Reproducibility in machine‐learning‐based research: Overview, barriers, and drivers
Open accessAbstract
Many research fields are currently reckoning with issues of poor levels of reproducibility. Some label it a “crisis,” and research employing or building machine learning (ML) models is no exception. Issues including lack of transparency, data or code, poor adherence to standards, and the sensitivity of ML training conditions mean that many papers are not even reproducible in principle. Where they are, though, reproducibility experiments have found worryingly low degrees of similarity with original results. Despite previous appeals from ML researchers on this topic and various initiatives from conference reproducibility tracks to the ACM's new Emerging Interest Group on Reproducibility and Replicability, we contend that the general community continues to take this issue too lightly. Poor reproducibility threatens trust in and integrity of research results. Therefore, in this article, we lay out a new perspective on the key barriers and drivers (both procedural and technical) to increased reproducibility at various levels (methods, code, data, and experiments). We then map the drivers to the barriers to give concrete advice for strategies for researchers to mitigate reproducibility issues in their own work, to lay out key areas where further research is needed in specific areas, and to further ignite discussion on the threat presented by these urgent issues.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
All experiments were conducted in the Google Colab environment with the same versions of the same libraries.
“All experiments were conducted in the Google Colab environment, in line with reported best-practices (Semmelrock et al., 2025), with the same versions of the same libraries.”
From CiteFusion: an ensemble framework for citation intent classification harnessing dual-model binary couples and SHAP analyses (2025) · cites this paper as Semmelrock et al., 2025
These factors are inherent to modern computational pipelines in deep learning.
“These factors are inherent to modern computational pipelines in deep learning (Desai et al., 2025; Goodman et al., 2016; Semmelrock et al., 2025)”
From CiteFusion: an ensemble framework for citation intent classification harnessing dual-model binary couples and SHAP analyses (2025) · cites this paper as Semmelrock et al., 2025
Some degree of run-to-run variability is unavoidable in deep learning frameworks, particularly when dealing with Language Models.
“recent works and official documentation emphasize that some degree of run-to-run variability is unavoidable in deep learning frameworks (Desai et al., 2025; Gundersen et al., 2018), in particular when dealing with Language Models (Semmelrock et al., 2025).”
From CiteFusion: an ensemble framework for citation intent classification harnessing dual-model binary couples and SHAP analyses (2025) · cites this paper as Semmelrock et al., 2025
Cites
Estimating the reproducibility of psychological science
Open Science Collaboration (2015) · Science · doi:10.1126/science.aac4716 · Cited by 8,515 · see citing works
Dwivedi, Y. K., Hughes, L., Ismagilova, E., Aarts, G., Coombs, C., Crick, T., Duan, Y., Dwivedi, R., Edwards, J., Eirug, A., et al. (2021) · International Journal of Information Management · doi:10.1016/j.ijinfomgt.2019.08.002 · Cited by 4,611 · see citing works
1,500 scientists lift the lid on reproducibility
Baker, M. (2016) · Nature · doi:10.1038/533452a · Cited by 4,209 · see citing works
MIMIC-IV, a freely accessible electronic health record dataset
Johnson, A. E. W., Bulgarelli, L., Shen, L., Gayles, A., Shammout, A., Horng, S., Pollard, T. J., Hao, S., Moody, B., Gow, B., et al. (2023) · Scientific Data · doi:10.1038/s41597-022-01899-x · Cited by 3,401 · see citing works
Model Cards for Model Reporting
Mitchell, M., Wu, S., Zaldivar, A., Barnes, P., Vasserman, L., Hutchinson, B., Spitzer, E., Raji, I. D., & Gebru, T. (2019) · Proceedings of the Conference on Fairness, Accountability, and Transparency · doi:10.1145/3287560.3287596 · Cited by 2,681 · see citing works
Checklist for Artificial Intelligence in Medical Imaging (CLAIM): a guide for authors and reviewers
Mongan, J., Moy, L., & Kahn, C. E. (2020) · Radiology: Artificial Intelligence · doi:10.1148/ryai.2020200029 · Cited by 1,443 · see citing works
Leakage and the reproducibility crisis in machine-learning-based science
Kapoor, S., & Narayanan, A. (2023) · Patterns · doi:10.1016/j.patter.2023.100804 · Cited by 1,387 · see citing works
What does research reproducibility mean?
