EMO: Episodic Memory Optimization for Few-Shot Meta-Learning

Yingjun Du, Jiayi Shen, Xiantong Zhen, Cees G.M. Snoek
Proceedings of The 2nd Conference on Lifelong Learning Agents, PMLR 232:1-20, 2023.

Abstract

Few-shot meta-learning presents a challenge for gradient descent optimization due to the limited number of training samples per task. To address this issue, we propose an episodic memory optimization for meta-learning, we call EMO, which is inspired by the human ability to recall past learning experiences from the brain’s memory. EMO retains the gradient history of past experienced tasks in external memory, enabling few-shot learning in a memory-augmented way. By learning to retain and recall the learning process of past training tasks, EMO nudges parameter updates in the right direction, even when the gradients provided by a limited number of examples are uninformative. We prove theoretically that our algorithm converges for smooth, strongly convex objectives. EMO is generic, flexible, and model-agnostic, making it a simple plug-and-play optimizer that can be seamlessly embedded into existing optimization-based few-shot meta-learning approaches. Empirical results show that EMO scales well with most few-shot classification benchmarks and improves the performance of optimization-based meta-learning methods, resulting in accelerated convergence.

Cite this Paper

BibTeX
@InProceedings{pmlr-v232-du23a, title = {EMO: Episodic Memory Optimization for Few-Shot Meta-Learning}, author = {Du, Yingjun and Shen, Jiayi and Zhen, Xiantong and Snoek, Cees G.M.}, booktitle = {Proceedings of The 2nd Conference on Lifelong Learning Agents}, pages = {1--20}, year = {2023}, editor = {Chandar, Sarath and Pascanu, Razvan and Sedghi, Hanie and Precup, Doina}, volume = {232}, series = {Proceedings of Machine Learning Research}, month = {22--25 Aug}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v232/du23a/du23a.pdf}, url = {https://proceedings.mlr.press/v232/du23a.html}, abstract = {Few-shot meta-learning presents a challenge for gradient descent optimization due to the limited number of training samples per task. To address this issue, we propose an episodic memory optimization for meta-learning, we call EMO, which is inspired by the human ability to recall past learning experiences from the brain’s memory. EMO retains the gradient history of past experienced tasks in external memory, enabling few-shot learning in a memory-augmented way. By learning to retain and recall the learning process of past training tasks, EMO nudges parameter updates in the right direction, even when the gradients provided by a limited number of examples are uninformative. We prove theoretically that our algorithm converges for smooth, strongly convex objectives. EMO is generic, flexible, and model-agnostic, making it a simple plug-and-play optimizer that can be seamlessly embedded into existing optimization-based few-shot meta-learning approaches. Empirical results show that EMO scales well with most few-shot classification benchmarks and improves the performance of optimization-based meta-learning methods, resulting in accelerated convergence.} }
Endnote
%0 Conference Paper %T EMO: Episodic Memory Optimization for Few-Shot Meta-Learning %A Yingjun Du %A Jiayi Shen %A Xiantong Zhen %A Cees G.M. Snoek %B Proceedings of The 2nd Conference on Lifelong Learning Agents %C Proceedings of Machine Learning Research %D 2023 %E Sarath Chandar %E Razvan Pascanu %E Hanie Sedghi %E Doina Precup %F pmlr-v232-du23a %I PMLR %P 1--20 %U https://proceedings.mlr.press/v232/du23a.html %V 232 %X Few-shot meta-learning presents a challenge for gradient descent optimization due to the limited number of training samples per task. To address this issue, we propose an episodic memory optimization for meta-learning, we call EMO, which is inspired by the human ability to recall past learning experiences from the brain’s memory. EMO retains the gradient history of past experienced tasks in external memory, enabling few-shot learning in a memory-augmented way. By learning to retain and recall the learning process of past training tasks, EMO nudges parameter updates in the right direction, even when the gradients provided by a limited number of examples are uninformative. We prove theoretically that our algorithm converges for smooth, strongly convex objectives. EMO is generic, flexible, and model-agnostic, making it a simple plug-and-play optimizer that can be seamlessly embedded into existing optimization-based few-shot meta-learning approaches. Empirical results show that EMO scales well with most few-shot classification benchmarks and improves the performance of optimization-based meta-learning methods, resulting in accelerated convergence.
APA
Du, Y., Shen, J., Zhen, X. & Snoek, C.G.. (2023). EMO: Episodic Memory Optimization for Few-Shot Meta-Learning. Proceedings of The 2nd Conference on Lifelong Learning Agents, in Proceedings of Machine Learning Research 232:1-20 Available from https://proceedings.mlr.press/v232/du23a.html.

Related Material