Integrating Artificial Neural Networks into Undergraduate Physics Laboratory: A Compound Pendulum Case Study

2026-05-14T10:47:57Z

Artificial Neural Networks (ANNs) are becoming important tools in physics research and education because they help in data analysis and complement traditional analytical methods. In this work, ANN modeling is introduced in a standard compound pendulum experiment used to determine the acceleration due to gravity, g. The aim is not to replace the conventional analytical method, but to demonstrate how machine learning can support experimental data analysis in undergraduate physics laboratories. Students first measure parameters such as effective length, time period, and angular displacement, and determine g using standard analytical methods with uncertainty analysis. These experimentally obtained data are then used to train and test an ANN model. The dataset is divided into training (70%), validation (15%), and testing (15%) groups. The experimentally determined value of gravitational acceleration was 1009.03 +/- 6.82 cm/s^2, while the ANN predicted a mean value of 1009.029858 cm/s^2 with a mean absolute error of 0.000592 cm/s^2. The close agreement between the experimental and ANN-predicted values shows that the ANN successfully learned the relationship between the pendulum parameters and g. However, the ANN prediction error should not be considered as an improvement in experimental accuracy because the model is trained using experimentally derived data. Instead, the ANN serves as a useful computational and educational tool that introduces students to regression, validation, overfitting, and data-driven analysis alongside traditional experimental physics.

A Pedagogical MKS-based Electromagnetic Unit Convention with $ε_0 = μ_0 = 1/c$

2026-05-13T13:58:42Z

We propose a pedagogical, rationalized MKS-based convention for electromagnetic quantities designed to reduce cognitive load in undergraduate undergraduate electromagnetism. By setting vacuum constants to $\varepsilon_0 = μ_0 = 1/c$, we preserve the familiar structure of Maxwell's equations while making the role of the speed of light explicit. In this convention, electrical units are expressed directly in terms of mechanical units (e.g.\ $[\mathrm{nuA}] = \sqrt{\mathrm{J/s}}$), effectively reducing the number of independent base units. A striking pedagogical consequence is that electrical resistance becomes dimensionless, capacitance and inductance acquire units of time, and radiation pressure reduces to $|\mathbf{E}\times \mathbf{B}|$, greatly simplifying dimensional analysis for circuits and fields. We introduce corresponding non-SI units (\textit{nu}-units), provide conversion relations to SI, and demonstrate the potential utility of this system through comparative ``before/after'' derivations of the wave equation, electromagnetic energy density, radiation pressure, and the Bohr atom. Preliminary empirical support is provided by student attitude surveys administered to $N_1 = 46$ and $N_2 = 39$ students in an undergraduate physics course, which showed a statistically significant improvement in the perceived clarity of the wave equation derivation after exposure to the nu-system ($p = 0.005$, Mann--Whitney $U$ test), and a majority preference for the dimensionless-resistance feature.

A Framework for institutional change in the age of AI

2026-05-12T21:11:13Z

Generative AI is rapidly reshaping STEM higher education. Not only are our educational practices changing, but how we think about educational transformation must adapt. Existing models of institutional change in STEM, aimed at interactive engagement, have largely followed an adoption logic: relatively stable, well-researched educational practices are evaluated and then scaled. These assumptions do not hold for generative AI, which is an arrival technology -- entering classrooms before a sufficient pedagogical evidence base could form. Building on recent decades of work on STEM institutional change, we propose a framework identifying six dimensions along which prior change models must be reconsidered in light of AI: three concerning the tools at the center of reform (the tool's evidence base, rate of change, and scope), and three concerning the people involved in change (faculty, change agents, and students). For each dimension, we examine how AI-era assumptions differ from those underlying prior interactive engagement reforms and derive design implications, including: privileging humble and local inquiries; organizing reform around pedagogical approaches rather than specific tools; repositioning change agents as facilitators of collective inquiry; and engaging students as partners in reform. Collectively, the six dimensions and design implications constitute a new framework for adapting change models to support institutions under conditions of genuine uncertainty. Finally, we illustrate how the framework may be applied through a brief case-study of a faculty workshop series carried out in a university physics department to support instructors adapting to this modern AI era.

