Breadcrumb
- Home
- AI Hub For Education
- Research Study Repository
- Evidence Explorer
- AI Tools
- Custom-built Research Tool
The Repository
Search and Filter
Custom-built research tool
Studies tagged with AI from their full text or abstract; see each study for details.
Showing 481 - 510 of 531 studies
Encouraging Responsible Use of Generative AI in Education: A Reward-Based Learning Approach
Aditi Singh, Abul Ehtesham, Saket Kumar, Gaurav Gupta, Tala Talaei Khoei. (06/2024). arXiv.
Text-based math inputs achieved higher accuracy across all categories than image-based inputs, reaching 85% accuracy for linear and quadratic equations.
Knowledge Distillation of LLMs for Automatic Scoring of Science Assessments
Ehsan Latif, Luyang Fang, Ping Ma, Xiaoming Zhai. (06/2024). arXiv.
The distilled student models scored responses with 2% to 3% higher accuracy than baseline neural networks and TinyBERT across the science and math datasets.
Beyond Answers: Large Language Model-Powered Tutoring System in Physics Education for Deep Learning and Precise Understanding
Zhoumingju Jiang, Mengjun Jiang. (06/2024). arXiv.
Physics-STAR increased students' average scores on complex information-based physics problems by 100% and increased their problem-solving efficiency by 5.95%.
AI AGENTS AND EDUCATION: SIMULATED PRACTICE AT SCALE
Ethan Mollick, Lilach Mollick, Natalie Bach, LJ Ciccarelli, Ben Przystanski, Daniel Ravipinto. (06/2024). arXiv.
A multi-agent architecture coordinates distinct AI roles to deliver direct instruction, adapt simulated investor practice, and generate student feedback alongside instructor summaries.
Ink and Algorithm: Exploring Temporal Dynamics in Human-AI Collaborative Writing
Kaixun Yang, Yixin Cheng, Linxuan Zhao, Mladen Raković, Zachari Swiecki, Dragan Gašević, Guanliang Chen. (06/2024). arXiv.
Writers clustered into four distinct temporal patterns of generative AI usage across their writing sessions.
Math Multiple Choice Question Generation via Human-Large Language Model Collaboration
Jaewook Lee, Digory Smith, Simon Woodhead, Andrew Lan. (05/2024). arXiv.
Educators deemed 70% of the question stems, keys, and explanations generated by GPT-4 as valid.
An Automatic Question Usability Evaluation Toolkit
Steven Moore, Eamon Costello, Huy A. Nguyen, John Stamper. (05/2024). arXiv.
SAQUET achieved a 94.13% accuracy rate across individual flaw classifications and agreed with human experts on overall question acceptability for 75.3% of items.
Large Language Models for In-Context Student Modeling: Synthesizing Student's Behavior in Visual Programming
Manh Hung Nguyen, Sebastian Tschiatschek, Adish Singla. (05/2024). arXiv.
A fine-tuned GPT-3.5 model matched human tutor performance on HoCMaze-18 and achieved overall quality scores of 0.64 on HoCMaze-4 and 0.82 on HoCMaze-18.
JiuZhang3.0: Efficiently Improving Mathematical Reasoning by Training Small Data Synthesis Models
Kun Zhou, Beichen Zhang, Jiapeng Wang, Zhipeng Chen, Wayne Xin Zhao, Jing Sha, Zhichao Sheng, Shijin Wang, Ji-Rong Wen. (05/2024). arXiv.
JiuZhang3.0 models achieved state-of-the-art results among open-source models on math benchmarks, including 52.8% on MATH and 89.8% on GSM8k.
Towards Educator-Driven Tutor Authoring: Generative AI Approaches for Creating Intelligent Tutor Interfaces
Tommaso Calò, Christopher J. MacLellan. (05/2024). arXiv.
AI assistance reduced the time required to build a complex arithmetic tutor interface by 68% and a simple sequential interface by 23%.
The AI Collaborator: Bridging Human-Ai Interaction In Educational And Professional Settings
Mohammad Amin Samadi, Spencer JaQuay, Nia Nixon, Jing Gu. (05/2024). arXiv.
The AI Collaborator system integrates Slack, a dedicated database, and GPT-4 to enable customizable conversational personas for team research.
Grade Like a Human: Rethinking Automated Assessment with Large Language Models
Wenjing Xie, Juxin Niu, Chun Jason Xue, Nan Guan. (05/2024). arXiv.
Fine-grained rubrics generated with score-distribution-aware sampling improved grading accuracy on complex operating system questions, reducing error and increasing correlation with human graders.
Desirable Characteristics for AI Teaching Assistants in Programming Education
Paul Denny, Stephen MacNeil, Jaromir Savelka, Leo Porter, Andrew Luxton-Reilly. (05/2024). arXiv.
A large majority of students agreed or strongly agreed that the AI assistant's responses for code writing and debugging tasks were both correct and helpful.
GPTutor: Great Personalized Tutor with Large Language Models for Personalized Learning Content Generation
Eason Chen, Jionghao Lin, Jia-En Lee, Kenneth R. Koedinger. (05/2024). arXiv.
GPTutor uses large language models and a serverless framework to generate personalized educational content and exercises from educator-uploaded curriculum and student interests.
Effective and Scalable Math Support: Experimental Evidence on the Impact of an AI- Math Tutor in Ghana
Owen Henkel, Hannah Horne-Robinson, Nessie Kozhakhmetova, Amanda Lee. (05/2024). arXiv.
Students receiving extra practice with the AI tutor achieved significantly higher math growth than control students, with an effect size of 0.36.
