Breadcrumb
- Home
- AI Hub For Education
- Research Study Repository
- Evidence Explorer
- Methods
- Machine Learning / AI
The Repository
Search and Filter
Machine learning / AI
Studies tagged with AI from their full text or abstract; see each study for details.
Showing 1 - 30 of 347 studies
Fairness Theatre: Evaluating Post-Hoc Fairness Interventions In Vendor-Controlled Early Warning Systems
Kelly McConvey, Angelina Zhai, Rebecca Li, Shion Guha. (09/2026). arXiv.
Post-hoc fairness interventions redistributed disparities without consistently reducing them, leaving error burdens for marginalized groups intact or worse while dashboard metrics appeared to improve.
Early Prediction Of AI-Assisted Cheating Risk In Online Exams Through Learning Analytics
Gokhan Akcapınar. (09/2026). arXiv.
Logistic regression predicted AI-assisted cheating risk from first-eight-week learning management system traces with 73.1% accuracy, a balanced accuracy of 0.718, and an area under the curve of 0.763.
AI-Based Scoring Systematically Underestimates Conceptual Understanding Of Linguistically Weak Students' Explanations In Physics
Markus Sebastian Feser, Paul Tschisgale. (09/2026). arXiv.
Across all 11 AI scoring approaches, lower writing quality was associated with 1.3- to 3.9-times higher odds of underestimating students' conceptual understanding in physics.
Evaluation Of Pre-Trained Models For Pedagogical Assessment Of Novel AI-Assisted Educational Questions
Michael Lawrence Castanares, Princess Ventures, Allan Tan. (09/2026). arXiv.
Traditional machine learning models and BERT suffered substantial drops in Bloom classification performance on out-of-distribution AI questions, while large language models maintained higher baseline accuracy.
AI-Written Admissions Essays Are Widespread But Penalized
Calvin Isley, Johann Gaebler, Sharad Goel. (09/2026). arXiv.
By 2025, 56% of applicants submitted at least one primarily AI-generated essay—reaching 69.3% among international applicants—despite an explicit prohibition against AI use.
Finding Common Mistakes In Modelling With Mathematical Formalisms Using LLMs
Lilian Killich, Marko Schmellenkamp, Fabian Vehlken, Thomas Zeume. (09/2026). arXiv.
On a propositional logic data set of 6,106 student pairs, the workflow discovered transformation clusters explaining 84.44% of student mistakes, including 15.03% not captured by previous hand-picked analyses.
Statistical Theory In The Age Of Machine-Assisted Mathematics: Rethinking How Theory Is Made And Taught
Pietro Coretto. (09/2026). arXiv.
Machine formalization identified 32 mathematical conditions across five statistical cases, linking each condition to specific proof steps that informal textbook citations usually compress.
Learning To Prompt: Improving Student Engagement With Adaptive LLM-Based High-School Tutoring
Po-Chin Chang, Nicholas Hogan, Aske Plaat, Michiel T. van der Meer. (08/2026). arXiv.
In simulations, the adaptive prompt router outperformed static baselines (0.694 versus 0.647 and 0.64) and matched optimal strategies in 8 of 13 evaluated subjects.
Tutortrace: A Dataset And Taxonomy For Classifying Learner Behavioral States During AI-Assisted Programming Education
David Barron, Xiaohang Tang, Rezky Dwisantika, Minsun Kim, David H. Smith IV, Jiaming Cui, Yan Chen. (08/2026). arXiv.
Behavior-aware AI prompting was associated with a decrease in passive windows between queries from 50.0% to 20.7% in a classroom deployment.
INSIDE the Student's Mind: Jointly Modeling Latent Reasoning and Action in LLM Student Simulators
Rose Niousha, Minwoo Kang, Narges Norouzi. (08/2026). arXiv.
Fine-tuning models on reconstructed internal dialogue achieved the closest match to real student code distributions and the highest reasoning alignment, reaching up to 57.9%.