Power, Prescription, and Postpositivism: Considerations for collecting and representing neurodiversity demographic information in physics education research

2026-05-12T18:01:39Z

Demographic data collection is essential in education research, as demographic data allows researchers to better describe the participant population they study and to contextualize findings. However, current research practices for neurodiversity demographics often rely on prescriptive methods (e.g., requiring participants to report official diagnoses) rather than allowing participants to self-identify. This approach can: a) not allow participants to express their intersecting identities in ways that are authentic; and b) limit trustworthiness and reliability of the data and interpretation. In addition, inconsistent dissemination and representation of demographic data across studies hinder the accessibility and usability of this work. Through a literature review of neurodivergent student experiences with learning and performing STEM, we identified widespread discrepancies in how demographic information is collected and reported. This paper explores how neurodivergent identities can be more accurately and inclusively represented in education research. We present findings of a thematic analysis on the ways neurodivergent demographic data collection is done in the literature using data from a systematic literature review on neurodivergent science, technology, engineering, and mathematics (STEM) learning and performance. We call on the PER community to contribute to the development of a framework that centers participant autonomy while supporting clarity, consistency, and future research use.

The Gendered Cost of Lower Grades: Women's Physics Perceived Recognition and Identity Suffer Disproportionately If They Earn Less Than A Grade

2026-05-12T05:52:59Z

Perceptions of disciplinary recognition and identity can be shaped by various forms of feedback and experiences. Here we focus on the potential effects of course grades on the perceievd recognition and physics identity of students. We analyze patterns in changes in physics identity and perceived recognition from pre course to post course across three cohorts of university students enrolled in calculus-based Physics 1 (N=1,681). Students not receiving A grade, on average, showed declines in physics identity and perceived recognition. Even a B grade resulted in declines, and the declines were nonlinear across lower grades. Changes in perceived recognition fully mediated the changes in identity. Importantly, women showed significantly larger declines in identity and perceived recognition, compared to men, if they got less than A grade. The gender moderation was specifically localized to changes in perceived recognition, with no further gender effects on identity beyond the cascading effects on perceived recognition.

The Solar System as a lab for the Law of Universal Gravitation

2026-05-11T17:19:17Z

The Law of Universal Gravitation is part of middle and high school's general physics and astronomy curricula. This topic is included in the most popular physics textbooks available as a fact whose origin remains in the detailed work of Sir Isaac Newton 300 years ago. Consequently, its mathematical form is presented as an equation without any deductive process. Nevertheless, deduction of the mathematical form of this law is an opportunity to discuss how a deductive process can be performed using the data available on the Internet from reliable sources.

Preparing Students for AI-Powered Materials Discovery: A Workflow-Aligned Framework for AI Literacy, Equity, and Scientific Judgment

2026-05-10T16:11:01Z

Artificial intelligence (AI) is reshaping education, scientific training, and materials discovery. In materials science, AI models increasingly support property prediction, experiment prioritization, and hypothesis generation; however, the limiting factor is no longer only algorithmic capability but also whether students and educators can use AI with domain-specific scientific judgment. This workshop-informed white paper and curriculum-oriented position article argues that AI education for AI-powered materials discovery must move beyond tool access and surface-level interaction with generative AI systems toward a workflow-aligned model of AI literacy. We connect AI literacy to materials-informatics competencies: data provenance, domain-specific featurization, model validation, uncertainty quantification, physics informed reasoning, reproducibility, and experimental feedback. We also emphasize outcome-oriented equity: institutions should evaluate not only access, participation, and engagement, but also whether AI-enabled instruction produces comparable learning gains, transfer of learning, confidence calibration, defined as the alignment with students confidence and the quality or correctness of their work, persistence, and research readiness across student subgroups. The paper synthesizes relevant evidence, identifies risks for learners such as cognitive off-loading and cognitive surrender, and provides a dual-track curriculum model and implementation recommendations such as curriculum guides and an assessment plan for courses, bootcamps, workshops, and program-level reform. The central goal is to prepare students to become better scientists, not merely more efficient users of AI tools.

