Communication efficiency, non-IID data, privacy and health-care applications: the federated learning papers cited most, with surveys and search strings.
Federated learning trains a shared model while the training data stay distributed across many clients. The oldest listed paper is Konečný et al. (2016) on communication efficiency. Two overviews organise it: Li et al. (2020), Challenges, Methods, and Future Directions, and Kairouz et al. (2021), Advances and Open Problems, written by 59 authors. The recurring technical problems are statistical heterogeneity (Zhao et al., 2018), communication cost and privacy guarantees (Wei et al., 2020).
How the literature is organised
Konečný et al. (2016) describe the setting, in which data remain distributed over a large number of clients with unreliable and relatively slow connections, and set out strategies for reducing communication. Within four years the idea had spread to edge computing, wireless networks and medicine, and overviews followed. Li et al. (2020) discuss the characteristics and challenges of federated learning, current approaches and future directions in a magazine-length article. Kairouz et al. (2021) is the long one, a monograph of advances and open problems. Lim et al. (2020) survey federated learning in mobile edge networks.
Much of the most-cited work is in IEEE journals on communications and the Internet of Things: adaptive federated learning under resource constraints (Wang et al., 2019), joint learning and communication over wireless networks (Chen et al., 2021) and energy efficiency (Yang et al., 2021).
Medicine is the most-cited application here. Rieke et al. (2020) set out the promise for digital health, Sheller et al. (2020) demonstrated multi-institutional collaboration without sharing patient data, and Dayan et al. (2021) trained a model on data from 20 institutes to predict the future oxygen requirements of symptomatic COVID-19 patients.
Main debates
The main technical debate concerns trade-offs among model performance, privacy and cost. Wei et al. (2020) show a trade-off between convergence performance and the level of privacy protection under differential privacy. A second question is whether one global model is even the right target when clients differ, which motivates personalised federated learning (Tan et al., 2023). A third concerns trust in data sharing between participants, where Lu et al. (2020) combine federated learning with blockchain for the industrial Internet of Things.
Where recent work is heading
Recent highly cited papers address efficiency through knowledge distillation (Wu et al., 2022), security applications such as anomaly detection for the Internet of Things (Mothukuri et al., 2022) and, in a survey, smart health care (Nguyen et al., 2023). Surveys have multiplied; Zhang et al. (2021) and Wen et al. (2023) are general ones.
Most-cited foundational papers
Published before 2021 and ranked by how often later work cites them. Read the abstract of each and the full text of the three or four closest to your question. Citation count measures attention, not quality, so treat this as a map of what the field has argued about rather than a ranking of what is true.
Yunlong Lu and 4 others (2020). IEEE Transactions on Industrial Informatics.
Cited by 1,265doi:10.1109/tii.2019.2942190
Most-cited papers since 2021
Primary studies and conceptual papers from 2021 onwards. A paper published in 2024 has had a few years to accumulate citations where the works in the section above have had decades, so compare these counts with each other rather than with the ones above.
Chuhan Wu and 4 others (2022). Nature Communications.
Cited by 623Open accessdoi:10.1038/s41467-022-29763-x
Recent reviews and meta-analyses
The fastest way into a literature. A good review gives you the structure of the field, a reference list to mine and, in its limitations section, the gaps other researchers have already spotted.
Dinh C. Nguyen and 7 others (2023). ACM Computing Surveys.
Cited by 801doi:10.1145/3501296
How big the literature is, and where it is published
OpenAlex indexes 37,733 works whose title matches this topic. The chart shows how many were published each year from 2005 to 2025; the current year is left out because it is incomplete.
Show the numbers as a table
Year
Works
2005
2
2006
1
2007
2
2008
1
2011
2
2016
3
2017
6
2018
46
2019
222
2020
987
2021
2,166
2022
3,413
2023
5,493
2024
7,370
2025
9,411
Journals behind the most-cited work
Counted across the 168 most-cited works on the topic, not across everything published. Browsing recent issues of the first two or three is a reliable way to find current work that has not yet been cited much.
IEEE Internet of Things Journal23 papers
IEEE Transactions on Industrial Informatics10 papers
IEEE Transactions on Wireless Communications8 papers
IEEE Transactions on Information Forensics and Security7 papers
ACM Computing Surveys5 papers
Future Generation Computer Systems5 papers
IEEE Transactions on Parallel and Distributed Systems5 papers
IEEE Access4 papers
Search strings to copy
Written for databases that accept Boolean operators (Scopus, Web of Science, ERIC, PubMed, EBSCO). Quotation marks keep a phrase together, an asterisk stands in for the end of a word, and OR groups go in brackets. British and American spellings are written out with OR rather than covered by a single-character wildcard, because those wildcards differ between databases: PubMed’s help page documents the asterisk only, and asks for at least four characters before it. Limit the search to title and abstract first; widen it only if you get too little. Our guide to starting a literature review covers how to record what you searched.
