pith. the verified trust layer for science. sign in

arxiv: 2601.15280 · v1 · pith:6KUQCWWCnew · submitted 2026-01-21 · 💻 cs.HC · cs.AI

LLM-based Multimodal Feedback Produces Equivalent Learning and Better Student Perceptions than Educator Feedback

Chloe Qianhui Zhao , Jie Cao , Jionghao Lin , Kenneth R. Koedinger This is my paper

Pith reviewed 2026-05-16 11:44 UTC · model grok-4.3

classification 💻 cs.HC cs.AI
keywords AI feedbackmultimodal feedbackLLMstudent perceptionslearning outcomescognitive loadonline educationfeedback quality
0
0 comments X p. Extension
Add this Pith Number to your LaTeX paper What is a Pith Number? 
\usepackage{pith}
\pithnumber{6KUQCWWC}

Prints a linked pith:6KUQCWWC badge after your title and writes the identifier into PDF metadata. Compiles on arXiv with no extra files. Learn more

The pith

AI multimodal feedback matches educator feedback in learning gains but improves student perceptions of clarity and motivation.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper tests an AI system that gives feedback by combining text explanations with relevant slide references and audio narration. The experiment compared it to fixed educator feedback using online participants solving questions. Students showed the same learning gains either way, but they rated the AI version higher for clarity, specificity, motivation, and lower mental effort. Engagement logs showed AI feedback encouraged more revisions on open-ended questions while educator feedback prompted more attempts on multiple-choice ones. The work points to a way for AI to deliver scalable feedback that keeps learning effective while raising student satisfaction.

Core claim

LLM-based multimodal feedback that integrates structured textual explanations with dynamic multimedia resources, including retrieved slide page references and streaming AI audio narration, achieves learning gains equivalent to fixed educator feedback while significantly outperforming it on perceived clarity, specificity, conciseness, motivation, satisfaction, and cognitive load, with comparable correctness, trust, and acceptance.

What carries the argument

The real-time AI-facilitated multimodal feedback system that combines structured textual explanations with retrieved most relevant slide pages and streaming AI audio narration.

If this is right

  • Feedback can be delivered instantly at scale without reducing measured learning outcomes.
  • Different question types trigger distinct engagement: educator feedback increases submissions on multiple-choice items while AI suggestions increase iterations on open-ended items.
  • Student satisfaction rises on clarity and motivation without lowering trust or acceptance of the feedback.
  • Instructor workload can decrease for routine feedback tasks while preserving equivalent learning gains.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same multimodal approach could extend to subjects with rich visual materials like diagrams or videos if slide-like resources are available.
  • Adding student history or prior errors might allow the system to generate even more targeted suggestions beyond the tested setup.
  • Platform designers could use the observed engagement differences to route question types to AI or human feedback selectively.

Load-bearing premise

Crowdsourced online participants using fixed pre-written educator feedback accurately represent real student learning processes and typical educator feedback quality in authentic educational settings.

What would settle it

A controlled study in a real classroom comparing the AI multimodal system to live personalized educator feedback, measuring pre-to-post learning gains on the same material and collecting perception ratings.

Figures

Figures reproduced from arXiv: 2601.15280 by Chloe Qianhui Zhao, Jie Cao, Jionghao Lin, Kenneth R. Koedinger.

Figure 1
Figure 1. Figure 1: A screenshot of the learner interface of the AI multimodal feedback system when embedding for an open-ended [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Comparison of the feedback composition in different conditions. [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Pre- and Post-Test Scores (%) Showing Comparable [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
read the original abstract