FOKE: A Personalized And Explainable Education Framework Integrating Foundation Models, Knowledge Graphs, And Prompt Engineering
Silan Hu, Xiaoning Wang. (05/2024). arXiv.
The FOKE framework integrates a hierarchical knowledge forest, multi-dimensional user profiling, and structured prompt engineering to generate tailored and explainable educational guidance.
App Planner: Utilizing Generative AI in K-12 Mobile App Development Education
David Kim, Prerna Ravi, Randi Williams, Daeun Yoo. (04/2024). arXiv.
High school students strongly preferred using App Planner over designing independently, reporting that it streamlined brainstorming, clarified subsequent steps, and delivered practical technical tips.
Exploring How Multiple Levels of GPT-Generated Programming Hints Support or Disappoint Novices
Ruiwei Xiao, Xinying Hou, John Stamper. (04/2024). arXiv.
Worked-example code hints most frequently supported success, leading to correct code modifications in 59.32% of instances.
iScore: Visual Analytics for Interpreting How Language Models Automatically Score Summaries
Adam Coscia, Langdon Holmes, Wesley Morris, Joon Suh Choi, Scott Crossley, Alex Endert. (03/2024). arXiv.
Interacting with the visual analytics tool led one learning engineer to improve their language model's scoring accuracy by three percentage points.
CuentosIE: can a chatbot about "tales with a message" help to teach emotional intelligence?
Antonio Ferrandez, Rocío Lavigne-Cervan, Jesus Peral, Ignasi Navarro-Soria, Angel Lloret, David Gil, Carmen Rocamora. (03/2024). arXiv.
Users reported an overall satisfaction score of 7.82 out of 10 for CuentosIE and rated their emotional intelligence improvement at 6.66.
Teach AI How to Code: Using Large Language Models as Teachable Agents for Programming Education
Hyoungwook Jin, Seonghee Lee, Hyungyu Shin, Juho Kim. (03/2024). arXiv.
AlgoBo's question-asking mode significantly increased the density of knowledge-building messages in dialogues during problem-solving, with an effect size of 0.71.
Does Feedback on Talk Time Increase Student Engagement? Evidence from a Randomized Controlled Trial on a Math Tutoring Platform
Dorottya Demszky, Rose Wang, Sean Geraghty, Carol Yu. (03/2024). ACM Digital Library.
The Talk Meter increased student talk ratios by 13% to 14% across both tutor-only and tutor-and-student feedback conditions.
Mathemyths: Leveraging Large Language Models to Teach Mathematical Language through Child-AI Co-Creative Storytelling
Chao Zhang, Xuechen Liu, Katherine Ziska, Soobin Jeon, Chi-Lin Yu, Ying Xu. (02/2024). arXiv.
Children showed significant improvements in overall math language scores, with learning gains from interacting with Mathemyths appearing comparable to those with a human partner.
Edu-ConvoKit An Open-Source Library for Education Conversation Data
Rose Wang, Dorottya Demszky. (02/2024). arXiv.
Edu-ConvoKit provides a unified, modular Python pipeline using dataframes to standardize the pre-processing, annotation, and analysis of educational dialogue transcripts.
A Piece of Theatre: Investigating How Teachers Design LLM Chatbots to Assist Adolescent Cyberbullying Education
Michael A. Hedderich, Natalie N. Bazarova, Wenting Zou, Ryun Shim, Xinda Ma, Qian Yang. (02/2024). arXiv.
Teachers approached chatbot design as playwrights, creating multi-participant role-play scenarios that allowed students to rehearse and explore prosocial bystander behaviors in safe environments.
Are LLMs Useful in the Poorest Schools? TheTeacher.AI in Sierra Leone
Jun Ho Choi, Oliver Garrod, Paul Atherton, Andrew Joyce-Gibbons, Daniel Björkegren, Miriam Mason-Sesay. (02/2024). arXiv.
Nearly half of all queries focused on clarifying concepts (48%), followed by lesson planning (21%), writing support (7%), professional development (6%), and classroom management (5%).
Learning Programming of Agent-based Modeling with LLM Companions: Experiences of Novices and Experts Using ChatGPT & NetLogo Chat
John Chen, Xi Lu, David Du, Michael Rejtig, Ruth Bagley, Mike Horn, Uri J. Wilensky. (01/2024). arXiv.
Experts reported more perceived benefits than novices and predominantly intended to continue using language model interfaces, whereas frustrated novices sought alternative learning resources.
Interactions with Prompt Problems: A New Way to Teach Programming with Large Language Models
James Prather, Paul Denny, Juho Leinonen, David H. Smith IV, Brent N. Reeves, Stephen MacNeil, Brett A. Becker, Andrew Luxton-Reilly, Thezyrie Amarouche, Bailey Kimmel. (01/2024). arXiv.
Students averaged 4.3 submissions per problem, often iterating by modifying a few words rather than lengthening prompts, and occasionally experimenting further after solving tasks.
Challenges and Opportunities of Moderating Usage of Large Language Models in Education
Lars Krupp, Steffen Steinert, Maximilian Kiefer-Emmanouilidis, Karina E. Avila, Paul Lukowicz, Jochen Kuhn, Stefan Küchemann, Jakob Karolus. (12/2023). arXiv.
Moderated chatbots reduced unreflected behavior like copy-pasting to 21% and 23% of queries, compared to 47% with an unmoderated chatbot.
White Paper: The Generative Education (GenEd) Framework Empowering Educators as AI-Enhanced Mentors
Daniel Leiker. (11/2023). arXiv.
The Generative Education framework transitions educators into AI-enhanced mentors who supply training materials and guidance while AI companions provide scalable, personalized tutoring.
