelligence and Computing, Intl Conf on Cloud and Big Data Computing, Intl Conf on Cyber Science and Technology Congress (DASC/PiCom/CBDCom/ CyberSciTech), Aug 2023

2022

1. Current methods and future research in the diagnosis of Alzheimer’s disease

   J Gaganpreet and F Mark

   Columbia Undergrad Sci J, Aug 2022

2018

1. Current opinions in organ allocation

   Contributors C4 Article (Appendix 1)

   American Journal of Transplantation, Aug 2018