Proceedings of the Genetic and Evolutionary Computation Conference Companion最新文献

Scaling Policy Gradient Quality-Diversity with Massive Parallelization via Behavioral Variations 基于行为变化的大规模并行化分级策略梯度质量多样性

Proceedings of the Genetic and Evolutionary Computation Conference Companion Pub Date : 2025-07-08 DOI: 10.1145/3712256.3726335

Konstantinos Mitsides, Maxence Faldor, Antoine Cully

{"title":"Scaling Policy Gradient Quality-Diversity with Massive Parallelization via Behavioral Variations","authors":"Konstantinos Mitsides, Maxence Faldor, Antoine Cully","doi":"10.1145/3712256.3726335","DOIUrl":"https://doi.org/10.1145/3712256.3726335","url":null,"abstract":"Quality-Diversity optimization comprises a family of evolutionary algorithms aimed at generating a collection of diverse and high-performing solutions. MAP-Elites (ME), a notable example, is used effectively in fields like evolutionary robotics. However, the reliance of ME on random mutations from Genetic Algorithms limits its ability to evolve high-dimensional solutions. Methods proposed to overcome this include using gradient-based operators like policy gradients or natural evolution strategies. While successful at scaling ME for neuroevolution, these methods often suffer from slow training speeds, or difficulties in scaling with massive parallelization due to high computational demands or reliance on centralized actor-critic training. In this work, we introduce a fast, sample-efficient ME based algorithm capable of scaling with massive parallelization, significantly reducing runtimes without compromising performance. Our method, ASCII-ME, unlike existing policy gradient quality-diversity methods, does not rely on centralized actor-critic training. It performs behavioral variations based on time step performance metrics and maps these variations to solutions using policy gradients. Our experiments show that ASCII-ME can generate a diverse collection of high-performing deep neural network policies in less than 250 seconds on a single GPU. Additionally, it operates on average, five times faster than state-of-the-art algorithms while maintaining competitive sample efficiency.","PeriodicalId":216331,"journal":{"name":"Proceedings of the Genetic and Evolutionary Computation Conference Companion","volume":"1 1","pages":"1282-1290"},"PeriodicalIF":0.0,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147904117","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

Enhancing MAP-Elites with Multiple Parallel Evolution Strategies 多并行进化策略增强地图精英

Proceedings of the Genetic and Evolutionary Computation Conference Companion Pub Date : 2024-07-08 DOI: 10.1145/3638529.3654089

Manon Flageat, Bryan Lim, Antoine Cully

{"title":"Enhancing MAP-Elites with Multiple Parallel Evolution Strategies","authors":"Manon Flageat, Bryan Lim, Antoine Cully","doi":"10.1145/3638529.3654089","DOIUrl":"https://doi.org/10.1145/3638529.3654089","url":null,"abstract":"With the development of fast and massively parallel evaluations in many domains, Quality-Diversity (QD) algorithms, that already proved promising in a large range of applications, have seen their potential multiplied. However, we have yet to understand how to best use a large number of evaluations as using them for random variations alone is not always effective. High-dimensional search spaces are a typical situation where random variations struggle to effectively search. Another situation is uncertain settings where solutions can appear better than they truly are and naively evaluating more solutions might mislead QD algorithms. In this work, we propose MAP-Elites-Multi-ES (MEMES), a novel QD algorithm based on Evolution Strategies (ES) designed to exploit fast parallel evaluations more effectively. MEMES maintains multiple (up to ~ 100) simultaneous ES processes, each with its own independent objective and reset mechanism designed for QD optimisation, all on just a single GPU. We show that MEMES outperforms both gradient-based and mutation-based QD algorithms on black-box optimisation and QD-Reinforcement-Learning tasks, demonstrating its benefit across domains. Additionally, our approach outperforms sampling-based QD methods in uncertain domains when given the same evaluation budget. Overall, MEMES generates reproducible solutions that are high-performing and diverse through large-scale ES optimisation on easily accessible hardware.","PeriodicalId":216331,"journal":{"name":"Proceedings of the Genetic and Evolutionary Computation Conference Companion","volume":"1 1","pages":"1082-1090"},"PeriodicalIF":0.0,"publicationDate":"2024-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147895250","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

