Sagnik Dakshit

Large language models (LLMs) are increasingly being used in education, yet their correctness alone does not capture the quality, reliability, or pedagogical validity of their problem-solving behavior, especially in mathematics, where multi-step logic, symbolic reasoning, and conceptual clarity are critical. Conventional evaluation methods largely focus on final answer accuracy and overlook the reasoning process. To address this gap, we introduce a novel interpretability framework for analyzing LLM-generated solutions using undergraduate calculus problems as a representative domain. Our approach combines reasoning flow extraction and decomposing solutions into semantically labeled operations and concepts with prompt ablation analysis to assess input salience and output stability. Using structured metrics such as reasoning complexity, phrase sensitivity, and robustness, we evaluated the model behavior on real Calculus I–III university exams and compared it with the performances of students enrolled in the courses. Our findings revealed that LLMs often produce syntactically fluent yet conceptually flawed solutions with reasoning patterns sensitive to prompt phrasing and input variation. This framework enables a fine-grained diagnosis of reasoning failures, supports curriculum alignment, and informs the design of interpretable AI-assisted feedback tools. The framework was evaluated on Gemma 3, an open-access large language model, across zero-shot, retrieval-augmented generation, and contextual retrieval configurations, using nine real undergraduate calculus examinations from three course levels. To our knowledge, this is the first paper to apply a combined reasoning flow decomposition and prompt ablation framework to real undergraduate calculus examinations, benchmarked against actual student cohort performance, laying the foundation for the transparent and responsible deployment of AI in STEM learning environments.

As artificial intelligence (AI) is increasingly used in classrooms, tutoring systems, and learning platforms, it is essential that these tools are not only powerful, but also easy to understand, fair, and supportive of real learning. Many current AI systems can generate fluent responses or accurate predictions, yet they often fail to clearly explain their decisions, reflect students’ cultural contexts, or give learners and educators meaningful control. This gap can reduce trust and limit the educational value of AI-supported learning. This paper introduces the PEARL framework, a human-centered approach for designing and evaluating explainable AI in education. PEARL is built around five core principles: Pedagogical Personalization (adapting support to learners’ levels and curriculum goals), Explainability and Engagement (providing clear, motivating explanations in everyday language), Attribution and Accountability (making AI decisions traceable and justifiable), Representation and Reflection (supporting fairness, diversity, and learner self-reflection), and Localized Learner Agency (giving learners control over how AI explains and supports them). Unlike many existing explainability approaches that focus mainly on technical performance, PEARL emphasizes how students, teachers, and administrators experience and make sense of AI decisions. The framework is demonstrated through simulated examples using an AI-based tutoring system, showing how PEARL can improve feedback clarity, support different stakeholder needs, reduce bias, and promote culturally relevant learning. The paper also introduces the PEARL Composite Score, a practical evaluation tool that helps assess how well educational AI systems align with ethical, pedagogical, and human-centered principles. This study includes a small exploratory mixed-methods user study (N = 17) evaluating example AI tutor interactions; no live classroom deployment was conducted. Together, these contributions offer a practical roadmap for building educational AI systems that are not only effective, but also trustworthy, inclusive, and genuinely supportive of human learning.

Simulations constitute a fundamental component of medical and nursing education and traditionally employ standardized patients (SP) and high-fidelity manikins to develop clinical reasoning and communication skills. However, these methods require substantial resources, limiting accessibility and scalability. In this study, we introduce CLiVR, a Conversational Learning system in Virtual Reality that integrates large language models (LLMs), speech processing, and 3D avatars to simulate realistic doctor-patient interactions. Developed in Unity and deployed on the Meta Quest 3 platform, CLiVR enables trainees to engage in natural dialogue with virtual patients. Each simulation is dynamically generated from a syndrome-symptom database and enhanced with sentiment analysis to provide feedback on communication tone. Through an expert user study involving medical school faculty (n=13), we assessed usability, realism, and perceived educational impact. Results demonstrated strong user acceptance, high confidence in educational potential, and valuable feedback for improvement. CLiVR offers a scalable, immersive supplement to SP-based training.