Providing timely, targeted, and multimodal feedback helps students quickly correct errors, build deep understanding and stay motivated, yet making it at scale remains a challenge. This study introduces a real-time AI-facilitated multimodal feedback system that integrates structured textual explanations with dynamic multimedia resources, including the retrieved most relevant slide page references and streaming AI audio narration. In an online crowdsourcing experiment, we compared this system against fixed business-as-usual feedback by educators across three dimensions: (1) learning effectiveness, (2) learner engagement, (3) perceived feedback quality and value. Results showed that AI multimodal feedback achieved learning gains equivalent to original educator feedback while significantly outperforming it on perceived clarity, specificity, conciseness, motivation, satisfaction, and reducing cognitive load, with comparable correctness, trust, and acceptance. Process logs revealed distinct engagement patterns: for multiple-choice questions, educator feedback encouraged more submissions; for open-ended questions, AI-facilitated targeted suggestions lowered revision barriers and promoted iterative improvement. These findings highlight the potential of AI multimodal feedback to provide scalable, real-time, and context-aware support that both reduces instructor workload and enhances student experience.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript introduces a real-time LLM-based multimodal feedback system that combines structured textual explanations with dynamic multimedia elements such as relevant slide references and streaming AI audio narration. It reports results from an online crowdsourcing experiment comparing this system to fixed educator feedback on three dimensions: learning effectiveness, learner engagement, and perceived feedback quality. The central claims are that AI multimodal feedback produces equivalent learning gains to educator feedback while significantly outperforming it on perceived clarity, specificity, conciseness, motivation, satisfaction, and cognitive load, with comparable correctness, trust, and acceptance; process logs also indicate distinct engagement patterns for multiple-choice versus open-ended questions.

Significance. If the equivalence and perception results hold under more rigorous controls, the work could support scalable, real-time feedback tools that reduce instructor workload in educational settings. The integration of multimodal elements and the reported differences in revision behavior for different question types represent potentially useful design insights for HCI and learning technologies. However, the current evidence base is limited by the study population and feedback format, reducing immediate generalizability.

major comments (2)
  1. [Online Crowdsourcing Experiment] The central equivalence claim on learning gains rests on an online crowdsourcing experiment whose validity is undermined by the use of non-student participants and fixed pre-written educator feedback; these choices do not proxy authentic student motivation, prior knowledge, or the context-sensitive, adaptive nature of live educator responses, as noted in the skeptic analysis of the weakest assumption.
  2. [Abstract and Results] The abstract and results summary report equivalence on learning gains and superiority on multiple perception metrics but supply no sample size, statistical tests, effect sizes, pre-registration details, or controls for confounds such as feedback length or timing; without these, the load-bearing claims cannot be fully evaluated for robustness.
minor comments (1)
  1. [System Description] Clarify the exact procedure for retrieving and presenting slide page references and streaming audio narration to allow replication.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which help us clarify the scope and strengthen the reporting of our work. We address each major point below and have made targeted revisions to improve transparency and acknowledge limitations.

read point-by-point responses

  1. Referee: [Online Crowdsourcing Experiment] The central equivalence claim on learning gains rests on an online crowdsourcing experiment whose validity is undermined by the use of non-student participants and fixed pre-written educator feedback; these choices do not proxy authentic student motivation, prior knowledge, or the context-sensitive, adaptive nature of live educator responses, as noted in the skeptic analysis of the weakest assumption.

    Authors: We agree that crowdsourced participants and fixed educator feedback reduce ecological validity relative to live classroom interactions with adaptive responses. This design was chosen to enable precise control, objective pre/post learning measures, and detailed process logging at scale. We have expanded the Limitations section to explicitly discuss the non-student sample, the use of pre-written feedback, and the need for future classroom studies with actual students and dynamic educator input. The reported equivalence is grounded in objective test scores rather than self-report, which we maintain provides initial evidence for the approach despite the acknowledged constraints. revision: partial

  2. Referee: [Abstract and Results] The abstract and results summary report equivalence on learning gains and superiority on multiple perception metrics but supply no sample size, statistical tests, effect sizes, pre-registration details, or controls for confounds such as feedback length or timing; without these, the load-bearing claims cannot be fully evaluated for robustness.

