Article
Computer Science, Artificial Intelligence
Thomas H. W. Back, Anna V. Kononova, Bas van Stein, Hao Wang, Kirill A. Antonov, Roman T. Kalkreuth, Jacob de Nobel, Diederick Vermetten, Roy de Winter, Furong Ye
Summary: This article discusses some major developments in the field of evolutionary algorithms over the past 30 years, including covariance matrix adaptation evolution strategy, multimodal optimization, surrogate-assisted optimization, multiobjective optimization, and automated algorithm design. The article emphasizes the need for fewer algorithms and proper benchmarking procedures to determine the usefulness of newly proposed algorithms.
EVOLUTIONARY COMPUTATION
(2023)
Article
Computer Science, Theory & Methods
Rui Han, Shilin Li, Xiangwei Wang, Chi Harold Liu, Gaofeng Xin, Lydia Y. Chen
Summary: Research shows that with the exponential growth of data generated by edge computing, the decentralized and Gossip-based training of deep learning models is gaining momentum. The EdgeGossip framework is designed to reduce the performance variation among heterogeneous edge platforms during training and achieve best possible model accuracy quickly. Implementing EdgeGossip based on popular Gossip algorithms has demonstrated an average reduction of model training time by 2.70 times with only 0.78% accuracy loss.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2021)
Article
Computer Science, Theory & Methods
Zhixing Yu, Kejing He, Xiuhong Zou
Summary: The paper introduces a novel distributed pool evolutionary algorithm model PEAB, which addresses the issues of the classical Pool Model through buffer setting, Reunion mechanism, and MP strategy. Experimental results demonstrate that PEAB has a faster convergence rate and stronger population control compared to EvoSpace.
PARALLEL COMPUTING
(2021)
Article
Mathematics
Faisal Altaf, Ching-Lung Chang, Naveed Ishtiaq Chaudhary, Muhammad Asif Zahoor Raja, Khalid Mehmood Cheema, Chi-Min Shu, Ahmad H. Milyani
Summary: The study utilized evolutionary and swarm computing paradigms to address the overparameterization issue in parameter estimation for nonlinear systems. By integrating the key term separation principle and genetic algorithms, the proposed approach effectively estimated the actual parameters of Hammerstein control autoregressive systems.
Article
Automation & Control Systems
Xiao-Qi Guo, Wei-Neng Chen, Feng-Feng Wei, Wen-Tao Mao, Xiao-Min Hu, Jun Zhang
Summary: Surrogate-assisted evolutionary algorithms have been proposed to solve data-driven optimization problems. However, most existing methods do not consider the challenges brought by the distribution of data at the edge of networks in the era of the Internet of Things. In this study, we propose edge-cloud co-EAs to address distributed data-driven optimization problems, where data are collected by edge servers.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Computer Science, Information Systems
Giovanni Acampora, Autilia Vitiello
Summary: This study introduces a new evolutionary algorithm utilizing an actual quantum processor, which employs quantum phenomena to achieve significant speed-up in computation. By implementing quantum concepts such as quantum chromosome and entangled crossover, the proposed algorithm efficiently executes genetic evolution on quantum devices to converge towards proper sub-optimal solutions of a given optimization problem. The experimental results show that the synergy between quantum and evolutionary computation leads to a promising bio-inspired optimization strategy.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Hardware & Architecture
Wenjie Zhu, Wufei Wu, Xingyu Yang, Gang Zeng
Summary: This study proposes a structure-aware task scheduling strategy for parallel application scheduling in heterogeneous distributed embedded systems, along with an improved energy pre-allocation algorithm. Experimental results demonstrate that the algorithm can reduce task scheduling length and energy consumption.
JOURNAL OF SYSTEMS ARCHITECTURE
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Mario Garcia-Valdez, Rene Marquez, Leonardo Trujillo, J. J. Merelo
Summary: The researchers studied how asynchronous distributed evolutionary algorithms solve the issue of non-synchronous nodes by dropping homogeneity and synchronicity assumption, and found that randomly varying parameters can impact the performance of the algorithm.
2021 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC 2021)
(2021)
Article
Computer Science, Information Systems
Ravi Reddy Manumachu, Hamidreza Khaleghzadeh, Alexey Lastovetsky
Summary: Accelerating the bi-objective optimization of applications for performance and energy is crucial. This work highlights the challenges of accelerating model-based methods for bi-objective optimization of data parallel applications. The proposed algorithms provide significant speedups over state-of-the-art solutions.
