DECN: Evolution Inspired Deep Convolution Network for Black-box Optimization

Liu Penghui, Kai Wu, Jing Liu

Keywords: Learning to Optimize, Black-box Optimization

Abstract: We design a deep evolutionary convolution network (DECN) for continuous black-box optimization to force the random population to move near the optimal solution, which is the goal of population-based optimization, such as evolutionary algorithm and evolutionary strategy. DECN is composed of two modules: convolution-based reasoning module (CRM) and selection module (SM), to move from hand-designed searching strategies to learned searching strategies in population-based optimization. CRM produces a population that is closer to the optimal solution based on the convolution operators, and SM removes poor solutions. We also design a proper loss function to support the training of DECN. The experimental results on unconstrained continuous optimization problems show that DECN can learn searching strategies and surpass population-based baselines. Moreover, DECN obtains good performance when transferred to optimization problems unseen during the training stage. In addition, DECN is friendly to the acceleration with Graphics Processing Units (GPUs) and runs 102 times faster than unaccelerated EA when evolving 32 populations, each containing 6400 individuals.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Optimization (eg, convex and non-convex optimization)