Article
Computer Science, Artificial Intelligence
F. M. Bianchi, Claudio Gallicchio, Alessio Micheli
Summary: A deep Graph Neural Network (GNN) model is proposed, which alternates between two types of layers to achieve a trade-off between computational efficiency and accuracy in graph embedding.
Article
Chemistry, Analytical
Laurent Chiasson-Poirier, Hananeh Younesian, Katia Turcot, Julien Sylvestre
Summary: This study focuses on the numerical implementation of a reservoir computing algorithm called echo state network (ESN) for gait event detection. The results show that ESN is robust and comparable to other state-of-the-art algorithms in various conditions.
Article
Mathematics, Applied
Jongha Jeon, Pilwon Kim, Bongsoo Jang, Yunho Kim
Summary: The study proposes a new method for image denoising, RCPDE, which combines reservoir computing with partial differential equations and outperforms traditional methods in the small data regime.
Article
Multidisciplinary Sciences
Ryan Pyle, Nikola Jovanovic, Devika Subramanian, Krishna V. Palem, Ankit B. Patel
Summary: Recent advancements in computing algorithms and hardware have led to increased interest in developing high-accuracy, low-cost surrogate models for simulating physical systems. The echo state network (ESN) technique has gained popularity within the weather and climate modeling community. A study found that state-of-the-art LSR-ESNs reduce to a polynomial regression model called D2R2, which outperforms other approaches significantly in computational savings.
PHILOSOPHICAL TRANSACTIONS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES
(2021)
Article
Meteorology & Atmospheric Sciences
Balasubramanya T. Nadiga
Summary: Reduced-order dynamical models are crucial in understanding climate predictability, with linear inverse modeling (LIM) and reservoir computing (RC) being valuable for improving predictive skills. RC shows promise in enhancing predictability studies by providing nonlinear approaches, especially in scenarios with limited data.
JOURNAL OF ADVANCES IN MODELING EARTH SYSTEMS
(2021)
Review
Energy & Fuels
Ivan S. S. Maksymov
Summary: This article reviews the recent advances in analogue and reservoir computing driven by the properties and energy of water waves. It suggests that these research areas have the potential to bring artificial intelligence closer to rural areas, allowing them to benefit from novel technologies that are already present in large cities. The physical reservoir computing systems discussed in the article can be used for designing and optimizing power grid networks and forecasting energy consumption at local and global scales. Therefore, this review article is of significant importance for readers interested in the innovative practical applications of artificial intelligence and machine learning.
Article
Physics, Multidisciplinary
Maxime Casanova, Barbara Dalena, Luca Bonaventura, Massimo Giovannozzi
Summary: We investigate the ability of an ensemble reservoir computing approach to predict the long-term behaviour of the phase-space region in hadron storage rings. Echo State Networks (ESN) are computationally effective recurrent neural networks that have been proven to be universal approximants of dynamical systems. Our results show that the proposed ESN approach can effectively predict the time evolution of the dynamic aperture, improving the predictions by analytical scaling laws and providing an efficient surrogate model.
EUROPEAN PHYSICAL JOURNAL PLUS
(2023)
Article
Multidisciplinary Sciences
Jingyao Liu, Jiajia Chen, Guijin Yan, Wengang Chen, Bingyin Xu
Summary: This paper proposes a novel clustering and dynamic recognition-based auto-reservoir neural network (CDbARNN) for short-term load forecasting (STLF) of industrial park microgrids. CDbARNN decomposes load sets into clusters using K-means clustering and uses a dynamic recognition technology to identify the cluster to which an input load series belongs. The input load series and load curves of the corresponding cluster are then used as input to the reservoir of CDbARNN for short-term forecasting.
Article
Environmental Sciences
Yslam D. Mammedov, Ezutah Udoncy Olugu, Guleid A. Farah
Summary: Wind power has become a significant research area in renewable energy as a response to increasing demand for global energy supply chain. The study introduces a weather prediction method which includes two models for wind speed and atmospheric system forecasting. The physics-informed model was found to outperform other methods in accuracy and reliability, demonstrating potential for application in wind energy analysis.
ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH
(2022)
Article
Physics, Fluids & Plasmas
Rodrigo Martinez-Pena, Juan-Pablo Ortega
Summary: This paper explores alternative representations in quantum reservoir computing (QRC) systems and establishes connections with the density matrix approach and representation in the space of observables. It shows that these representations allow for independent discussion of fading memory property (FMP) and echo state property (ESP), and sheds light on fundamental questions in QRC theory in finite dimensions.