Goodman, S. N., Fanelli, D., & Ioannidis, J. P. A. (2016) · Science Translational Medicine · doi:10.1126/scitranslmed.aaf5027 · Cited by 1,293 · see citing works
The economics of reproducibility in preclinical research
Freedman, L. P., Cockburn, I. M., & Simcoe, T. S. (2015) · PLOS Biology · doi:10.1371/journal.pbio.1002165 · Cited by 1,055 · see citing works
The natural selection of bad science
Smaldino, P. E., & McElreath, R. (2016) · Royal Society Open Science · doi:10.1098/rsos.160384 · Cited by 860 · see citing works
Replicability, Robustness, and Reproducibility in Psychological Science
Nosek, B. A., Hardwicke, T. E., Moshontz, H., Allard, A., Corker, K. S., Dreber, A., Fidler, F., Hilgard, J., Kline Struhl, M., Nuijten, M. B., et al. (2022) · Annual Review of Psychology · doi:10.1146/annurev-psych-020821-114157 · Cited by 770 · see citing works
Investigating the replicability of preclinical cancer biology
Errington, T. M., Mathur, M., Soderberg, C. K., Denis, A., Perfito, N., Iorns, E., & Nosek, B. A. (2021) · eLife · doi:10.7554/elife.71601 · Cited by 393 · see citing works
Hardwicke, T. E., Mathur, M. B., MacDonald, K., Nilsonne, G., Banks, G. C., Kidwell, M. C., Hofelich Mohr, A., Clayton, E., Yoon, E. J., Henry Tessler, M., et al. (2018) · Royal Society Open Science · doi:10.1098/rsos.180448 · Cited by 336 · see citing works
Reproducible Research Practices and Transparency across the Biomedical Literature
Iqbal, S. A., Wallach, J. D., Khoury, M. J., Schully, S. D., & Ioannidis, J. P. A. (2016) · PLOS Biology · doi:10.1371/journal.pbio.1002333 · Cited by 310 · see citing works
State of the Art: Reproducibility in Artificial Intelligence
Gundersen, O. E., & Kjensmo, S. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.11503 · Cited by 293 · see citing works
Holistic Evaluation of Language Models
Liang, P., Bommasani, R., Lee, T., Tsipras, D., Soylu, D., Yasunaga, M., Zhang, Y., Narayanan, D., Wu, Y., Kumar, A., et al. (2022) · arXiv · doi:10.48550/arxiv.2211.09110 · Cited by 255 · see citing works
Mitigating Bias in Radiology Machine Learning: 1. Data Handling
Rouzrokh, P., Khosravi, B., Faghani, S., Moassefi, M., Vera Garcia, D. V., Singh, Y., Zhang, K., Conte, G. M., & Erickson, B. J. (2022) · Radiology: Artificial Intelligence · doi:10.1148/ryai.210290 · Cited by 145 · see citing works
Han, S., Olonisakin, T. F., Pribis, J. P., Zupetic, J., Yoon, J. H., Holleran, K. M., Jeong, K., Shaikh, N., Rubio, D. M., & Lee, J. S. (2017) · PLOS ONE · doi:10.1371/journal.pone.0183591 · Cited by 131 · see citing works
Liesenfeld, A., Lopez, A., & Dingemanse, M. (2023) · Proceedings of the 5th International Conference on Conversational User Interfaces · doi:10.1145/3571884.3604316 · Cited by 111 · see citing works
Computational Reproducibility via Containers in Psychology
Clyburne-Sherin, A., Fei, X., & Green, S. A. (2019) · Meta-Psychology · doi:10.15626/mp.2018.892 · Cited by 69 · see citing works
Open Science at the generative AI turn: An exploratory analysis of challenges and opportunities
Hosseini, M., Horbach, S. P. J. M., Holmes, K., & Ross-Hellauer, T. (2025) · Quantitative Science Studies · doi:10.1162/qss_a_00337 · Cited by 40 · see citing works
Self-citationReviewerGPT? An Exploratory Study on Using Large Language Models for Paper Reviewing
Liu, R., & Shah, N. B. (2023) · arXiv · doi:10.48550/arxiv.2306.00622 · Cited by 23 · see citing works
Navarro Leija, O. S., Shiptoski, K., Scott, R. G., Wang, B., Renner, N., Newton, R. R., & Devietti, J. (2020) · Proceedings of the Twenty-Fifth International Conference on Architectural Support for Programming Languages and Operating Systems · doi:10.1145/3373376.3378519 · Cited by 22 · see citing works
Lessons from the Trenches on Reproducible Evaluation of Language Models
Biderman, S., Schoelkopf, H., Sutawika, L., Gao, L., Tow, J., Abbasi, B., Aji, A. F., Ammanamanchi, P. S., Black, S., Clive, J., et al. (2024) · arXiv · doi:10.48550/arxiv.2405.14782 · Cited by 9 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
Paolini, L., Vahdati, S., Di Iorio, A., Wardenga, R., Heibi, I., & Peroni, S. (2025) · Scientometrics · doi:10.1007/s11192-025-05418-8 · Cited by 8 · see citing works
- Authors
- Semmelrock, H., Ross‐Hellauer, T., Kopeinik, S., Theiler, D., Haberl, A., Thalmann, S., & Kowald, D.
- Affiliations
- Graz University of Technology (AT)
- Know Center Research GmbH (Austria) (AT)
- University of Graz (AT)
- Corresponding author
- Kowald, Dominik
- ORCID
- Venue
- AI Magazine
- Publisher
- Wiley
- Published
- 2025-04-14
- Volume
- 46
- Issue
- 2
- ISSN
- 0738-4602 (ISSN-L)
- 2371-9621
- OpenAlex
- W4409415146
- Cited by
- 118
- References
- 24
- Citations by year
- 2026: 88 · 2025: 29 · 2024: 1
- FWCI
- 77.19
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- AI Magazine (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion, cc-by)
- arXiv (Cornell University) (cc-by)
- Funders
- HORIZON EUROPE Framework Programme (101094817)
- Topics
- Social Sciences › Decision Sciences › Information Systems and Management › Scientific Computing and Data Management
- Physical Sciences › Computer Science › Information Systems › Research Data Management Practices
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Data Classification
- Keywords
- reproducibility, machine learning research, research integrity, code availability, data availability, transparency, research trust
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
