About

Peter Youngs is the Chair of the Department of Curriculum, Instruction and Special Education at the University of Virginia's School of Education and Human Development. He conducts research on the effects of educational policy and school social context on teaching and learning in the core academic subjects. Youngs has a special focus on the relationship between policy and practice in the areas of teacher education and induction. Currently, he is engaged in research on ways that neural networks can be used to automatically classify instructional activities in video. Additionally, Youngs serves as co-editor of the forthcoming second volume of the Handbook of Education Policy Research.

Research topics

• Sociology
• Mathematics education
• Psychology
• Pedagogy
• Medicine
• Political Science
• Social psychology
• Medical education

Selected publications

• Commentary

  Teacher Education and Special Education The Journal of the Teacher Education Division of the Council for Exceptional Children · 2026-01-11

  article1st authorCorresponding
  PublisherDOI

• Exploring Automated Recognition of Instructional Activity and Discourse from Multimodal Classroom Data

  2026-03-06

  articleOpen access

  Observation of classroom interactions can provide concrete feedback to teachers, but current methods rely on manual annotation, which is resource-intensive and hard to scale. This work explores AI-driven analysis of classroom recordings, focusing on multimodal instructional activity and discourse recognition as a foundation for actionable feedback. Using a densely annotated dataset of 164 hours of video and 68 lesson transcripts, we design parallel, modality-specific pipelines. For video, we evaluate zero-shot multimodal LLMs, fine-tuned vision–language models, and self-supervised video transformers on 24 activity labels. For transcripts, we fine-tune a transformer-based classifier with contextualized inputs and compare it against prompting-based LLMs on 19 discourse labels. To handle class imbalance and multi-label complexity, we apply per-label thresholding, context windows, and imbalance-aware loss functions. The results show that fine-tuned models consistently outperform prompting-based approaches, achieving macro-F1 scores of 0.577 for video and 0.460 for transcripts. These results demonstrate the feasibility of automated classroom analysis and establish a foundation for scalable teacher feedback systems.

  PublisherOA PDFDOI

• Contrasting Opportunities to Learn Across Domains in Special Education Teacher Preparation: A Mixed Methods Study

  Teachers College Record The Voice of Scholarship in Education · 2025-03-01 · 2 citations

  articleSenior author

  Background/Context: Special educators must be prepared to provide expert instruction and collaborate with a range of professionals to provide students with disabilities with equitable and inclusive educational opportunities. Teacher preparation is an important part of providing special education teacher candidates (SETCs) with robust opportunities to learn (OTL) both the instructional and collaborate aspects of their role. Purpose/Research Questions: In this paper, we describe SETCs’ OTL about their instructional and collaborative roles, addressing the following research questions: (1) To what extent do SETCs report OTL instructional and collaborative practices? (2) Are there significant differences in their experiences of OTL instructional and collaborative practices? (3) In what ways are SETCs’ reported OTL instructional practice distinct from their OTL collaborative practice? Research Design: We used a sequential mixed explanatory design to fulfill our research aim. In Phase 1, we drew on survey data from 154 SETCs across six traditional teacher preparation programs about their OTL instructional and collaborative practice. To explore differences identified in Phase 1 and to better understand SETCs’ OTL in teacher preparation, in Phase 2 we analyzed interview data from 20 SETCs from the survey sample. Coding was informed by Grossman and colleagues’ (2009) pedagogies of practice framework. Conclusions/Recommendations: Quantitative analyses revealed SETCs experienced more extensive OTL about instructional practice than collaborative practice. Qualitative analyses underscored that, whereas OTL instructional practice reflected the three elements of pedagogies of practice (i.e., representations, decompositions, and approximations of practice), OTL collaborative practice lacked structure and specificity. Especially given the struggles novice special educators report regarding collaboration in their first years in the field, findings offer valuable insight into the preparation of special educators for the complex, coordinated, and collaborative work necessary to support inclusive education and improved outcomes for students with disabilities.

  PublisherDOI

• How beginning elementary teachers’ mathematics instructional quality is associated with similarities between student teaching and first-year teaching assignments

  Teaching and Teacher Education · 2025-06-10

  article1st authorCorresponding
  PublisherDOI

• Causal State Space Model for Video Understanding

  IEEE Signal Processing Letters · 2025-01-01

  article

  We present a causal state space model (CSSM) for video understanding that couples a learned causal DAG with latent state dynamics. Latent factors form DAG nodes, enabling explicit cause–effect modeling over time; the state-space form provides efficient sequence inference, while the graph adds interpretability and robustness to distribution shifts. We learn the latent graph and inject its adjacency into the transition operator. On HMDB-51, UCF-101, and HAR, CSSM improves accuracy over strong baselines and supports counterfactual reasoning about video events.