Article
Computer Science, Artificial Intelligence
Denis Kleyko, Edward Paxon Frady, Mansour Kheffache, Evgeny Osipov
Summary: In this study, an approximation of echo state networks (ESNs) based on hyperdimensional computing is proposed, which can be efficiently implemented on digital hardware. The proposed intESN replaces the recurrent matrix multiplication with an efficient cyclic shift operation and utilizes a vector containing only a few integers as the reservoir. Experimental results show that the proposed intESN approach is effective and more energy efficient compared to conventional ESNs.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Kiril Simov, Petia Koprinkova-Hristova, Alexander Popov, Petya Osenova
Summary: Reservoir computing (RC) is considered biologically plausible and has been applied to Word Sense Disambiguation (WSD) tasks in natural language processing (NLP). A novel deep bi-directional ESN (DBiESN) structure and a new approach for exploiting reservoirs' steady states were proposed in this research.
COGNITIVE COMPUTATION
(2023)
Article
Computer Science, Artificial Intelligence
Benjamin Paassen, Alexander Schulz, Terrence C. Stewart, Barbara Hammer
Summary: In this work, the authors successfully achieved some of the computational capabilities of DNCs using an echo state network with efficient training. The model can recognize all regular languages and performs comparably to fully trained deep versions on typical benchmark tasks for DNCs.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Laercio de Oliveira Junior, Florian Stelzer, Liang Zhao
Summary: This study investigates the effectiveness of ESNs with clustered adjacency matrices (CESNs) and applies them to signal denoising. The findings suggest that CESNs and deep CESNs can compete with deep ESNs in various tasks.
Article
Statistics & Probability
Lukas Gonon, Lyudmila Grigoryeva, Juan-pablo Ortega
Summary: This work focuses on the approximation capabilities of randomly generated internal weights in single-hidden-layer feed-forward and recurrent neural networks. The results show that as long as the unknown function, functional, or dynamical system is sufficiently regular, it is possible to draw the internal weights of the random neural network from a generic distribution and quantify the error in terms of the number of neurons and hyperparameters. This provides a mathematical explanation for the empirical success of echo state networks in learning dynamical systems.
ANNALS OF APPLIED PROBABILITY
(2023)
Article
Computer Science, Artificial Intelligence
Kiril Simov, Petia Koprinkova-Hristova, Alexander Popov, Petya Osenova
Summary: Reservoir computing (RC) is considered biologically plausible and has been applied to Word Sense Disambiguation (WSD) tasks in natural language processing (NLP). A novel deep bi-directional ESN (DBiESN) structure and a new approach for exploiting reservoirs' steady states were proposed in this research.
COGNITIVE COMPUTATION
(2023)
Article
Instruments & Instrumentation
Bruna D. M. Lopes, Luis C. B. Silva, Isidro M. Blanquet, Petia Georgieva, Carlos A. F. Marques
Summary: Aquaculture is a fundamental sector of the food industry, and the need for monitoring various parameters and using novel anomaly detection methods for prediction and disaster prevention to maintain sustainability and profitability is highlighted in this study.
REVIEW OF SCIENTIFIC INSTRUMENTS
(2021)
Article
Computer Science, Information Systems
Simona Nedelcheva, Sofiya Ivanovska, Mariya Durchova, Petia Koprinkova-Hristova
Summary: The paper introduces a supercomputer parallel implementation of a brain inspired model combining a Python module and NEST Simulator. The study compares simulations on different numbers of nodes and parallel processes to evaluate time consumption of the model.
CLUSTER COMPUTING-THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS
(2022)
Article
Engineering, Electrical & Electronic
Fabio D. L. S. Coutinho, Hugerles S. Silva, Petia Georgieva, Arnaldo S. R. Oliveira
Summary: In this paper, a CNN-based framework is proposed to tackle the problem of cascaded channels estimation. The results show that the CNN-based framework performs well in terms of bit error rate and mean squared error, especially for increasing number of links and modulation order.