    Authors: We have revised the abstract to include sample size, the specific statistical tests used for equivalence and superiority claims, effect sizes, and a reference to the pre-registration. The full results section already reports these details along with controls for feedback length (word-count matching) and timing (immediate delivery for both conditions); we now explicitly summarize the controls in the abstract and methods for clarity. These additions allow direct evaluation of the claims without altering the underlying data or conclusions. revision: yes

Circularity Check

0 steps flagged

No circularity: direct empirical measurements from controlled experiment

full rationale

The paper reports results from an online crowdsourcing experiment that directly measures learning gains, engagement patterns, and perception metrics by comparing AI multimodal feedback against fixed educator feedback. No equations, fitted parameters, predictions, or derivations appear in the manuscript. All reported outcomes (equivalent learning gains, superior perceptions on clarity/specificity/etc.) are presented as raw experimental observations rather than quantities defined in terms of the inputs or reduced via self-citation chains. The study is self-contained against its own data collection protocol with no load-bearing self-referential steps.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim depends on the validity of the experimental contrast between the AI system and fixed educator feedback, plus the assumption that crowdsourced performance generalizes to actual students.

axioms (1)
  • domain assumption Crowdsourced online participants and pre-written educator feedback serve as valid proxies for real classroom learning and typical educator responses.
    The study generalizes from this online experiment to broader educational impact.

pith-pipeline@v0.9.0 · 5511 in / 1416 out tokens · 31400 ms · 2026-05-16T11:44:50.077228+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

44 extracted references · 44 canonical work pages · 2 internal anchors

  1. [1]

    Ahmad Ari Aldino, Yi-Shan Tsai, Siddarth Gupte, Michael Henderson, Debarshi Nath, Dragan Gašević, and Guanliang Chen. 2025. Analytics of Learner-Centered Feedback: A Large-Scale Case Study in Higher Education.Computers & Education 237 (2025), 105360. doi:10.1016/j.compedu.2025.105360

    work page doi:10.1016/j.compedu.2025.105360 2025

  2. [2]

    Lodge, Marie Boden, and Hassan Khosravi

    Omar Alsaiari, Nilufar Baghaei, Hatim Lahza, Jason M. Lodge, Marie Boden, and Hassan Khosravi. 2025. Emotionally enriched AI-generated feedback: Supporting student well-being without compromising learning.Computers & Education239 (2025), 105363. doi:10.1016/j.compedu.2025.105363

    work page doi:10.1016/j.compedu.2025.105363 2025

  3. [3]

    2025.Claude Sonnet 4.5 System Card

    Anthropic. 2025.Claude Sonnet 4.5 System Card. Technical Report. Anthropic PBC

    work page 2025

  4. [4]

    Moraes, Fernanda Oliveira, and Carla A

    Juliana Barros, Laura O. Moraes, Fernanda Oliveira, and Carla A. D. M. Delgado

    work page

  5. [5]

    InArtificial Intelligence in Education, Alexandra I

    Large Language Models Generating Feedback for Students of Introductory Programming Courses. InArtificial Intelligence in Education, Alexandra I. Cristea, Erin Walker, Yu Lu, Olga C. Santos, and Seiji Isotani (Eds.). Springer Nature Switzerland, Cham, 421–433

    work page

  6. [6]

    Bruner, Jacqueline J

    Jerome S. Bruner, Jacqueline J. Goodnow, and George A. Austin. 1956.A Study of Thinking. John Wiley and Sons, New York, NY

    work page 1956

  7. [7]

    Koedinger, and Jionghao Lin

    Jie Cao, Chloe Qianhui Zhao, Xian Chen, Shuman Wang, Christian Schunn, Kenneth R. Koedinger, and Jionghao Lin. 2025. From First Draft to Final Insight: A Multi-agent Approach for Feedback Generation. InArtificial Intelligence in Education, Alexandra I. Cristea, Erin Walker, Yu Lu, Olga C. Santos, and Seiji Isotani (Eds.). Springer Nature Switzerland, Cham...

    work page 2025

  8. [8]

    David Carless, , and David Boud. 2018. The development of student feedback lit- eracy: enabling uptake of feedback.Assessment & Evaluation in Higher Education 43, 8 (Nov. 2018), 1315–1325. doi:10.1080/02602938.2018.1463354

    work page doi:10.1080/02602938.2018.1463354 2018

  9. [9]

    Anderson Pinheiro Cavalcanti, Arthur Barbosa, Ruan Carvalho, Fred Freitas, Yi- Shan Tsai, Dragan Gašević, and Rafael Ferreira Mello. 2021. Automatic feedback in online learning environments: A systematic literature review.Computers and Education: Artificial Intelligence2 (2021), 100027

    work page 2021

  10. [10]