  PublisherDOI

• Evidence of the indirect transmission of emotions from teachers to students in mathematics: The mediating role of instructional quality.

  School Psychology · 2025-07-21

  articleOpen accessSenior author

  = 443) self-reported mathematics emotions and engagement. Participants were recruited from 14 public elementary students in a single state in the Southwestern United States. Schools across this state varied considerably in schoolwide socioeconomic status and racial/ethnic makeup. Path models with cluster robust standard errors revealed an initial association between teachers' Time 1 (early fall) mathematics anxiety and their students' Time 3 (mid-winter) mathematics engagement, as well as two indirect effects of teachers' Time 2 (mid-fall) instructional quality on students' Time 3 outcomes: Instructional quality fully mediated the initial association between teachers' mathematics anxiety and students' mathematics engagement and played an indirect role in the association among teachers' mathematics anxiety and students' mathematics enjoyment. Effect sizes were small, ranging from .03 to .04. Results can inform efforts by education researchers and practitioners to incorporate foci on emotions in future research and systems of teacher and student support. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

  PublisherDOI

• A dynamic fractional generalized deterministic annealing for rapid convergence in deep learning optimization

  npj Artificial Intelligence · 2025-10-01

  articleOpen access

  Optimization is central to classical and modern machine learning. This paper introduces Dynamic Fractional Generalized Deterministic Annealing (DF-GDA), a physics-inspired algorithm that boosts stability and speeds convergence across a wide range of models, especially deep networks. Unlike traditional methods such as Stochastic Gradient Descent, which may converge slowly or become trapped in local minima, DF-GDA employs an adaptive, temperature-controlled schedule that balances global exploration with precise refinement. Its dynamic fractional-parameter update selectively optimizes model components, improving computational efficiency. The method excels on high-dimensional tasks, including image classification, and also strengthens simpler classical models by reducing local-minimum risk and increasing robustness to noisy data. Extensive experiments on sixteen large, interdisciplinary datasets, including image classification, natural language processing, healthcare, and biology, show that DF-GDA consistently outperforms both state-of-the-art and traditional optimizers in convergence speed and accuracy, offering a powerful alternative for critical large-scale, complex problems across diverse scientific and industrial settings today.

  PublisherOA PDFDOI

• A dynamic predictive transformer with temporal relevance regression for action detection

  Pattern Recognition · 2025-04-14 · 4 citations

  article
  PublisherDOI

• Semanticbox: Bounding Box-Guided Caption Enhanced Action Recognition for Instructional Videos

  2025-09-14

  article

  Multimodal action recognition within complex scenes requires a comprehensive understanding of the entire scene, encompassing both the visual and audio aspects of the video. Contrastive Learning Image Pretraining (CLIP) is a well-known backbone for multi-modal action recognition tasks as seen in ActionCLIP and its variants. However, these models are subject to a major weakness: overemphasis on the background. SemanticBox integrates bounding boxes into the video action recognition CLIP-style paradigm to add visual clues that boost the model’s classification performance. Additionally, a pretrained generative classifier is added to provide rich frame descriptions, enhancing the textual feature semantics and offering an additional performance boost. SemanticBox achieves impressive performance on a complex instructional video dataset characterized by background clutter, achieving comparable Recall@2 to state-of-the-art CLIP-based models and outperforming them in Top-1 and Top-2 accuracy, F1 score, and mean average precision (mAP).

  PublisherDOI

• Teacher Preparation:

  2025-07-17

  book-chapter1st authorCorresponding
  PublisherDOI

Recent grants

• Using Neural Networks for Automated Classification of Elementary Mathematics Instructional Activities

  NSF · $1.5M · 2020–2025

• A Study of Elements of Teacher Preparation Programs that Interact with Candidates' Characteristics to Support Novice Elementary Teachers to Enact Ambitious Mathematics Instruction

  NSF · $1.8M · 2015–2020

Frequent coauthors

• Scott T. Acton

  University of Virginia

  31 shared

• Matthew Korban

  University of Virginia

  29 shared

• G. S. Watson

  Old Dominion University

  22 shared

• Jonathan B. Foster

  Old Dominion University

  16 shared

• Scout Crimmins

  University of Virginia

  16 shared

• Jihyun Kim

  10 shared

• Kenneth A. Frank

  Michigan State University

  10 shared

• Madeline Mavrogordato

  Michigan State University

  8 shared

Labs

• EdPolicyWorksPI

Education

• PhD, Educational Policy Studies

  University of Wisconsin-Madison

  2003