United States

Machine learning algorithms represent an exciting new class of quantitative methods to understand physics classes and students. Recent work has applied these algorithms to classify students as those likely to receive an A or B or students were likely to receive an C. D, F in a physics class. The metrics used become unreliable when the outcome variable is substantially unbalanced. This talk further explores the classification and extends those methods to predicting whether a student will receive a D or F. This study also investigated the previous finding that demographic variables such as gender, underrepresented minority status, and first-generation status had low variable importance for predicting class outcomes; down sampling revealed that this result was not due to the underrepresentation of these students in the class studied.

AAPT SM 20

Extending Machine Learning to Predict Unbalanced Physics Course Outcome

at-risk students

prediction

machine learning

Naval Information Warfare Center Pacific is partnering with AFCEA San Diego chapter to host the workshop.

The NAML workshop is an opportunity for researchers from Navy research and development centers and our partner organizations to exchange technical ideas and to come together as a team.

The goals of NAML are:

* To spread awareness of current machine learning research relevant to Naval applications

* To connect machine learning researchers with experts in military requirements

* To build and strengthen collaborations within the defense research community

**Watch this video to hear about the AFCEA purpose.**

<iframe width=1280 height=706 frameborder="0" scrolling="no" src="https://screencast-o-matic.com/player/creleVVcvRh?v=6&ff=1&title=0&controls=1" allowfullscreen="true"></iframe>

NAML 2021

technical paper

The 2020 AAPT Virtual Summer Meeting brought together physics educators from around the globe to learn about the latest in physics theory, research methodology and education from leaders in their field. The program featured plenary speakers, sessions, workshops and many special activities.

For more information please visit the 2020 AAPT Virtual Summer Meeting page at: https://www.aapt.org/Conferences/SM2020/index.cfm.


The 2020 AAPT Virtual Summer Meeting brought together physics educators from around the globe to learn about the latest in physics theory, research methodology and education from leaders in their field. 

This work examines peer-quizzing with metacognition in calculus based introductory physics. Students take individual quizzes and then redo part of the quiz with a partner. Because the students to choose which problems to redo, they must reflect on their confidence level on different concepts (questions). The results show there is a clear benefit to introducing weekly quizzes. While peer quizzes do not result in a clear improvement in the final concept assessment score, there is improvement in the normalized gain. Further, there is a clear signal that students who score above average on their quiz redos will also show above average assessment gains. These results show that while the intervention does not improve the overall grade, it helps individual students at all grade levels improve their understanding in relationship to their peers. It also suggests that interventions for students with low redo scores will have significant impact.

Utilizing Peer-Instruction and Metacognition on Quizzes to Improve Concept Learning

Science education studies framed around funds of knowledge, mental models and student resource framework appear to have three general characteristics: they are learner centered, tend to deviate from the deficit student model of teaching that present learners as empty vessels, but rather present learners as individual beings with different sets of skills and needs. However, there is not a consensus on the ways, as far as we know, in which these three concepts relate with each other. We conduct a systematic inquiry of the literature available for how funds of knowledge, mental models, and resource framework are respectively used in science education, and present a scheme relating these three frameworks that can be utilized when evaluating students’ ideas about physical concepts such as energy. Considering how those frameworks connect with one another might provide additional avenues to further explore how students' conceptual understanding of the physical world are formed.

Characterizing Conceptual Understanding in Introductory Physics using Funds of Knowledge, Mental Models and Resource Framework

You and/or your department are interested in introducing computational modeling into your introductory physics courses, but want some concrete ideas of how to accomplish this. You want it to be student friendly; i.e., not requiring much prior student programming experience. You want it to be instructor friendly; i.e., not requiring huge time investment for the faculty to implement. You want it to be affordable; i.e., not requiring a huge investment by the department, faculty or student. This presentation will illustrate a way to accomplish introducing computational modeling into introductory physics courses.

Computational Physics in Introductory Physics

As a first step in a larger study of student difficulties in upper-division thermodynamics and statistical mechanics, we present the results of think-aloud interviews with graduate students on a set of entropy related questions. The four interview questions were developed to probe student understanding of entropy as a pressure towards equilibrium, as a quantity maximized in equilibrium, as a connection between microstates and macrostates, and as a macroscopic state-function. Exploring graduate students' understanding entropy and their ability to solve problems and reason with entropic arguments will provide insights into how physicists develop a mature understanding of entropy as a physical quantity. Specifically, we hope to see if new conceptual difficulties emerge as students progress to graduate school, and whether difficulties seen in undergraduate courses persist, evolve, or cease to present issues in the graduate setting.