    Koedinger

    Eason Chen, Xinyi Tang, Aprille Xi, Chenyu Lin, Conrad Borchers, Shivang Gupta, Jionghao Lin, and Kenneth R. Koedinger. 2025. VTutor for High-Impact Tutoring at Scale: Managing Engagement and Real-Time Multi-Screen Monitoring with P2P Connections. InProceedings of the Twelfth ACM Conference on Learning @ Scale (L@S ’25). Association for Computing Machiner...

    work page doi:10.1145/3698205.3733948 2025

  11. [11]

    Hagit Gabbay and Anat Cohen. 2024. Combining LLM-Generated and Test-Based Feedback in a MOOC for Programming. InProceedings of the Eleventh ACM Conference on Learning @ Scale. ACM, Atlanta GA USA, 177–187. doi:10.1145/ 3657604.3662040

    work page arXiv 2024

  12. [12]

    Ge Gao, Amelia Leon, Andrea Jetten, Jasmine Turner, Husni Almoubayyed, Stephen Fancsali, and Emma Brunskill. 2025. Predicting Long-Term Student Out- comes from Short-Term EdTech Log Data. InProceedings of the 15th International Learning Analytics and Knowledge Conference (LAK ’25). Association for Comput- ing Machinery, New York, NY, USA, 631–641. doi:10....

    work page doi:10.1145/3706468.3706552 2025

  13. [13]

    Yunfan Gao, Yun Xiong, Xinyu Gao, Kangxiang Jia, Jinliu Pan, Yuxi Bi, Yi Dai, Ji- awei Sun, Meng Wang, and Haofen Wang. 2024. Retrieval-Augmented Generation for Large Language Models: A Survey. doi:10.48550/arXiv.2312.10997

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2312.10997 2024

  14. [14]

    Google Gemini Team. 2025. Gemini 2.5: Pushing the Frontier with Advanced Rea- soning, Multimodality, Long Context, and Next Generation Agentic Capabilities. https://arxiv.org/abs/2507.06261

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  15. [15]

    Jodi Goodman, Robert Wood, and Margaretha Hendrickx. 2004. Feedback Speci- ficity, Exploration, and Learning.The Journal of applied psychology89 (April 2004), 248–62. doi:10.1037/0021-9010.89.2.248

    work page doi:10.1037/0021-9010.89.2.248 2004

  16. [16]

    John Hattie and Helen Timperley. 2007. The Power of Feedback.Review of Educational Research77, 1 (March 2007), 81–112

    work page 2007

  17. [17]

    Yueqiao Jin, Kaixun Yang, Lixiang Yan, Vanessa Echeverria, Linxuan Zhao, Rior- dan Alfredo, Mikaela Milesi, Jie Xiang Fan, Xinyu Li, Dragan Gasevic, and Roberto Martinez-Maldonado. 2025. Chatting with a Learning Analytics Dashboard: The Role of Generative AI Literacy on Learner Interaction with Conventional and Scaffolding Chatbots. InProceedings of the 1...

    work page doi:10.1145/3706468.3706545 2025

  18. [18]

    Jelena Jovanovic, Negin Mirriahi, Dragan Gašević, Shane Dawson, and Abelardo Pardo. 2019. Predictive power of regularity of pre-class activities in a flipped classroom.Computers & Education134 (2019), 156–168. doi:10.1016/j.compedu. 2019.02.011

    work page doi:10.1016/j.compedu 2019

  19. [19]

    Lukas Jürgensmeier and Bernd Skiera. 2024. Generative AI for scalable feedback to multimodal exercises.International Journal of Research in Marketing41, 3 (2024), 468–488. doi:10.1016/j.ijresmar.2024.05.005

    work page doi:10.1016/j.ijresmar.2024.05.005 2024

  20. [20]