Article
Computer Science, Information Systems
Dapeng Lan, Amir Taherkordi, Frank Eliassen, Lei Liu, Stephane Delbruel, Schahram Dustdar, Yang Yang
Summary: This article introduces a system framework, EDGE VISION, for computer vision applications on heterogeneous edge computing platforms. It proposes two scheduling algorithms, minimum latency task scheduling and minimum cost task scheduling, to minimize processing latency and system cost.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Theory & Methods
Enda Yu, Dezun Dong, Xiangke Liao
Summary: This paper proposes a standard for systematically classifying communication optimization algorithms in distributed deep learning systems based on mathematical modeling, which is a novel contribution in the field. The authors categorize existing works into four categories based on communication optimization strategies and discuss potential future challenges and research directions.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Giovanni Acampora, Ferdinando Di Martino, Alfredo Massa, Roberto Schiattarella, Autilia Vitiello
Summary: This paper introduces the concept of Distributed Noisy-Intermediate Scale Quantum (D-NISQ) as a reference computational model to design innovative frameworks for quantum devices to interact and solve complex problems collaboratively. Through two case studies, a multi-threaded implementation of the D-NISQ model demonstrates greater reliability in solving problems through quantum computation.
INFORMATION FUSION
(2023)
Article
Computer Science, Artificial Intelligence
Jesus Guillermo Falcon-Cardona, Raquel Hernandez Gomez, Carlos A. Coello Coello, Ma. Guadalupe Castillo Tapia
Summary: This paper presents a survey of parallel implementations of multi-objective evolutionary algorithms (pMOEAs), discussing their significance in tackling computationally expensive applications, describing taxonomy and methods review, and proposing open questions for further research.
SWARM AND EVOLUTIONARY COMPUTATION
(2021)
Article
Computer Science, Theory & Methods
Sultan Alamro, Tian Lan, Suresh Subramaniam
Summary: Data-intensive computing frameworks divide job workload into fixed-size chunks for parallel processing on distributed machines. However, the variability and uncertainty in processing time can lead to performance degradation. This paper proposes Forseti, a processing scheme that dynamically adjusts data chunk size based on machine heterogeneity and dynamic execution environment. Forseti also utilizes virtual machine reuse to reduce startup and initialization costs. Experimental results demonstrate significant performance improvements of Forseti compared to other baselines in terms of job completion time.
JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING
(2023)
Article
Computer Science, Theory & Methods
Feng Li, Fengguang Song
Summary: Optimizing deployment plans for in-situ workflows in geographically distributed heterogeneous computing environments is challenging. This study presents a heuristic-based solver using the SNL algorithm, which produces effective deployment plans with significantly reduced problem-solving time compared to the CPLEX optimization method.
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE
(2023)
Article
Computer Science, Software Engineering
Leonardo Trujillo, Jose Manuel Munoz Contreras, Daniel E. Hernandez, Mauro Castelli, Juan J. Tapia
Summary: Geometric Semantic Genetic Programming (GSGP) is an efficient machine learning method based on evolutionary computation, which performs search operations at the level of program semantics. This paper introduces GSGP-CUDA, the first CUDA implementation of GSGP that exploits the parallelism of GPUs, resulting in significant speedups during the model training process. Additionally, the implementation allows seamless inference over new data using the best evolved model.
Article
Computer Science, Artificial Intelligence
Cristian Sandoval, Oliver Cuate, Luis C. Gonzalez, Leonardo Trujillo, Oliver Schutze
Summary: In this study, a regression-based approach using Genetic Programming is proposed to approximate the hypervolume (HV) value and improve computational efficiency. The approach achieves low errors and high correlation in multiple-objective problems, and demonstrates significantly faster computation compared to standard methods.
APPLIED SOFT COMPUTING
(2022)
Editorial Material
Computer Science, Artificial Intelligence
Gustavo Olague, Mario Koppen, Oscar Cordon
Summary: This article introduces the field of Evolutionary Computer Vision (ECV), which is at the intersection of computer vision (CV) and evolutionary computation (EC). ECV utilizes evolutionary algorithms and metaheuristic approaches combined with analytical methods to achieve human-competitive results. It aims to design software and hardware solutions for challenging CV problems and enhance our understanding of visual processing in nature.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2023)
Article
Automation & Control Systems
Debbrota Paul Chowdhury, Sambit Bakshi, Chiara Pero, Gustavo Olague, Pankaj Kumar Sa
Summary: This article introduces an Industry 4.0 compliant ear biometric recognition method based on DenseNet. Compared to other biometric traits, ear recognition has been challenging due to limited images and the potential of deep learning is still unexplored. The proposed DenseNet achieves state-of-the-art results on challenging benchmarks and popular ear databases, showing better performance than existing methods. With fewer parameters and fast processing, this method can ensure privacy preservation over the Internet of Biometric Things.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Multidisciplinary Sciences
Mario Garcia-Valdez, Alejandra Mancilla, Oscar Castillo, Juan Julian Merelo-Guervos
Summary: In this work, a distributed and asynchronous bio-inspired algorithm is proposed to speed up the design process of a controller by executing simulations in parallel. The algorithm uses a multi-population multi-algorithmic approach with isolated populations interacting asynchronously using a distributed message queue. The results demonstrate the speedup benefit of the proposed algorithm and the advantages of mixing populations with distinct metaheuristics.