Performance and failure modes of AI chatbots on a novel concept inventory on relativity in classical mechanics

2026-05-10T15:30:11Z

AI chatbots are increasingly used by students as study tools in physics, raising practical questions about their reliability on conceptual tasks. Existing evaluations of large language models (LLMs) on physics concept inventories rely almost exclusively on instruments that have been publicly available for years and likely appear in model training data, making it difficult to disentangle physics competence from familiarity with the test items themselves. We address this issue by evaluating three frontier LLMs (GPT-5.2, Gemini 3 Pro, Gemini 3 Flash) on the Classical Relativity Concept Inventory (CRCI), a recently developed and validated 21-item instrument on Galilean relativity that was not publicly available at the time of testing. Each item was administered 30 times per model, and all 1890 responses were qualitatively coded along three dimensions: visual interpretation, physics reasoning, and coordination. Mean accuracy was 97% for Gemini 3 Flash, 89% for Gemini 3 Pro, and 73% for GPT-5.2, compared to 62% for the student sample (N = 267). However, all three models fail completely on a small number of items. The qualitative analysis shows that these failures stem predominantly from misinterpretations of visual content rather than from deficits in physics knowledge, and that LLM errors differ structurally from those of students: when models err, they converge on a single distractor with high consistency, whereas student errors are more broadly distributed. These findings indicate that chatbot reliability on conceptual physics is item-dependent and unpredictable, with direct implications for how concept inventories are administered.

Using Consumer Cameras to Observe Scintillation Light from Radiation

2026-05-10T13:07:28Z

For a long time, the cloud chamber was the only educational tool available for measuring radiation. In recent years, simple radiation detectors combining scintillators with silicon photomultipliers have become increasingly common for these purposes. However, students are not able to see the scintillation light, the core process of radiation measurements with scintillators. Therefore, we explored the possibility of detecting scintillation light using two general-purpose cameras. In addition, we examined how differences in the spatial distribution relate to radiation types and energies. Scintillation light were able to be measured by a general-use camera, and their spatial distribution indicates radiation energy. This method could be utilized as an accessible imaging setup to compare radiation properties in a classroom.

Graduate Training in Quantum Information Science and Engineering: Lessons, Challenges, and a Roadmap from the NSF Research Traineeship Programs

2026-05-08T21:42:45Z

Since 2019, eighteen NSF Research Traineeship (NRT) awards in quantum information science and engineering (QISE) and adjacent fields have been funded, constituting the largest NSF-coordinated investment in graduate QISE training in the United States. Synthesizing lessons from our programs, we work through the central tensions that every QISE graduate program must negotiate: between depth in a home discipline and breadth across the field, between structured instruction and open-ended experiential and hands-on learning, and between training individual specialists and cultivating teams that collectively cover all areas of QISE. We describe the structural and pedagogical innovations the NRT programs have developed in response, assess what is working and what remains unresolved, and sketch 12 open problems the community will need to address as QISE graduate education scales beyond the well-resourced research universities where it has up till now been mainly concentrated. Eight concrete recommendations follow: (1) adopt the startup model of team-based training as an organizing philosophy; (2) invest immediately in sensing and communication curriculum development; (3) build student agency into program governance, not just activities; (4) establish structural mechanisms for industrial engagement rather than depending on goodwill; (5) design for sustainability from year one; (6) develop graduate-level textbooks spanning all three QISE pillars: computing, sensing, and communications; (7) establish shared outcome assessment instruments across programs; and (8) develop structured mechanisms for faculty professional development in QISE.

From Floors to Electrons: Using a Building Analogy and Cartooning to Teach Quantum Numbers

2026-05-08T18:16:32Z

Aspects of quantum physics are no longer confined to the upper years of a physics degree. Concepts like superposition or entanglement that were once reserved for second- or third-year undergraduate courses now deserve attention earlier in a student's curriculum. Technology is changing at a pace that requires engaged citizens to understand some of the quantum basics if they are to make sense of the world. This paper offers a cartoon building analogy that teachers can use to introduce quantum numbers to their students.