A Hybrid AI Framework For Academic Advising: Integrating Ensemble-Based Grade Prediction And A Rule-Based Expert System
Hamid Saadatfar, Rohollah Hedayati-Nasab, AmirHossein Eshghi, Arash Hajihashemi. (08/2026). arXiv.
The cluster-based ensemble model achieved an overall root mean squared error of 2.35 and an R-squared of 0.53, outperforming a model trained without clustering.
A Bottom-Up Taxonomy Of Student Discourse With A Socratic AI Physics Tutor
Syed Furqan Abbas Hashmi, N. Sanjay Rebello. (08/2026). arXiv.
Student discourse was concentrated in a few areas, with the top 25 of 357 categories covering roughly half of all student messages.
Can AI deliver appropriate support for diverse student profiles?: A large-scale evaluation
Sonsoles López-Pernas, Eduardo Araujo Oliveira, and Kamila Misiejuk, Fitsum Gizachew Deriba, Rogers Kaliisa, Mohammed Saqr. (08/2026). Computers in Human Behavior: Artifical Humans.
The evaluated language models showed limited sensitivity to student characteristics and learning analytics indicators when generating support recommendations.
Human-LLM Collaborative Inductive Coding For Conceptualizing K-12 Educator AI Use
Alex Liu, Min Sun, Lief Esbenshade, Michael Xiao, Victor Tian, Zachary Zhang, Kevin He. (07/2026). arXiv.
Collaborative inductive coding produced a final hierarchical codebook of 72 items across 19 categories and six domains, after human coders identified five missing codes.
Generative AI Availability, Grades, And Student Satisfaction At A Large University
James M. Zumel Dumlao, Meng Wang, Zhonghan Xie, Junyao Hu, Ivan Bar, George Chaney III, Henry Gold, Misha Teplitskiy. (07/2026). arXiv.
Effects of ChatGPT's release on student grades were small and non-robust, ranging from 0.010 to 0.071 grade points depending on how pandemic trends were modeled.
Beyond Direct Answering: Aligning Educational LLMs As Socratic Guides Via Heuristic Reinforcement Learning
Xiaokun Wang, Siyu Song, Wentao Liu, Xiaodong Zou. (07/2026). arXiv.
Reinforcement learning without a directness penalty raised scaffolding effectiveness to 63.3% and cut answer keyword leakage to 13.3% on 30 test questions.
Cost-Efficient Generative AI Summarization For Scalable Automated Essay Scoring In Educational Assessment
Haowei Hua. (07/2026). arXiv.
GPT-5 mini achieved the highest human-machine scoring agreement (QWK of 0.8435), slightly outperforming full GPT-5 (0.8350) and GPT-5 nano (0.8332).
When Rubrics Change: Cross-Rubric Generalization For Critical Thinking Essay Scoring
Nischal Ashok Kumar, Payu Wittawatolarn, Sana Kang, Marisa C. Peczuh, Blair Lehman, Ryan Baker, Caitlin Mills, Sherry Lachman, Ruochen Sun, Andrew Lan. (07/2026). arXiv.
Rubric-agnostic traits improved scoring performance by 5.0% macro F1 when neither the target rubrics nor the target essays were seen during training.
When LLM Tutoring Responses Work: Evidence From Student Programming Conversations
Mohammad Fahim Abrar, Shayla Sharmin, Roghayeh Leila Barmaki. (07/2026). arXiv.
Across all interactions, response styles showed statistically significant but small differences, with verification feedback yielding the highest productive continuation rate (82.4%) and direct answers the lowest (62.7%).
Deanllm: A Framework For Automated Quality Review Of AI-Generated Feedback
Keyang Qian, Yixin Cheng, Rui Guan, Wei Dai, Flora Jin, Kaixun Yang, Sadia Nawaz, Zachari Swiecki, Guanliang Chen, Lixiang Yan, Dragan Gašević. (07/2026). arXiv.