Empirical linkage learning for non-binary discrete search spaces in the optimization of a large-scale real-world problem 非二元离散搜索空间在大规模现实问题优化中的经验链接学习

Proceedings of the Genetic and Evolutionary Computation Conference Companion Pub Date : 2022-07-09 DOI: 10.1145/3520304.3534059

M. Przewozniczek, M. Komarnicki

引用次数: 0

On the impact of body material properties on neuroevolution for embodied agents: the case of voxel-based soft robots 身体材料特性对具身代理神经进化的影响:以体素为基础的软体机器人为例

Proceedings of the Genetic and Evolutionary Computation Conference Companion Pub Date : 2022-07-09 DOI: 10.1145/3520304.3533967

E. Medvet, Giorgia Nadizar, F. Pigozzi

{"title":"On the impact of body material properties on neuroevolution for embodied agents: the case of voxel-based soft robots","authors":"E. Medvet, Giorgia Nadizar, F. Pigozzi","doi":"10.1145/3520304.3533967","DOIUrl":"https://doi.org/10.1145/3520304.3533967","url":null,"abstract":"Artificial agents required to perform non-trivial tasks are commonly controlled with Artificial Neural Networks (ANNs), which need to be carefully fine-tuned. This is where ANN optimization comes into play, often in the form of Neuroevolution (NE). Among artificial agents, the embodied ones are characterized by a strong body-brain entanglement, i.e., a strong interdependence between the physical properties of the body and the controller. In this work, we aim at characterizing said interconnection, experimentally evaluating the impact body material properties have on NE for embodied agents. We consider the case of Voxel-based Soft Robots (VSRs), a class of simulated modular soft robots which achieve movement through the rhythmical contraction and expansion of their modules. We experiment varying several physical properties of VSRs and assess the effectiveness of the evolved controllers for the task of locomotion, together with their robustness and adaptability. Our results confirm the existence of a deep body-brain interrelationship for embodied agents, and highlight how NE fruitfully exploits the physical properties of the agents to give rise to a wide gamut of effective and adaptable behaviors.","PeriodicalId":216331,"journal":{"name":"Proceedings of the Genetic and Evolutionary Computation Conference Companion","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124392946","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 3

Cooperative attack-defense evolution of large-scale agents: a multi-population high-dimensional mean-field game approach 大规模智能体的协同攻防进化:多种群高维平均场博弈方法

Proceedings of the Genetic and Evolutionary Computation Conference Companion Pub Date : 2022-07-09 DOI: 10.1145/3520304.3528912

Guofang Wang, Xiao Zhang, Wang Yao, Lu Ren

{"title":"Cooperative attack-defense evolution of large-scale agents: a multi-population high-dimensional mean-field game approach","authors":"Guofang Wang, Xiao Zhang, Wang Yao, Lu Ren","doi":"10.1145/3520304.3528912","DOIUrl":"https://doi.org/10.1145/3520304.3528912","url":null,"abstract":"The traditional optimization and control technologies deal with the dynamic interactions between individuals separately, with the increase in the agents' number, the modeling process of cooperative attack-defense problems tends to be complex, and the difficulty of solving the optimal strategy will increase significantly. Moreover, to carry out more accurate real-time control of agents, the state variables used to characterize their kinematics are usually high-dimensional. To overcome these challenges, we formulate the cooperative attack-defense evolution of large-scale agents as a multi-population high-dimensional stochastic mean-field game (MPHD-MFG). Numerical methods for MPHD-MFGs are practically non-existent, because, the heterogeneity of the multi-population model increases the complexity of sequential games, and grid-based spatial discretization leads to dimension explosion. Thus, we propose a generative adversarial network-based method, where we use a coupled alternating neural network composed of multiple generators and multiple discriminators, to tractably solve MPHD-MFGs. Simulation experiments are carried out for various attack-defense scenarios, the results verify the feasibility and effectiveness of our proposed model and algorithm.","PeriodicalId":216331,"journal":{"name":"Proceedings of the Genetic and Evolutionary Computation Conference Companion","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117148193","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 3

