Unsupervised Model Personalization While Preserving Privacy and Scalability: An Open Problem
By Matthias De Lange, Xu Jia, Sarah Parisot, Ales Leonardis, G. Slabaugh, Tinne Tuytelaars. 2020
26 citations recorded by the source corpus.
This work investigates the task of unsupervised model personalization, adapted to continually evolving, unlabeled local user images. We consider the practical scenario where a high capacity server interacts with a myriad of resource-limited edge devices, imposing strong requirements on scalability and local data privacy. We aim to address this challenge within the continual learning paradigm and provide a novel Dual User-Adaptation framework (DUA) to explore the problem. This framework flexibly disentangles user-adaptation into model personalization on the server and local data regularization on the user device, with desirable properties regarding scalability and privacy constraints. First, on the server, we introduce incremental learning of task-specific expert models, subsequently aggregated using a concealed unsupervised user prior. Aggregation avoids retraining, whereas the user prior conceals sensitive raw user data, and grants unsupervised adaptation. Second, local user-adaptation incorporates a domain adaptation point of view, adapting regularizing batch normalization parameters to the user data. We explore various empirical user configurations with different priors in categories and a tenfold of transforms for MIT Indoor Scene recognition, and classify numbers in a combined MNIST and SVHN setup. Extensive experiments yield promising results for data-driven local adaptation and elicit user priors for server adaptation to depend on the model rather than user data. Hence, although user-adaptation remains a challenging open problem, the DUA framework formalizes a principled foundation for personalizing both on server and user device, while maintaining privacy and scalability.
Subfields: Federated & Distributed Learning, Privacy & Security
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Federated & Distributed Learning
In Privacy & Security
Citations in the corpus
- Federated Class-Incremental Learning (2022, background): “[18] sacrifices the local optimality for rapid convergence, while [12,22] aim to improve the performance of local models.”
- How Well Does Self-Supervised Pre-Training Perform with Streaming Data? (2021, background): “Instead, real-world data are usually accessed in a streaming fashion, e.g., data are generated and collected sequentially chunk by chunk (Delange et al., 2021), or even decentralized and stored in different servers (Lange et al., 2020); such a learning setup is known as sequential training (ST).”
- Rehearsal revealed: The limits and merits of revisiting samples in continual learning (2021, neutral): “In this section we compare rehearsal to pioneering work in this family, Elastic Weight Consolidation [18], inspiring many other works [42, 1, 23, 20].”
- Continual Prototype Evolution: Learning Online from Non-Stationary Data Streams (2020, background): “The main body of works presumes the presence of tasks, which divide the data streams into large discrete subsets, and are indicated to the learner with a task identifier [25, 30, 53, 1, 10, 52].”
- Federated Continual Learning with Weighted Inter-client Transfer (2020, background): “This is an opposite direction from personalized federated learning methods (Fallah et al., 2020; Lange et al., 2020; Deng et al., 2020) which put more emphasis on the performance of local models.”
- A Continual Learning Survey: Defying Forgetting in Classification Tasks (2019, background): “Aljundi et al. [13] suggest unsupervised importance estimation, allowing increased flexibility and online user adaptation as in [21].”