Strategies for Collecting Multi-Institutional Data in Discipline-Based Education Research

2026-05-08T13:54:14Z

Multi-institutional studies are critical for advancing discipline-based education research (DBER) because they allow us to determine where and for whom research findings are applicable. Despite this benefit, such studies remain relatively rare due to the complexities of coordinating data collection across different institutions. In this paper, we describe key challenges and propose actionable strategies for implementing multi-institutional DBER studies. We focus on navigating Institutional Review Board procedures, recruiting participants from a range of institution types, standardizing data sources across institutions, and managing logistics. We also provide an applied example of these strategies from a national research project in which we collected concept inventory data, social network surveys, and classroom observations from 31 introductory physics instructors at 28 institutions in the United States.

Using Computational Physics Essays to Facilitate Engineering Students' Computational Thinking

2026-05-07T23:29:50Z

Background: As traditional coding tasks in education become increasingly vulnerable to the use of Generative AI, there is a critical need for authentic, project-based assessments that evaluate students' scientific inquiry. To address this need, we adapted the existing Computational Essay framework to create the Computational Physics Essay (CPE). Administered as a culminating capstone project, the CPE required introductory engineering students to use Python within Jupyter Notebooks to iteratively model real-world physics systems. We analyzed a random sample of CPE submissions (N = 100) using a customized 20-item rubric based on Weintrop's computational thinking (CT) taxonomy. Results: The project-based constraint successfully elicited a high variety of CT practices. Students demonstrated high proficiency in Modeling and Systems Thinking, with 99% successfully investigating complex systems as a whole. Furthermore, the use of CT practices strongly correlated (\r{ho}= 0.75) with expert ratings of the overall quality of the CPE. While some students showed expected novice weaknesses in software modularity, the CPE successfully shifted their epistemic frame toward physical sensemaking. Conclusions: Situating computation within real-world capstone projects provides a robust framework for assessing CT, bridging the gap between programming and scientific argumentation in introductory engineering students.

LaTA: A Drop-in, FERPA-Compliant Local-LLM Autograder for Upper-Division STEM Coursework

2026-05-06T20:06:06Z

Large-language-model (LLM) graders promise to relieve the grading burden of upper-division STEM courses, but most deployments to date send student work to third-party APIs, violating FERPA and exposing institutions to data risk while requiring substantial assignment modification. We present $\textbf{LaTA}\ (\textit{LaTeX Teaching Assistant})$, a drop-in, open-source autograder that runs entirely on commodity on-premises hardware and assumes a LaTeX-native workflow already adopted by many engineering and physics courses. LaTA implements a four-stage pipeline (ingest, segment, grade, report) using a locally hosted open-weight chain-of-thought LLM grader (gpt-oss:120b) that compares student work to an instructor-authored reference solution and applies a YAML rubric with binary per-item scoring. We deployed LaTA in Winter~2026 in ME 373 (Mechanical Engineering Methods) at Oregon State University, grading every weekly assignment for approximately 200 students on a single Mac Studio at \$0 marginal cost per assignment and 1--3 minutes of wall-clock time per submission, enabling regrading of corrected assignments and greatly expanded TA office hour offerings. The instructor-confirmed grading-error rate held at roughly $0.02$--$0.04\%$ per rubric line item across the term. Relative to the same instructor's previous traditionally-graded cohort, the LaTA-graded cohort outperformed by approximately $11\%$ on the midterm exam and $8\%$ on the final exam, and reported large gains in self-assessed confidence on every stated learning objective ($N = 159$ survey responses, $Δ\geq +1.49$ Likert points, $p < 10^{-27}$ on every comparison). We release the code under AGPLv3.

Determining Viscosity of a Liquid with Smartphone Sensors: A Classroom-Friendly Approach Using Damped Oscillations

2026-05-06T16:42:48Z

This study presents a classroom-friendly method for measuring the coefficient of viscosity of a liquid using a smartphone s accelerometer sensor. A metallic ball tied with a spring-mass system and submerged in mustard oil undergoes damped oscillations due to viscous forces. The Phyphox app is used to record the temporal variation of acceleration, from which the damping constant is calculated to determine the coefficient of viscosity of the oil. The experimentally obtained value is further validated using the Tracker app, and this value is shown to be in close agreement with the standard literature. This method provides an accurate, low-cost experiment ideal for educational settings, utilizing smartphone sensors for viscosity measurement.