Clayton Cohn
Postdoctoral Associate, National Tutoring Observatory, Cornell University
About Me
Greetings! My name is Clayton Cohn. I am a Postdoctoral Associate in the National Tutoring Observatory within the Future of Learning Lab at Cornell University. I recently completed my PhD in Computer Science at Vanderbilt University, where my dissertation focused on designing LLM-powered conversational agents for STEM+C learning and assessment.
My research sits at the intersection of artificial intelligence in education (AIED), natural language processing (NLP), learning sciences, and educational data mining. I design, evaluate, and interpret AI/ML methods and agents for understanding and supporting teaching and learning in authentic educational settings.
My current work examines how tutor-student interaction data can be modeled to support scalable, interpretable misconception diagnosis and tutoring analytics.
I like to stay active, so I play hockey and golf and also lift weights. I'm a committed Francophile and enjoy improving my French language skills and learning more about French culture. I also enjoy volunteering, live music, and playing guitar.
More than anything, I like to challenge myself. I like to solve problems and learn new things. I like to engage in constructive discourse and hear opposing points of view. I like to broaden my horizons and expose myself to activities outside of my comfort zone.
Publications
Designing LLM-Powered Conversational Agents for STEM+C Learning and Assessment (Cohn, Doctoral Dissertation, Vanderbilt University)
in press
Evidence-Decision-Feedback: Theory-Driven Adaptive Scaffolding for LLM Agents (Cohn et al., AIED)
Jun. 2026
Jun. 2026
"He's Gonna Say It's Trash": Students' Emotional Experiences with AI Feedback in Collaborative Scientific Inquiry (Humburg et al., CSCL)
in press
Feeling like a Scientist, Engineer, and Computer Scientist: Exploring Students' Learning and Disciplinary Identity in an Integrated STEM+C Curriculum (Srivastava et al., ICLS)
in press
Quality of Collaboration in Computational Modeling: Implications for Middle-School STEM+C Learning (Jain et al., CSCL)
in press
Apr. 2026
Apr. 2026
Apr. 2026
A Theory of Adaptive Scaffolding for LLM-Based Pedagogical Agents (Cohn et al., AAAI)
Mar. 2026
Feb. 2026
Centering Youth Perspectives in Agent Design: Piloting an Activity to Scaffold Discussions on Ethics & Agents (Srivastava et al., ICCE Workshop)
Dec. 2025
Dec. 2025
Personalizing Student-Agent Interactions Using Log-Contextualized Retrieval Augmented Generation (RAG) (Cohn et al., AIED Workshop)
Jul. 2025
LearnLens: An AI-Enhanced Dashboard to Support Teachers in Open-Ended Classrooms (Srivastava and Jain et al., AIED Workshop)
Jul. 2025
Exploring Agentic Multimodal Late Fusion With LLMs for Embodied Learning (Fonteles and Cohn et al., EDM Workshop)
Jul. 2025
Eye movements as predictors of student experiences during nursing simulation learning events (Lee Mason et al., CRPI)
Jul. 2025
Exploring the Design of Pedagogical Agent Roles in Collaborative STEM+C Learning (Cohn et al., CSCL)
Jun. 2025
Jun. 2025
Comparing Text Augmentation by GPT-3.5 and Llama3 for Evaluating Student Responses (Cochran et al., IJAIED)
Apr. 2025
Sep. 2024
Towards A Human-in-the-Loop LLM Approach to Collaborative Discourse Analysis (Cohn et al., AIED)
Jul. 2024
Mar. 2024
Mar. 2024
Using BERT to Identify Causal Structure in Students' Scientific Explanations (Cochran et al., IJAIED)
Dec. 2023
Jul. 2023
Improving Automated Evaluation of Student Text Responses Using GPT-3.5 for Text Data Augmentation (Cochran et al., AIED) — Nominated for Best Student Paper
Jul. 2023
Jun. 2023
Improving NLP Model Performance on Small Educational Data Sets Using Self-Augmentation (Cochran et al., CSEDU)
Apr. 2023
Identifying Gaze Behavior Evolution via Temporal Fully-Weighted Scanpath Graphs (Davalos et al., LAK)
Mar. 2023
Using the DiCoT framework for Integrated Multimodal Analysis in Mixed-Reality Training Environments (Vatral et al., Frontiers in AI)
Jul. 2022
Improving Automated Evaluation of Formative Assessments with Text Data Augmentation (Cochran et al., AIED)
Jul. 2022
BERT Efficacy On Scientific and Medical Datasets: A Systematic Literature Review (Cohn, Master's Thesis, DePaul University)
Nov. 2020
Current Projects
National Tutoring Observatory: Sequential Misconception Diagnosis in Tutoring Interactions
Jun. 2026 - Present
- Investigating whether sequences of tutor and student moves provide stronger evidence of student misconceptions than isolated student utterances.
