Multi-Objective Meta Learning
Authors: Feiyang YE, Baijiong Lin, Zhixiong Yue, Pengxin Guo, Qiao Xiao, Yu Zhang
NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Empirically, we show the effectiveness of the proposed MOML framework in several meta learning problems, including few-shot learning, domain adaptation, multi-task learning, and neural architecture search. |
| Researcher Affiliation | Academia | 1 Department of Computer Science and Engineering, Southern University of Science and Technology 2 University of Technology Sydney 3 Eindhoven University of Technology 4 Peng Cheng Laboratory |
| Pseudocode | Yes | Algorithm 1 Optimization Algorithm for MOML |
| Open Source Code | Yes | The source code of MOML is available at https://github.com/Baijiong-Lin/MOML. |
| Open Datasets | Yes | Experiments are conducted on two FSL benchmark datasets, mini-Image Net [61] and CUB-200-2011 (referred to as CUB) [62]. ... Experiments are conducted on the Office-31 dataset [48] ... Experiments are conducted on the NYUv2 [53], Office-31 and Office-Home [60] datasets. |
| Dataset Splits | Yes | Here each Di is partitioned into two subsets: the training dataset Dtr i and the validation dataset Dval i ... The entire dataset is split into a training dataset denoted by Dtr and a validation dataset denoted by Dval. |
| Hardware Specification | No | The paper does not specify the types of GPUs, CPUs, or other hardware used to run the experiments. |
| Software Dependencies | No | The paper mentions using 'gradient-based multi-objective optimization methods' and specifically 'Multiple Gradient Descent Algorithm (MGDA) [7]', but it does not provide specific version numbers for any software libraries, frameworks (e.g., PyTorch, TensorFlow), or programming languages used. |
| Experiment Setup | Yes | The Res Net-50 is used as the backbone and we do not use data augmentation. ... Due to page limit, experimental settings and experimental results on the CUB dataset are put in Appendix C. ... Due to page limit, details of baselines and experimental settings are put in Appendix D. ... Due to page limit, the introduction of datasets, details on experimental settings, and experimental results on the Office-31 and Office-Home datasets are put in Appendix E. ... Due to page limit, the introduction of datasets, details on experimental settings, and experimental results are put in Appendix F. |