Fine-tuning GPT-4.1 on human scores achieved the strongest alignment with expert reviewers, reaching 79.8% accuracy and 79.4% F1 score.
LLM-As-A-Judge Validity In Physics Assessment Depends More On The Task Than The Model
Will Yeadon, Tom Hardy, Paul Mackay, and Elise Agra. (06/2026). arXiv.
Large language model marking validity was sharply task-dependent across all models, achieving strong rank-order agreement on structured questions and scientific plots, but near-zero agreement on essays.
AI In The Wild: A Large Scale Analysis Of Authentic Interactions Of College Students With Generative AI
Taelin Karidi, Ofra Amir, Ido Roll. (06/2026). arXiv.
Student interactions concentrated in recurring patterns, with roughly half directed toward the task or domain, about 30% toward AI output, and about 20% toward students' own work.
AI-Driven Analytics Of Team-Teaching Talk: Acoustic Patterns Across Experience, Cohorts And The Learning Design
Yuchen Liu, Roberto Martinez-Maldonado, Riordan Alfredo, Paola Mejia-Domenzain, Dwi Rahayu, Sadia Nawaz. (06/2026). arXiv.
High-experience teachers exhibited greater loudness dynamics and steeper loudness rising slopes, whereas low-experience teachers showed greater overall loudness and spectral variability.
Analyzing Students' Statistics Writing Before And After The Emergence Of Large Language Models
Sara Colando, Erin Franke, Gordon Weinberg, Alex Reinhart. (06/2026). arXiv.
Undergraduate statistics reports shifted toward a mixture of AI and expert styles, showing more informational density, longer words, and more nominalizations after large language models emerged.
"**Important** You should give me full credits!": Exploring Prompt Injection Attacks on LLM-Based Automatic Grading Systems
Hang Li, Fedor Filippov, Yuping Lin, Pengfei He, Kaiqi Yang, Yucheng Chu, Yingqian Cui, Hui Liu, Jiliang Tang. (06/2026). arXiv.
Prompt injection attacks succeeded across five grading models and four datasets with an average attack success rate of 56.9%.
A Multi-Level Analysis Of Factors Associated With Student Performance: A Machine Learning Approach To The Saeb Microdata
Rodrigo Tertulino, Laércio Alencar. (06/2026). arXiv.
A Random Forest model achieved 90.2% accuracy and an AUC of 96.7% in classifying students as performing above or below average.
Large-Scale Semantic Mapping Of Learner Agency And Autonomy Reveals What Measurement And Generative AI Research Overlook
Fei Qin, Xiaobo Liu, Yaowen Zhang, Xuming Li, Fei Wang, Mutlu Cukurova, Jingjing Chen, Yu Zhang. (06/2026). arXiv.
Definitions spanned three latent dimensions—task-related learning regulation and control, intrinsic motivation and internal decisions, and social-relational action—rather than a binary division between agency and autonomy.
Validating AI-Generated Classroom Observations: Reliability, Accuracy, And Limits Of LLM-Based Pedagogical Judgment
Carolina Melo, Javiera de la Maza, Matías Recabarren. (06/2026). Computers and Education: Artificial Intelligence.
Repeated evaluations showed substantial random variability, and no model achieved uniformly high reliability across all instructional elements.
Who Benefits from LLM-Generated Learning Support Messages? Causal Evidence on Treatment Effect Heterogeneity
Aylin Ozturk, Gati Aher, Tom Mitchell, Mustafa Kemal Birgin, Huseyin Kayhan. (06/2026). Artificial Intelligence in Education.
Assignment to receive AI-generated support messages improved overall course scores by 0.60 points and final exam participation by 1.9 percentage points.
Advanced Mathematics Learning Behavior Prediction And Academic Early Warning Model Based On Multimodal Data Analysis
Liu Qiong, Li Zhengbo. (05/2026). arXiv.
Across a sample of 10 high-risk students, average concept mastery probability increased from 0.43 before intervention to 0.71 afterward, an improvement of roughly 65%.