    Britnie Delinger Kane and Brooks Rosenquist. 2019. Relationships Between Instructional Coaches’ Time Use and District- and School-Level Policies and Expectations.American Educational Research Journal56, 5 (2019), 1718–1768. doi:10.3102/0002831219826580

    work page doi:10.3102/0002831219826580 2019

  21. [21]

    Enkelejda Kasneci, Kathrin Sessler, Stefan Küchemann, Maria Bannert, Daryna Dementieva, Frank Fischer, Urs Gasser, Georg Groh, Stephan Günnemann, Eyke Hüllermeier, Stephan Krusche, Gitta Kutyniok, Tilman Michaeli, Claudia Nerdel, Jürgen Pfeffer, Oleksandra Poquet, Michael Sailer, Albrecht Schmidt, Tina Seidel, Matthias Stadler, Jochen Weller, Jochen Kuhn,...

    work page doi:10.1016/j.lindif 2023

  22. [22]

    Hieke Keuning, Johan Jeuring, and Bastiaan Heeren. 2018. A Systematic Literature Review of Automated Feedback Generation for Programming Exercises.ACM Trans. Comput. Educ.19, 1 (Sept. 2018). doi:10.1145/3231711

    work page doi:10.1145/3231711 2018

  23. [23]

    Gyeonggeon Lee, Lehong Shi, Ehsan Latif, Yizhu Gao, Arne Bewersdorff, Matthew Nyaaba, Shuchen Guo, Zhengliang Liu, Gengchen Mai, Tianming Liu, and Xi- aoming Zhai. 2025. Multimodality of AI for Education: Toward Artificial General Intelligence.IEEE Transactions on Learning Technologies18 (2025), 666–683. doi:10.1109/TLT.2025.3574466

    work page doi:10.1109/tlt.2025.3574466 2025

  24. [24]

    Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, and Douwe Kiela. 2021. Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks

    work page 2021

  25. [25]

    Koedinger

    Jionghao Lin, Eason Chen, Ashish Gurung, and Kenneth R. Koedinger. 2024. MuFIN: A Framework for Automating Multimodal Feedback Generation using Generative Artificial Intelligence. InProceedings of the Eleventh ACM Conference on Learning @ Scale. ACM, Atlanta GA USA, 550–552

    work page 2024

  26. [26]

    Richard E. Mayer. 2020.Multimedia Learning(3 ed.). Cambridge University Press

    work page 2020

  27. [27]

    Nicol and Debra Macfarlane-Dick

    David J. Nicol and Debra Macfarlane-Dick. 2006. Formative assessment and self-regulated learning: a model and seven principles of good feedback practice. Studies in Higher Education31, 2 (2006), 199–218. doi:10.1080/03075070600572090

    work page doi:10.1080/03075070600572090 2006

  28. [28]

    2025.GPT-5 System Card

    OpenAI. 2025.GPT-5 System Card. System Card / Technical Report. OpenAI. https://cdn.openai.com/gpt-5-system-card.pdf

    work page 2025

  29. [29]

    OpenAI. 2025. Introducing gpt-realtime and Realtime API updates for production voice agents. https://openai.com/index/introducing-gpt-realtime/

    work page 2025

  30. [30]

    Allan Paivio. 1990. Dual Coding Theory. InMental Representations: A dual coding approach. Oxford University Press

    work page 1990

  31. [31]

    Jaeuk Park. 2024. Students’ perceptions toward providing video-enhanced mul- timodal feedback on oral presentations.Computer Assisted Language Learning (May 2024), 1–27. doi:10.1080/09588221.2024.2344546

    work page doi:10.1080/09588221.2024.2344546 2024

  32. [32]

    Michael Prince. 2004. Does Active Learning Work? A Review of the Research. Journal of Engineering Education93, 3 (2004), 223–231

    work page 2004

  33. [33]

    Tracii Ryan, Michael Henderson, and Michael Phillips. 2019. Feedback modes matter: Comparing student perceptions of digital and non-digital feedback modes in higher education.British Journal of Educational Technology50, 3 (2019), 1507–

    work page 2019

  34. [34]

    doi:10.1111/bjet.12749

    work page doi:10.1111/bjet.12749

  35. [35]