Core problems
"federated learning" AND ("non-IID" OR heterogene* OR "communication efficien*" OR (personalised OR personalized) OR convergence)
Privacy and security
"federated learning" AND ("differential privacy" OR "secure aggregation" OR poisoning OR "inference attack*" OR "homomorphic encryption")
Health care
"federated learning" AND (healthcare OR medical OR clinical OR hospital*) AND (privacy OR "multi-institutional")
Explore federated learning in Wonders →Opens Wonders on this question. Once you are signed in it becomes a project with suggested sub-topics and keywords you can edit before searching. The trial is 14 days.
Sub-topics to narrow into
A thesis-sized question usually sits inside one of these, combined with a population or a setting.
Statistical heterogeneity (non-IID data)
Zhao et al. (2018) showed the accuracy loss; Zhu et al. (2021) survey the remedies.
Communication and resource constraints
Konečný et al. (2016), Wang et al. (2019) and work on wireless networks by Chen et al. (2021) and Yang et al. (2021).
Privacy and security
Wei et al. (2020) on differential privacy; Mothukuri et al. (2021) survey threats.
Personalisation
Tan et al. (2023) review personalised federated learning.
Rieke et al. (2020), Sheller et al. (2020) and the 20-institution COVID-19 study by Dayan et al. (2021).
How to cite these papers
Every paper above has a DOI, a part of the reference that is easy to leave out. Here is one of them, “Federated Learning: Challenges, Methods, and Future Directions” (2020), in the two styles students ask about most:
APA 7th edition
Li, T., Sahu, A. K., Talwalkar, A., & Smith, V. (2020). Federated learning: Challenges, methods, and future directions. IEEE Signal Processing Magazine, 37(3), 50–60. https://doi.org/10.1109/msp.2020.2975749
MLA 9th edition
Li, Tian, et al. “Federated Learning: Challenges, Methods, and Future Directions.” IEEE Signal Processing Magazine, vol. 37, no. 3, 2020, pp. 50–60, https://doi.org/10.1109/msp.2020.2975749.
Check the details against the article itself before you submit: databases, including the one behind this page, sometimes carry the online-first year rather than the volume year. Full rules and more examples are in our guides to APA, MLA, Chicago, Harvard, Vancouver and ABNT, with the rest in the citation guides. You can also format a reference from its DOI with our free citation tools.
Frequently asked questions
Where is the original federated learning paper?
McMahan and colleagues' 2016 paper is titled Communication-Efficient Learning of Deep Networks from Decentralized Data (arXiv:1602.05629). Its title does not contain the words federated learning, so a title search misses it. Konečný et al. (2016), which is listed and has McMahan as second author, dates from the same year.
Does federated learning guarantee privacy?
Not by itself, which is why there is a literature on it. Wei et al. (2020) add differential privacy to federated learning and show a trade-off: better convergence performance means a lower level of privacy protection. Mothukuri et al. (2021) survey security and privacy. When you read a privacy claim, look for the threat model it assumes.
What is non-IID data and why does it matter?
In a federation, each client's data come from a different distribution: one hospital sees different patients from another. Zhao et al. (2018) showed that accuracy fell by up to 55% for neural networks trained on highly skewed non-IID data, where each client held a single class. Zhu et al. (2021) survey the work on this problem.
How this page was made
The lists come from OpenAlex, an open index of scholarly works whose data are published under a CC0 licence, queried on September 21, 2026 for works whose title matches "federated learning". Only works with a DOI are listed. Each one was checked against the publisher’s own record at Crossref or DataCite (title, year, first author, journal, volume and pages), and in three cases, where the publisher deposited no byline, against PubMed; anything OpenAlex or Crossref flags as retracted was left out, and an editor took out results that matched the words but not the subject. Citation counts are OpenAlex’s on that date and are usually lower than Google Scholar’s, which counts more kinds of document. Ranking by citations tells you what a field has relied on, not what is correct; several heavily cited papers on any topic are cited because later work disputes them. Books without a DOI are missing, which matters in fields where the founding text is a book.
Try Wonders for free