Simulated annealing for optimization of graphs and sequences

Optimization of discrete structures aims at generating a new structure with the better property given an existing one, which is a fundamental problem in machine learning. Different from the continuous optimization, the realistic applications of discrete optimization (e.g., text generation) are very challenging due to the complex and long-range constraints, including both syntax and semantics, in discrete structures. In this work, we present SAGS, a novel Simulated Annealing framework for Graph and Sequence optimization. The key idea is to integrate powerful neural networks into metaheuristics (e.g., simulated annealing, SA) to restrict the search space in discrete optimization. We start by defining a sophisticated objective function, involving the property of interest and pre-defined constraints (e.g., grammar validity). SAGS searches from the discrete space towards this objective by performing a sequence of local edits, where deep generative neural networks propose the editing content and thus can control the quality of editing. We evaluate SAGS on paraphrase generation and molecule generation for sequence optimization and graph optimization, respectively. Extensive results show that our approach achieves state-of-the-art performance compared with existing paraphrase generation methods in terms of both automatic and human evaluations. Further, SAGS also significantly outperforms all the previous methods in molecule generation.? † These authors contributed equally to this work. ‡ This paper is an extension to Liu et al. (2020b), published at ACL 2020. There is more than 40% new material, including graph optimization algorithms and several experiments on molecule generation. The code of our work is available at: https://github.com/liuxg16/UPSA ∗ To whom correspondence should be addressed. Email: songsen@mail.tsinghua.edu.cn ? This article is an accepted manuscript of Neurocomputing and under the CC-BY-NC-ND license. The formal publication of this manuscript is: Xianggen Liu, Pengyong Li, Fandong Meng, Hao Zhou, Huasong Zhong, Jie Zhou, Lili Mou, Sen Song. (2021). Simulated annealing for optimization of graphs and sequences. Neurocomputing, 465:310-324. https://doi.org/10.1016/j.neucom.2021.09.003 1 ar X iv :2 11 0. 01 38 4v 1 [ cs .L G ] 1 O ct 2 02 1