Large Language Models (LLMs) are increasingly being used in education, yet their correctness alone does not capture the quality, reliability, or pedagogical validity of their problem-solving behavior, especially in mathematics, where multistep logic, symbolic reasoning, and conceptual clarity are critical. Conventional evaluation methods largely focus on final answer accuracy and overlook the reasoning process. To address this gap, we introduce a novel interpretability framework for analyzing LLM-generated solutions using undergraduate calculus problems as a representative domain. Our approach combines reasoning flow extraction and decomposing solutions into semantically labeled operations and concepts with prompt ablation analysis to assess input salience and output stability. Using structured metrics such as reasoning complexity, phrase sensitivity, and robustness, we evaluated the model behavior on real Calculus I to III university exams. Our findings revealed that LLMs often produce syntactically fluent yet conceptually flawed solutions, with reasoning patterns sensitive to prompt phrasing and input variation. This framework enables fine-grained diagnosis of reasoning failures, supports curriculum alignment, and informs the design of interpretable AI-assisted feedback tools. This is the first study to offer a structured, quantitative, and pedagogically grounded framework for interpreting LLM reasoning in mathematics education, laying the foundation for the transparent and responsible deployment of AI in STEM learning environments.

The growing presence of large language models (LLMs) and connected sensors or the Internet of Things (IoT) offers substantial potential across various tasks and fields, albeit with accompanying challenges. On the other hand, as advances in spatial computing redefine immersive interactions, ambient intelligence becomes essential for enabling environments that intuitively respond to user behavior without explicit commands. By fusing real-time sensory input with LLM-driven reasoning, ambient intelligence transforms spatially embedded systems into adaptive, context-aware, and human-centered digital experiences. In fact, according to reports,1 two prominent technological trends of 2025 are spatial computing and ambient invisible intelligence.

The emergence of Large Language Models (LLMs) has significantly impacted the field of Natural Language Processing and has transformed conversational tasks across various domains because of their widespread integration in applications and public access. The discussion surrounding the application of LLMs in education has raised ethical concerns, particularly concerning plagiarism and policy compliance. Despite the prowess of LLMs in conversational tasks, the limitations of reliability and hallucinations exacerbate the need to guardrail conversations, motivating our investigation of RAG in computer science higher education. We developed Retrieval Augmented Generation (RAG) applications for the two tasks of virtual teaching assistants and teaching aids. In our study, we collected the ratings and opinions of faculty members in undergraduate and graduate computer science university courses at various levels, using our personalized RAG systems for each course. This study is the first to gather faculty feedback on the application of LLM-based RAG in education. The investigation revealed that while faculty members acknowledge the potential of RAG systems as virtual teaching assistants and teaching aids, certain barriers and features are suggested for their full-scale deployment. These findings contribute to the ongoing discussion on the integration of advanced language models in educational settings, highlighting the need for careful consideration of ethical implications and the development of appropriate safeguards to ensure responsible and effective implementation.

One of the hindrances in the widespread acceptance of deep learning–based decision support systems in healthcare is bias. Bias in its many forms occurs in the datasets used to train and test deep learning models and is amplified when deployed in the real world, leading to challenges such as model drift. Recent advancements in the field of deep learning have led to the deployment of deployable automated healthcare diagnosis decision support systems at hospitals as well as tele-medicine through IoT devices. Research has been focused primarily on the development and improvement of these systems leaving a gap in the analysis of the fairness. The domain of FAccT ML (fairness, accountability, and transparency) accounts for the analysis of these deployable machine learning systems. In this work, we present a framework for bias analysis in healthcare time series (BAHT) signals such as electrocardiogram (ECG) and electroencephalogram (EEG). BAHT provides a graphical interpretive analysis of bias in the training, testing datasets in terms of protected variables, and analysis of bias amplification by the trained supervised learning model for time series healthcare decision support systems. We thoroughly investigate three prominent time series ECG and EEG healthcare datasets used for model training and research. We show the extensive presence of bias in the datasets leads to potentially biased or unfair machine-learning models. Our experiments also demonstrate the amplification of identified bias with an observed maximum of 66.66%. We investigate the effect of model drift due to unanalyzed bias in datasets and algorithms. Bias mitigation though prudent is a nascent area of research. We present experiments and analyze the most prevalently accepted bias mitigation strategies of under-sampling, oversampling, and the use of synthetic data for balancing the dataset through augmentation. It is important that healthcare models, datasets, and bias mitigation strategies should be properly analyzed for a fair unbiased delivery of service.