Practical Implementation of Federated Learning for Detecting Backdoor Attacks in a Next-word Prediction Model
Abstract This article details the development of a next-word prediction model utilizing federated learning and introduces a mechanism for detecting backdoor attacks. Federated learning enables multiple devices to collaboratively train a shared model while retaining data locally. However, this decent...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-024-82079-2 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832594785115832320 |
|---|---|
| author | Jimmy K. W. Wong Ki Ki Chung Yuen Wing Lo Chun Yin Lai Steve W. Y. Mung |
| author_facet | Jimmy K. W. Wong Ki Ki Chung Yuen Wing Lo Chun Yin Lai Steve W. Y. Mung |
| author_sort | Jimmy K. W. Wong |
| collection | DOAJ |
| description | Abstract This article details the development of a next-word prediction model utilizing federated learning and introduces a mechanism for detecting backdoor attacks. Federated learning enables multiple devices to collaboratively train a shared model while retaining data locally. However, this decentralized approach is susceptible to manipulation by malicious actors who control a subset of participating devices, thereby biasing the model’s outputs on specific topics, such as a presidential election. The proposed detection mechanism aims to identify and exclude devices with anomalous datasets from the training process, thereby mitigating the influence of such attacks. By using the example of a presidential election, the study demonstrates a positive correlation between the proportion of compromised devices and the degree of bias in the model’s outputs. The findings indicate that the detection mechanism effectively reduces the impact of backdoor attacks, particularly when the number of compromised devices is relatively low. This research contributes to enhancing the robustness of federated learning systems against malicious manipulation, ensuring more reliable and unbiased model performance. |
| format | Article |
| id | doaj-art-a95d288f9ad7490eb479b8548ac5e633 |
| institution | Kabale University |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-a95d288f9ad7490eb479b8548ac5e6332025-01-19T12:22:38ZengNature PortfolioScientific Reports2045-23222025-01-0115111110.1038/s41598-024-82079-2Practical Implementation of Federated Learning for Detecting Backdoor Attacks in a Next-word Prediction ModelJimmy K. W. Wong0Ki Ki Chung1Yuen Wing Lo2Chun Yin Lai3Steve W. Y. Mung4Research and Development Office, The Education University of Hong KongDepartment of Mathematics and Information Technology, The Education University of Hong KongResearch and Development Office, The Education University of Hong KongResearch and Development Office, The Education University of Hong KongResearch and Development Office, The Education University of Hong KongAbstract This article details the development of a next-word prediction model utilizing federated learning and introduces a mechanism for detecting backdoor attacks. Federated learning enables multiple devices to collaboratively train a shared model while retaining data locally. However, this decentralized approach is susceptible to manipulation by malicious actors who control a subset of participating devices, thereby biasing the model’s outputs on specific topics, such as a presidential election. The proposed detection mechanism aims to identify and exclude devices with anomalous datasets from the training process, thereby mitigating the influence of such attacks. By using the example of a presidential election, the study demonstrates a positive correlation between the proportion of compromised devices and the degree of bias in the model’s outputs. The findings indicate that the detection mechanism effectively reduces the impact of backdoor attacks, particularly when the number of compromised devices is relatively low. This research contributes to enhancing the robustness of federated learning systems against malicious manipulation, ensuring more reliable and unbiased model performance.https://doi.org/10.1038/s41598-024-82079-2 |
| spellingShingle | Jimmy K. W. Wong Ki Ki Chung Yuen Wing Lo Chun Yin Lai Steve W. Y. Mung Practical Implementation of Federated Learning for Detecting Backdoor Attacks in a Next-word Prediction Model Scientific Reports |
| title | Practical Implementation of Federated Learning for Detecting Backdoor Attacks in a Next-word Prediction Model |
| title_full | Practical Implementation of Federated Learning for Detecting Backdoor Attacks in a Next-word Prediction Model |
| title_fullStr | Practical Implementation of Federated Learning for Detecting Backdoor Attacks in a Next-word Prediction Model |
| title_full_unstemmed | Practical Implementation of Federated Learning for Detecting Backdoor Attacks in a Next-word Prediction Model |
| title_short | Practical Implementation of Federated Learning for Detecting Backdoor Attacks in a Next-word Prediction Model |
| title_sort | practical implementation of federated learning for detecting backdoor attacks in a next word prediction model |
| url | https://doi.org/10.1038/s41598-024-82079-2 |
| work_keys_str_mv | AT jimmykwwong practicalimplementationoffederatedlearningfordetectingbackdoorattacksinanextwordpredictionmodel AT kikichung practicalimplementationoffederatedlearningfordetectingbackdoorattacksinanextwordpredictionmodel AT yuenwinglo practicalimplementationoffederatedlearningfordetectingbackdoorattacksinanextwordpredictionmodel AT chunyinlai practicalimplementationoffederatedlearningfordetectingbackdoorattacksinanextwordpredictionmodel AT stevewymung practicalimplementationoffederatedlearningfordetectingbackdoorattacksinanextwordpredictionmodel |