Multi-Objective Reinforced Evolution in Mobile Neural Architecture Search
By Xiangxiang Chu, Ruijun Xu, Hailong Ma. 2019
118 citations recorded by the source corpus.
Fabricating neural models for a wide range of mobile devices demands for a specific design of networks due to highly constrained resources. Both evolution algorithms (EA) and reinforced learning methods (RL) have been dedicated to solve neural architecture search problems. However, these combinations usually concentrate on a single objective such as the error rate of image classification. They also fail to harness the very benefits from both sides. In this paper, we present a new multi-objective oriented algorithm called MoreMNAS (Multi-Objective Reinforced Evolution in Mobile Neural Architecture Search) by leveraging good virtues from both EA and RL. In particular, we incorporate a variant of multi-objective genetic algorithm NSGA-II, in which the search space is composed of various cells so that crossovers and mutations can be performed at the cell level. Moreover, reinforced control is mixed with a natural mutating process to regulate arbitrary mutation, maintaining a delicate balance between exploration and exploitation. Therefore, not only does our method prevent the searched models from degrading during the evolution process, but it also makes better use of learned knowledge. Our experiments conducted in Super-resolution domain (SR) deliver rivalling models compared to some state-of-the-art methods with fewer FLOPS.
Subfields: Computer Vision and Pattern Recognition, Neural Architecture Search & AutoML, Evolutionary Computation
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computer Vision and Pattern Recognition
In Neural Architecture Search & AutoML
In Evolutionary Computation
Citations in the corpus
- GreedyNAS: Towards Fast One-Shot NAS With Greedy Supernet (2020, background): “One remedy is by dint of evolutionary algorithms [11] or reinforced version (e.g., MoreMNAS [5]), which takes the supernet as an off-the-shelf evaluator.”
- Hierarchical Neural Architecture Search for Single Image Super-Resolution (2020, methodology): “To further show the effectiveness of the proposed method, we also conduct visual comparisons between HNAS and three SR methods, such as CARN [31], MoreMNAS [32] and FALSR [33].”
- FairNAS: Rethinking Evaluation Fairness of Weight Sharing Neural Architecture Search (2019, methodology): “We choose a multi-objective NAS method [5] to serve our need.”
- Fast, Accurate and Lightweight Super-Resolution with Neural Architecture Search (2019, methodology): “In addition, it dominates DRCN [Kim et al. , 2016b] and MoreMNAS-A [Chu et al. , 2019] over three objectives on four datasets.”
- Efficient Transformer for Single Image Super-Resolution (2021, methodology): “As for neural architecture search-based methods, MoreMNAS [6] presents a new multi-objective oriented algorithm by leveraging good virtues from both evolution algorithms and reinforced learning methods.”
- A Survey on Evolutionary Neural Architecture Search (2020, background): “[121] preferred that while a crossover mainly contributes to exploitation, a mutation is usually aimed to introduce exploration.”