A. Cagla Acun

A. Cagla Acun

Assistant Professor in Artificial Intelligence and Data Engineering at Hacettepe University

Posts by Collection

portfolio

publications

A data science pipeline for educational data: A case study using learning catalytics in the active learning classroom

Published in M.S. thesis, University of Louisville, 2017

Master thesis developing a data science pipeline for classroom response system data collected in an active learning classroom.

Recommended citation: Acun Sener, A. C. (2017). "A data science pipeline for educational data: A case study using learning catalytics in the active learning classroom." M.S. thesis, University of Louisville. https://ir.library.louisville.edu/etd/2758/

Using a Data Science Pipeline for Course Data: A Case Study Analyzing Heterogeneous Student Data in Two Flipped Classes

Published in ASEE Annual Conference and Exposition, 2019

A four-stage data science pipeline applied to classroom response system, homework, exam and demographic data from two flipped engineering-mathematics sections.

Recommended citation: Acun Sener, A. C., Hieb, J. L., & Nasraoui, O. (2019). "Using a Data Science Pipeline for Course Data: A Case Study Analyzing Heterogeneous Student Data in Two Flipped Classes." ASEE Annual Conference and Exposition. https://peer.asee.org/using-a-data-science-pipeline-for-course-data-a-case-study-analyzing-heterogeneous-student-data-in-two-flipped-classes

A Data Science Approach to Flagging Non-Retention in Engineering Enrollment Data

Published in 2020 ASEE Virtual Annual Conference, 2020

Proposes an enrollment-gap criterion, three consecutive semesters not enrolled, as a data-driven short-term measure of non-retention in engineering.

Recommended citation: Boujelbene, M., Damak, K., Acun Sener, A. C., Hieb, J. L., Bego, C. R., Ralston, P. A., & Nasraoui, O. (2020). "A Data Science Approach to Flagging Non-Retention in Engineering Enrollment Data." 2020 ASEE Virtual Annual Conference. https://peer.asee.org/a-data-science-approach-to-flagging-non-retention-in-engineering-enrollment-data

In-training explainability frameworks: A method to make black-box machine learning models more explainable

Published in 2023 IEEE/WIC International Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT), 2023

Introduces pre-hoc and co-hoc explainability, preserving black-box prediction accuracy while drawing explanations from an inherently interpretable white-box model.

Recommended citation: Acun, C., & Nasraoui, O. (2023). "In-training explainability frameworks: A method to make black-box machine learning models more explainable." 2023 IEEE/WIC International Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT), 230-237. https://ieeexplore.ieee.org/document/10350149

GAI-Enhanced Assignment Framework: A Case Study on Generative AI Powered History Education

Published in NeurIPS 2023 Workshop on Generative AI for Education, 2023

A case study of ChatGPT role-play assignments in an undergraduate History of Science course, introducing the Reference-Check Protocol as a safeguard for accuracy and academic integrity.

Recommended citation: Acun, C., & Acun, R. (2023). "GAI-Enhanced Assignment Framework: A Case Study on Generative AI Powered History Education." NeurIPS 2023 Workshop on Generative AI for Education. https://gaied.org/neurips2023/files/48/48_paper.pdf

ChatGPT for conversational recommendation: Refining recommendations by reprompting with feedback

Published in arXiv preprint arXiv:2401.03605, 2024

Evaluates ChatGPT as a top-n conversational recommender, showing that reprompting with user feedback improves relevancy and that popularity bias can be mitigated through prompt engineering.

Recommended citation: Spurlock, K. D., Acun, C., Saka, E., & Nasraoui, O. (2024). "ChatGPT for conversational recommendation: Refining recommendations by reprompting with feedback." arXiv preprint arXiv:2401.03605. https://arxiv.org/abs/2401.03605

In-training explainability frameworks to make black-box machine learning models more explainable

Published in Ph.D. dissertation, University of Louisville, 2024

Doctoral dissertation introducing pre-hoc and co-hoc explainability, with global and local variants evaluated on credit risk scoring, movie recommendation and robotic grasp failure detection.

Recommended citation: Acun, A. C. (2024). "In-training explainability frameworks to make black-box machine learning models more explainable." Ph.D. dissertation, University of Louisville. https://ir.library.louisville.edu/etd/4375/

Enhancing Robotic Grasp Failure Prediction Using A Pre-hoc Explainability Framework

Published in 2024 IEEE 20th International Conference on Automation Science and Engineering (CASE), 2024

Adds pre-hoc explainability to robot grasp failure prediction with Factorization Machines, learning the explainable model before the black-box model rather than after it.

Recommended citation: Acun, C., Ashary, A., Popa, D. O., & Nasraoui, O. (2024). "Enhancing Robotic Grasp Failure Prediction Using A Pre-hoc Explainability Framework." 2024 IEEE 20th International Conference on Automation Science and Engineering (CASE), 1993-1998. https://ieeexplore.ieee.org/document/10711481

Robot failure mode prediction with deep learning sequence models

Published in Neural Computing and Applications, 2025

Recurrent neural network models for robot grasp failure prediction, introducing generative-interpretive grasp prediction and adaptive time to failure analysis.

Recommended citation: Damak, K., Boujelbene, M., Acun, C., Alvanpour, A., Das, S. K., Popa, D. O., & Nasraoui, O. (2025). "Robot failure mode prediction with deep learning sequence models." Neural Computing and Applications, 37(6), 4291-4302. https://link.springer.com/article/10.1007/s00521-024-10856-1

Comparative Analysis of Post Hoc Explainable Methods for Robotic Grasp Failure Prediction

Published in Electronics, 2025

A comparison of Tree-SHAP, LIME and TreeInterpreter on grasp failure predictions, covering agreement between methods, dependence on model type and computational efficiency.

Recommended citation: Alvanpour, A., Acun, C., Spurlock, K., Robinson, C. K., Das, S. K., Popa, D. O., & Nasraoui, O. (2025). "Comparative Analysis of Post Hoc Explainable Methods for Robotic Grasp Failure Prediction." Electronics, 14(9), 1868. https://www.mdpi.com/2079-9292/14/9/1868

Pre Hoc and Co Hoc Explainability: Frameworks for Integrating Interpretability into Machine Learning Training for Enhanced Transparency and Performance

Published in Applied Sciences, 2025

Pre hoc and co hoc frameworks that embed an inherently interpretable white-box model into black-box training, improving global and local fidelity over LIME without sacrificing accuracy.

Recommended citation: Acun, C., & Nasraoui, O. (2025). "Pre Hoc and Co Hoc Explainability: Frameworks for Integrating Interpretability into Machine Learning Training for Enhanced Transparency and Performance." Applied Sciences, 15(13), 7544. https://www.mdpi.com/2076-3417/15/13/7544

Task Allocation for Nursing Robots Using Explainable Machine Learning with Factorization Machines

Published in 2025 IEEE 21st International Conference on Automation Science and Engineering (CASE), 2025

An explainable pre-hoc Factorization Machine framework that predicts optimal task allocation between nurses and robotic nursing assistants with 94% accuracy.

Recommended citation: Acun, C., Ashary, A., Popa, D. O., & Nasraoui, O. (2025). "Task Allocation for Nursing Robots Using Explainable Machine Learning with Factorization Machines." 2025 IEEE 21st International Conference on Automation Science and Engineering (CASE), 1044-1049. https://ieeexplore.ieee.org/document/11164103

talks

teaching