Efficient Meta Learning via Minibatch Proximal Update

Authors: Pan Zhou, Xiaotong Yuan, Huan Xu, Shuicheng Yan, Jiashi Feng

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on several few-shot regression and classification tasks demonstrate the advantages of our method over state-of-the-arts.
Researcher Affiliation Collaboration Learning & Vision Lab, National University of Singapore, Singapore B-DAT Lab, Nanjing University of Information Science & Technology, Nanjing, China Alibaba and Georgia Institute of Technology, USA YITU Technology, Shanghai, China
Pseudocode Yes Algorithm 1 SGD for Meta-Minibatch Prox
Open Source Code Yes The code is available at https://panzhous.github.io.
Open Datasets Yes mini Image Net [5] and tiered Image Net [43]
Dataset Splits Yes Following [6, 10], we use the split proposed in [5], which consists of 64 classes for training, 16 classes for validation and the remaining 20 classes for testing.
Hardware Specification No The paper does not provide specific details about the hardware used for running experiments.
Software Dependencies No The paper mentions optimizers like SGD and Adam, but does not provide specific version numbers for any software libraries or dependencies used in the experiments.
Experiment Setup Yes For our Meta-Minibatch Prox, we set λ = 0.5 and use SGD to solve the inner subproblem with 15 steps of iteration with learning rate 0.02. For the learning rate ηs in Meta-Minibatch Prox, we decrease it at each iteration as ηs = α(1 s/S) where the total iteration number S in Algorithm 1 and α are set to 30, 000 and 0.8, respectively.