Article
Education & Educational Research
Juan J. Merelo, Pedro A. Castillo, Antonio M. Mora, Francisco Barranco, Noorhan Abbas, Alberto Guillen, Olia Tsivitanidou
Summary: This article examines the application of messaging platforms in higher education and the experiences and perceptions of teachers. A survey was conducted to gather teachers' preferences and opinions on messaging platforms and chatbots, as well as their views on how these tools can enhance student learning. The survey provides insights into teachers' needs and the various educational use cases where these tools could be valuable. The analysis also explores how teachers' opinions on tool usage vary based on gender, experience, and specialization. The key findings emphasize the factors that contribute to the adoption of messaging platforms and chatbots in higher education institutions to achieve desired learning outcomes.
EDUCATION AND INFORMATION TECHNOLOGIES
(2023)
Article
Computer Science, Information Systems
Rocio Ochoa-Montiel, Humberto Sossa, Gustavo Olague, Carlos Sanchez-Lopez
Summary: An evolutionary vision approach is used for the automatic recognition of AML leukemia images in this study. Unlike common approaches, the feature extraction process in the presented model is transparent, and the obtained solutions are interpretable by human users.
COMPUTACION Y SISTEMAS
(2023)
Article
Computer Science, Artificial Intelligence
Leonardo Trujillo, Joel Nation, Luis Munoz, Edgar Galvan
Summary: This study proposes a novel method to determine the compatibility of two problems for transfer learning, and for the first time, studies within genetic programming. By comparing the feature space representations of problems, a similarity measure is computed, and the results show significant distinction between compatible and non-compatible problems for transfer learning.
Article
Multidisciplinary Sciences
Gerardo Ibarra-Vazquez, Maria Soledad Ramirez-Montoya, Mariana Buenestado-Fernandez, Gustavo Olague
Summary: This study used machine learning models to analyze open education competency data and predict the competency levels based on students' perceptions of knowledge, skills, and attitudes related to open education. The results showed that students' perceptions provided satisfactory data for building machine learning models to predict competency levels.
Article
Mathematics, Interdisciplinary Applications
Enrique Naredo, Candelaria Sansores, Flaviano Godinez, Francisco Lopez, Paulo Urbano, Leonardo Trujillo, Conor Ryan
Summary: Robotics technology has made significant advancements in various fields, particularly in manufacturing and navigation. This research aims to explore how training scenarios affect the learning process for autonomous navigation tasks, with a focus on whether the initial conditions have a positive or negative impact on developing general controllers. The study aims to optimize the training process and improve the quality of autonomous navigation controllers.
MATHEMATICAL AND COMPUTATIONAL APPLICATIONS
(2023)
Article
Automation & Control Systems
J. Enriquez-Zarate, S. Gomez-Penate, C. Hernandez, Francisco Villarreal-Valderrama, R. Velazquez, Leonardo Trujillo
Summary: This article presents the design of a nonlinear hybrid controller for an underactuated Duffing oscillator with 2 degrees of freedom. The controller aims to reduce the frequency-response to specific resonant-frequencies while maintaining its robustness to external disturbances. Simulation results show that the proposed control scheme can reduce the system's response to external vibrations up to 83.88%.
OPTIMAL CONTROL APPLICATIONS & METHODS
(2023)
Article
Computer Science, Information Systems
Dalia A. Rodriguez, Julia Diaz-Escobar, Arnoldo Diaz-Ramirez, Leonardo Trujillo
Summary: Violence against women is a significant social issue, and social media contains a large amount of misogynistic content. This study introduces a BERT architecture to automatically detect misogynistic tweets in Spanish, achieving good results. Manual error analysis revealed misogynistic bias in the dataset, and a debiased model outperformed existing literature on misogyny detection.
SOCIAL NETWORK ANALYSIS AND MINING
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Alejandra Mancilla, Oscar Castillo, Mario Garcia Valdez
Summary: In this work, a distributed platform is proposed to execute multi-population metaheuristics. Two metaheuristics, Genetic Algorithms and Particle Swarm Optimization, are used as proof of concept. The algorithms are implemented asynchronously using a queue-based architecture. The study optimizes the parameters of a fuzzy controller and demonstrates the benefits of mixing algorithm populations and integrating migration processes.
INTELLIGENT AND FUZZY SYSTEMS: DIGITAL ACCELERATION AND THE NEW NORMAL, INFUS 2022, VOL 1
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Rocio Ochoa-Montiel, Humberto Sossa, Gustavo Olague, Carlos Sanchez-Lopez
Summary: This study analyzes the performance of three commonly used classifiers in the brain programming symbolic learning model, showing that MLP and SVM classifiers are robust to noisy data, with MLP demonstrating the most stable behavior in the symbolic learning model.
PATTERN RECOGNITION, MCPR 2022
(2022)