- Modeling misconception diagnosis as an interactional and temporal inference problem using tutor-student dialogue context.
National Tutoring Observatory: Fine-Tuning and Labeling Strategies for Misconception Detection
Jun. 2026 - Present
- Comparing outcome-labeled fine-tuning with richer multi-feature labeling to evaluate how supervision design affects model performance, interpretability, and pedagogical usefulness.
- Studying how label granularity, dialogue context, and tutoring-specific features shape the reliability and explanatory value of AI-based misconception detection.
Copa: A Multimodal Collaborative Peer Agent for STEM+C Learning
Sep. 2024 - Jun. 2026
- Designed and evaluated a collaborative “knowledgeable peer” agent integrating multimodal inputs, RAG, dialogue management, and learner modeling to deliver personalized support in STEM+C learning environments.
Education
Doctor of Philosophy, Computer Science from Vanderbilt University
Aug. 2021 - Jun. 2026
- Dissertation: Designing LLM-Powered Conversational Agents for STEM+C Learning and Assessment.
- Focused on NLP, LLMs, learner modeling, formative assessment, discourse analysis, and pedagogical agents for learning environments.
Master of Science, Computer Science from DePaul University
Jan. 2017 - Nov. 2020
- Concentration in Artificial Intelligence
- Graduated with Distinction
Bachelor of Science, Computer Science from DePaul University
Sep. 2013 - Nov. 2016
- Concentration in Software Development
- Cum Laude
Experience
Postdoctoral Associate at Cornell University
Jun. 2026 - Present
- Developing AI/ML methods for large-scale tutoring analytics in the National Tutoring Observatory within the Future of Learning Lab.
- Building NLP/LLM workflows for transforming tutor-student dialogue into interpretable learner evidence for tutoring research and AI-supported learning analytics.
- Modeling dialogue context, temporal interaction patterns, and tutoring-specific features to support scalable analysis of authentic tutoring sessions.
Research Assistant at Vanderbilt University
Jul. 2021 - Jun. 2026
- Researched and developed applied NLP and conversational AI methods for learning environments, including formative assessment scoring and feedback, discourse analysis, learner modeling, retrieval-augmented generation, and pedagogical agent design.
iOS Developer at The UpNext, Inc.
Jan. 2016 - Apr. 2017
- Designed, built, and launched iOS music discovery app.
Fellowships
University Graduate Fellow at Vanderbilt University
Aug. 2021 - Jun. 2026
"The University Graduate Fellowship is awarded by the Graduate School to highly qualified students entering study for the PhD or the master's degree."
Russell G. Hamilton Scholar at Vanderbilt University
Aug. 2021 - Jun. 2026
"Russell G. Hamilton scholarships are awarded to highly qualified students intending to pursue the PhD degree."
Volunteering
Dismas House in Nashville, Tennessee
- Organized meals for and provided guidance to halfway house residents transitioning out of prison.
Challah for Hunger at Vanderbilt University
- Participated in semi-monthly challah bakes to aid in fundraising for hunger-fighting organizations.