    Tracii Ryan, Michael Henderson, Kris Ryan, and Gregor Kennedy. 2021. Designing learner-centred text-based feedback: a rapid review and qualitative synthesis. Assessment & Evaluation in Higher Education46, 6 (Aug. 2021), 894–912. doi:10. 1080/02602938.2020.1828819

    work page arXiv 2021

  36. [36]

    Sylvio Rüdian, Julia Podelo, Jakub Kužílek, and Niels Pinkwart. 2025. Feedback on Feedback: Student’s Perceptions for Feedback from Teachers and Few-Shot LLMs. InProceedings of the 15th International Learning Analytics and Knowledge Conference (LAK ’25). Association for Computing Machinery, New York, NY, USA, 82–92. doi:10.1145/3706468.3706479

    work page doi:10.1145/3706468.3706479 2025

  37. [37]

    Kathrin Seßler, Arne Bewersdorff, Claudia Nerdel, and Enkelejda Kasneci. 2025. Towards Adaptive Feedback with AI: Comparing the Feedback Quality of LLMs and Teachers on Experimentation Protocols. https://arxiv.org/abs/2502.12842

    work page arXiv 2025

  38. [38]

    Insub Shin, Su Bhin Hwang, Yun Joo Yoo, Sooan Bae, and Rae Yeong Kim. 2025. Comparing Student Preferences for AI-Generated and Peer-Generated Feedback in AI-driven Formative Peer Assessment. InProceedings of the 15th International Learning Analytics and Knowledge Conference (LAK ’25). Association for Comput- ing Machinery, New York, NY, USA, 159–169. doi:...

    work page doi:10.1145/3706468.3706488 2025

  39. [39]

    Valerie J. Shute. 2008. Focus on Formative Feedback.Review of Educational Research78, 1 (2008), 153–189. doi:10.3102/0034654307313795

    work page doi:10.3102/0034654307313795 2008

  40. [40]

    Matthias Stadler, Maria Bannert, and Michael Sailer. 2024. Cognitive ease at a cost: LLMs reduce mental effort but compromise depth in student scientific inquiry. Computers in Human Behavior160 (2024), 108386. doi:10.1016/j.chb.2024.108386

    work page doi:10.1016/j.chb.2024.108386 2024

  41. [41]

    John Sweller. 2011. CHAPTER TWO - Cognitive Load Theory. Psychology of Learning and Motivation, Vol. 55. Academic Press, 37–76

    work page 2011

  42. [42]

    Devika Venugopalan, Ziwen Yan, Conrad Borchers, Jionghao Lin, and Vin- cent Aleven. 2025. Combining Large Language Models with Tutoring Sys- tem Intelligence: A Case Study in Caregiver Homework Support. InProceed- ings of the 15th International Learning Analytics and Knowledge Conference (LAK ’25). Association for Computing Machinery, New York, NY, USA, 3...

    work page doi:10.1145/3706468.3706516 2025

  43. [43]

    Zehan Wang, Ke Lei, Chen Zhu, Jiawei Huang, Sashuai Zhou, Luping Liu, Xize Cheng, Shengpeng Ji, Zhenhui Ye, Tao Jin, and Zhou Zhao. 2025. T2A-Feedback: Improving Basic Capabilities of Text-to-Audio Generation via Fine-grained AI Feedback. InProceedings of the 63rd Annual Meeting of the Association for Com- putational Linguistics (Volume 1: Long Papers), W...

    work page doi:10.18653/v1/2025.acl- 2025

  44. [44]

    In: Experimental Al- gorithms, 7th International Workshop, WEA 2008, Provincetown, MA, USA, May 30-June 1, 2008, Proceedings

    Chloe Qianhui Zhao, Jie Cao, Eason Chen, Kenneth R. Koedinger, and Jionghao Lin. 2025. SlideItRight: Using AI to Find Relevant Slides and Provide Feedback for Open-Ended Questions. InArtificial Intelligence in Education, Alexandra I. Cristea, Erin Walker, Yu Lu, Olga C. Santos, and Seiji Isotani (Eds.). Vol. 15880. Springer Nature Switzerland, Cham, 378–3...

    work page doi:10.1007/978-3- 2025