Investigating how graduate students connect microstates and macrostates with entropy

Computers and algorithms have fundamentally reshaped our world in the last few decades. Data is the new oil. NGSS has codified computational thinking as a science practice. Yet programming has yet to penetrate compulsory K-12 schooling in a meaningful way. A number of high school Modeling Physics teachers have been teaching their students to program for two decades. Although teaching programming basics cut into the finite number of minutes they would otherwise have devoted to physics concepts, they valued the trade-off—equipping students with a tool for visualizing abstractions (e.g., EM fields), and for more tightly connecting mathematical, graphical and diagrammatic representations of conceptual models. Five years ago AMTA and AAPT joined forces to adapt HS physics teachers’ pioneering work in computing for use in 9th grade Physics First classrooms. This is a report on where we are presently and what we have learned along the way.

Computational Modeling Physics First: A Progress Report

The usage of toys to teach physics is an interesting approach to promote engagement and creativity. Traditionally, toys have been widely used in qualitative demonstrations. However, it is frequently difficult to extract quantitative results in physics experiments involving toys. One possible strategy to address this difficulty is the use of smartphone sensors. In this work we investigate the dynamics of a traditional toy, the yoyo, theoretically and experimentally, using smartphone’ sensors. In particular, using the gyroscope, the angular velocity was measured. The experimental results were complemented thanks to a digital video analysis. As the yoyo is a ubiquitous, simple and traditional toy this simple proposal could encourage students to experiment with everyday objects and modern technologies. More information: http://smarterphysics.blogspot.com/

The physics of a yoyo using a smartphone

Physics Resources and Instructional Strategies for Motivating Students (PRISMS) is a high school physics curriculum resource that has been used extensively in UNI teacher preparation and professional development. The original PRISMS materials were a collection of 130 high interest activities related to the real-life experiences of high school physics students designed to develop student conceptual understanding and to cultivate student scientific reasoning and problem skills. The enhanced and revised version, called PRISMS PLUS, focuses on complete learning cycles that provide fully integrated experiences that enable students to develop not only their problem solving and inquiry skills but also deep, long-lasting understanding of physics concepts. This presentation will focus on how PER has been used to develop PRISMS PLUS and adapt it to teacher preparation and professional development including alignment with the Next Generation Science Standards.

Using PER to Develop and Adapt PRISMS PLUS

Quantum mechanics is the centerpiece of modern physics. Teaching quantum mechanics have been difficult because of complex numbers, Hilbert space, and interpretations. We show that a real formulation of quantum mechanics without complex numbers and Hilbert space is included in Schrödinger’s 1926 papers. The legitimacy and advantages of such a real formulation were justified in 1960s. High-caliber quantum mechanics textbooks in real formulation existed since 1970s in France. Experimental observations of atomic and molecular wavefunctions using scanning tunneling microscopy necessitates a real formulation. Practically, in chemical physics and condensed-matter physics, real wavefunctions are sufficient, especially regarding to density-functional theory (DFT). Outlined here is a freshman-sophomore course of real quantum mechanics covering a wide range of applications to dispense with complex numbers and Hilbert space. As shown, the derivation of the real Schrödinger equations is transparent, the mathematics is elementary, and the interpretation is intuitive.

Teach intuitive and practical quantum mechanics with real Schrödinger equations

Addressing complex societal problems requires STEM workers with expertise such that they are able to recognize, understand, and effectively reason out solutions using principles of their disciplines. This need emphasizes the importance of skills-based learning. This presentation will provide details about the development of an introductory physics lab curriculum, which was designed around the theory-evidence coordination (TEC) framework to advance select reasoning skills through scaffolded lab activities. Best practices for using the TEC framework to modify existing curriculum will be provided along with how one might assess subsequent impact on student development of scientific reasoning. We will also discuss our interest in pilot sites for our curriculum.

Transforming Lab Instruction to Promote Essential Scientific Reasoning Abilities

In the undergraduate lab experience, computation and experimentation are typically presented as distinct lab experiences. This approach ignores the important interplay that can exist between these two different ways of understanding the physical world. At Carleton College, we developed one two-week lab that treats computation and experimentation as interconnected, equal partners in students’ understanding of the natural world. I will discuss this lab that focuses on finding the half-lives of two simultaneously decaying isotopes and discuss how we encourage students to think about the interplay between computation and experimentation. Additionally, I will discuss efforts to promote lab skill development and inclusion broadly in this sophomore-level lab course and particularly how these inclusion efforts are manifested in the lab I will present in this talk.

Premium content

Extending Machine Learning to Predict Unbalanced Physics Course Outcome

Downloads

Next from AAPT SM 20

Utilizing Peer-Instruction and Metacognition on Quizzes to Improve Concept Learning

Similar lecture

Machine Learning in the Prediction of Displacement Hull Form Resistance