Coevolutionary computation for adversarial deep learning 对抗性深度学习的协同进化计算

Proceedings of the Genetic and Evolutionary Computation Conference Companion Pub Date : 2022-07-09 DOI: 10.1145/3520304.3533651

J. Toutouh, Una-May O’Reilly

引用次数: 0

A genetic algorithm for classifying metagenomic data 一种宏基因组数据分类的遗传算法

Proceedings of the Genetic and Evolutionary Computation Conference Companion Pub Date : 2022-07-09 DOI: 10.1145/3520304.3533944

Jolanta Kawulok, M. Kawulok

{"title":"A genetic algorithm for classifying metagenomic data","authors":"Jolanta Kawulok, M. Kawulok","doi":"10.1145/3520304.3533944","DOIUrl":"https://doi.org/10.1145/3520304.3533944","url":null,"abstract":"The goal of metagenomic analysis is to extract relevant information concerning the organisms that have left their genetic traces in an environmental sample. Each sample is subject to nucleotide sequencing, and obtained DNA fragments are decomposed into k-mers---short sequences of k nucleotides. Based on the found k-mers and their occurrence frequencies, it is possible to identify the organisms present in the sample---this allows for further analysis, but requires using large taxonomic datasets. Alternatively, depending on the specific goal of the analysis, the whole sample may be classified directly based on its k-mer profile. However, this is challenging due to a large number of possible k-mers, and choosing the most valuable ones remains an open research problem. In this paper, we propose a new technique that exploits a genetic algorithm for selecting a subset of k-mer features that are used for classification. We report our initial, yet promising results obtained for the problem of detecting the type 2 diabetes from human gut metagenomic samples. We expect that the proposed classification framework will enhance the capabilities of metagenomic analysis, without the need for performing costly taxonomic classification.","PeriodicalId":216331,"journal":{"name":"Proceedings of the Genetic and Evolutionary Computation Conference Companion","volume":"103 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117216977","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

GECCO 2022 tutorial on benchmarking multiobjective optimizers 2.0 GECCO 2022多目标优化器2.0基准测试教程

Proceedings of the Genetic and Evolutionary Computation Conference Companion Pub Date : 2022-07-09 DOI: 10.1145/3520304.3533635

D. Brockhoff, Tea Tušar

引用次数: 0

Towards an interactive ranking operator for NSGA-II 面向NSGA-II的交互式排序算子

Proceedings of the Genetic and Evolutionary Computation Conference Companion Pub Date : 2022-07-09 DOI: 10.1145/3520304.3529041

C. Rosa, W. Freire, A. Amaral, T. Colanzi

{"title":"Towards an interactive ranking operator for NSGA-II","authors":"C. Rosa, W. Freire, A. Amaral, T. Colanzi","doi":"10.1145/3520304.3529041","DOIUrl":"https://doi.org/10.1145/3520304.3529041","url":null,"abstract":"Knowing the Decision Makers (DM) preferences during the search process may help the algorithms to find solutions adequate to the DM profile. For this purpose, some interactive approaches allow the DM to evaluate solutions during the search process, rating them with scores. These scores represent the adequacy level of the solutions concerning the DM preferences and should influence the evolution of the search algorithm. Nevertheless, in these interactive approaches, this prioritization is partial and/or specific for a problem domain. In this context, this work aims to propose an initial version of an interactive ranking operator for NSGA-II, whose goal is to support the ranking of solutions considering both the DM preferences and the fitness of the solutions, for any software engineering domain. For validating the proposal, we incorporate this operator in an interactive approach for Product Line Architecture design. The results pointed out that the new ranking operator can properly deal with the DM preferences, giving a greater chance of surviving to those solutions with higher scores without compromising the diversity of solutions.","PeriodicalId":216331,"journal":{"name":"Proceedings of the Genetic and Evolutionary Computation Conference Companion","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123809555","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 2

Ant colony optimization for feature selection via a filter-randomized search heuristic 基于滤波随机搜索启发式的蚁群特征选择优化

Proceedings of the Genetic and Evolutionary Computation Conference Companion Pub Date : 2022-07-09 DOI: 10.1145/3520304.3528817

Alberto Ortega, J. J. Escobar, M. Damas, Andrés Ortiz, Jesús González

引用次数: 0