Fed$n$nP: Federated Unlearning With Multiple Client Set Partitions
Bibliographic record
Abstract
Federated learning (FL) has garnered increased attention in the field of distributed machine learning and privacy computing. In the FL setup, effective and efficient unlearning algorithms are required to remove the impact of specific training data from the trained model, called federated unlearning. However, traditional machine unlearning algorithms face limitations in FL systems because the client data is private and even non-IID. In this paper, we propose a new federated unlearning algorithm called Fed<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i>P. Our approach involves dividing the client set into subsets using multiple different partitions. We then train constituent models for each client subset within these partitions using existing FL algorithms and aggregate the results of constituent models for predictions. With multiple partitions, Fed<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i>P limits the influence of the data to be erased within its belonging subsets, while it also improves the accuracy of the aggregated prediction. Based on the multiple-partition framework, we design partition creation methods to effectively enhance the prediction accuracy. Furthermore, we propose a cost reduction method to reduce the cost of training/retraining. Our extensive experiments on various datasets and model architectures demonstrate that Fed<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i>P improves prediction accuracy while well-controls the additional cost.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".