Sign-MAML: Efficient Model-Agnostic Meta-Learning by SignSGD
This work addresses the computational bottleneck in meta-learning for researchers and practitioners, offering a more efficient method for few-shot learning, though it is incremental as it builds on existing MAML frameworks.
The authors tackled the computational inefficiency of Model-Agnostic Meta-Learning (MAML) by proposing Sign-MAML, a first-order algorithm that uses signSGD for bilevel optimization, resulting in improved accuracy-efficiency tradeoffs in few-shot image classification tasks.
We propose a new computationally-efficient first-order algorithm for Model-Agnostic Meta-Learning (MAML). The key enabling technique is to interpret MAML as a bilevel optimization (BLO) problem and leverage the sign-based SGD(signSGD) as a lower-level optimizer of BLO. We show that MAML, through the lens of signSGD-oriented BLO, naturally yields an alternating optimization scheme that just requires first-order gradients of a learned meta-model. We term the resulting MAML algorithm Sign-MAML. Compared to the conventional first-order MAML (FO-MAML) algorithm, Sign-MAML is theoretically-grounded as it does not impose any assumption on the absence of second-order derivatives during meta training. In practice, we show that Sign-MAML outperforms FO-MAML in various few-shot image classification tasks, and compared to MAML, it achieves a much more graceful tradeoff between classification accuracy and computation efficiency.