# Active Learning with Multiple Kernels

arXiv: https://arxiv.org/abs/2005.03188 · HTML: https://mlsanity.com/paper/2005.03188/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 7 May 2020.

## Built on by (most cited first, 6 of 6)

- [Tighter Regret Analysis and Optimization of Online Federated Learning](https://mlsanity.com/paper/2205.06491/) (2022)
- [Distributed Online Learning with Multiple Kernels](https://mlsanity.com/paper/2102.12733/) (2021)
- [Decentralized Kernel Ridge Regression Based on Data-Dependent Random Feature](https://mlsanity.com/paper/2405.07791/) (2024)
- [Multiple Kernel-Based Online Federated Learning](https://mlsanity.com/paper/2102.10861/) (2021)
- [Graceful Degradation and Related Fields](https://mlsanity.com/paper/2106.11119/) (2021)
- [Unsupervised Active Learning via Natural Feature Progressive Framework](https://mlsanity.com/paper/2510.04939/) (2025)

## Abstract (arXiv)

Online multiple kernel learning (OMKL) has provided an attractive performance in nonlinear function learning tasks. Leveraging a random feature approximation, the major drawback of OMKL, known as the curse of dimensionality, has been recently alleviated. In this paper, we introduce a new research problem, termed (stream-based) active multiple kernel learning (AMKL), in which a learner is allowed to label selected data from an oracle according to a selection criterion. This is necessary in many real-world applications as acquiring true labels is costly or time-consuming. We prove that AMKL achieves an optimal sublinear regret, implying that the proposed selection criterion indeed avoids unuseful label-requests. Furthermore, we propose AMKL with an adaptive kernel selection (AMKL-AKS) in which irrelevant kernels can be excluded from a kernel dictionary 'on the fly'. This approach can improve the efficiency of active learning as well as the accuracy of a function approximation. Via numerical tests with various real datasets, it is demonstrated that AMKL-AKS yields a similar or better performance than the best-known OMKL, with a smaller number of labeled data.