Enhancing Meta Learning via Multi-Objective Soft Improvement Functions

Authors: Runsheng Yu, Weiyu Chen, Xinrun Wang, James Kwok

ICLR 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Empirical studies on various machine learning settings demonstrate that the proposed method is efficient, and achieves better performance than the baselines, particularly on improving the performance of the poorly-performing tasks and thus alleviating the compromising phenomenon. In this section, we perform experiments on few-shot regression (Section 5.1), few-shot classification (Section 5.2), and reinforcement learning (Section 5.3).
Researcher Affiliation Academia 1Department of Computer Science and Engineering, Hong Kong University of Science and Technology 2School of Computer Science and Engineering, Nanyang Technological University
Pseudocode Yes Algorithm 1: Soft Improvement Multi-Objective Meta Learning (SIMOL) Algorithms 2 and 3 show the pseudo-codes for SIMOL-based MAML and PN, respectively.
Open Source Code No Our implementations are based on the popular open-source meta-learning library Learn2Learn (Arnold et al., 2020).
Open Datasets Yes In this section, we perform 5-way-1-shot and 5-way-5-shot classification on the mini Image Net (Ravi & Larochelle, 2016) and tiered Image Net data (Ren et al., 2018).
Dataset Splits Yes Following (Finn et al., 2017), we split the mini Image Net dataset into a meta-training set with 64 classes, a meta-validation set with 16 classes, and a meta-testing set with 20 classes. Similarly, as in (Zhou et al., 2019), we split the tiered Image Net dataset into a meta-training set with 351 classes, a meta-validation set with 97 classes, and a meta-test set with 160 classes.
Hardware Specification Yes All experiments are run on a Ge Force RTX 2080 Ti GPU and Intel(R) Xeon(R) CPU E5-2680.
Software Dependencies No Our implementations are based on the popular open-source meta-learning library Learn2Learn (Arnold et al., 2020).
Experiment Setup Yes We use Adam (Kingma & Ba, 2014), with an initial learning rate of 0.01, to update the base learners for 5 steps in the inner-loop. ... The mini-batch size is 16. ... The initial learning rate for the outer loop is 0.001. The learning rates for the meta networks are 0.003 for MAML-based methods and 0.001 for PN-based methods, respectively. The learning rate of the reweighting network is 0.08 for SIMOL and 0.0008 for SIMOL+PN. The mini-batch size is 32.