DIGITAL SIGNAL PROCESSING
(2022)
Editorial Material
Computer Science, Information Systems
Petia Koprinkova-Hristova, Mirjana Ivanovic, Banu Diri
JOURNAL OF INFORMATION AND TELECOMMUNICATION
(2022)
Article
Materials Science, Multidisciplinary
Ruben Lourenco, Antonio Andrade-Campos, Petia Georgieva
Summary: This study explores using machine learning techniques to improve the accuracy of material constitutive models in metal plasticity, including parameter identification inverse methodology, constitutive model corrector, data-driven constitutive model, and general implicit constitutive model. Training these methods requires a large amount of data on material behavior, necessitating the use of non-homogeneous strain field and complex strain path tests measured with digital image correlation techniques.
Proceedings Paper
Computer Science, Artificial Intelligence
Petia Koprinkova-Hristova, Dimitar Penkov, Simona Nedelcheva, Svetlozar Yordanov, Nikola Kasabov
Summary: The paper introduces a novel hierarchical recurrent neural network architecture for real-time classification and interpretation of EEG data. It combines two dynamic pools of neurons - one based on NeuCube three-dimensional structure of spiking neurons and another Echo state neural network (ESN) reservoir, to classify continuously extracted brain signals. The proposed method shows improved classification accuracy and interpretability of the EEG data.
2023 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, IJCNN
(2023)
Article
Computer Science, Information Systems
Daniel Canedo, Joao Fonte, Luis Goncalves Seco, Marta Vazquez, Rita Dias, Tiago Do Pereiro, Joao Hipolito, Fernando Menendez-Marsh, Petia Georgieva, Antonio J. R. Neves
Summary: Mapping potential archaeological sites using remote sensing and AI is an efficient tool for project planning and fieldwork. This paper discusses the use of LiDAR data and data-centric AI to identify burial mounds. The challenge of exploring the landscape and identifying archaeological sites accurately is addressed by proposing a novel data-centric AI approach that preprocesses LiDAR data, annotates known burial mounds, and uses object embedding techniques for augmentation. The proposed approach achieved a 72.53% positive rate, reducing false positives and aiding archaeologists in the ground-truthing phase.
Proceedings Paper
Engineering, Electrical & Electronic
Fabio D. L. Coutinho, Hugerles S. Silva, Petia Georgieva, Arnaldo Oliveira
Summary: This paper proposes a CNN-based algorithm for joint estimation of channel and phase noise in OFDM relay systems. The algorithm infers intercarrier interference caused by phase noise and investigates the impact of cascaded channels. The results show that the CNN-based approach outperforms practical estimation methods and significantly improves the bit error rate.
2022 IEEE LATIN-AMERICAN CONFERENCE ON COMMUNICATIONS (LATINCOM)
(2022)
Proceedings Paper
Computer Science, Theory & Methods
Petia Koprinkova-Hristova, Simona Nedelcheva, Nadejda Bocheva
Summary: This paper presents in-silico investigations on the effects of brain lesions in different parts of the human visual system. A hierarchical spike timing neural network model was implemented in the NEST simulator to reproduce the performance of visual tasks with reinforcement learning. The model's structure and connectivity were designed based on available information about the organization of corresponding brain areas. The simulations were conducted by feeding the model with dynamic visual stimuli, and the observed deterioration of visual task performance due to different brain lesions were summarized and commented upon.
CONTEMPORARY METHODS IN BIOINFORMATICS AND BIOMEDICINE AND THEIR APPLICATIONS
(2022)
Proceedings Paper
Computer Science, Information Systems
Ricardo Pousa, Petia Georgieva, Jose Pina, Pedro Cruz, Paulo Andre
Summary: This study suggests online monitoring of optical fiber network transmission quality using machine learning, fitting three regression models to estimate current and long-term QoT indicators, and providing real data from online measurable, equipment-agnostic features through realistic network experimental scenarios.
2021 EUROPEAN CONFERENCE ON OPTICAL COMMUNICATION (ECOC)
(2021)
Article
Robotics
Diogo Carneiro, Filipe Silva, Petia Georgieva
Summary: This study introduces a Robot Anticipation Learning System that anticipates the trajectory of a flying ball by observing the thrower's hand motion, improving the catch rate by up to 20% compared to traditional methods that rely solely on information acquired during the flight phase.
Article
Engineering, Multidisciplinary
Daniel Canedo, Pedro Fonseca, Petia Georgieva, Antonio J. R. Neves
Summary: Floor-cleaning robots are equipped with advanced vision systems using the YOLOv5 framework for spot detection, along with a synthetic data generator to address the lack of real data, improving the models' ability to differentiate between dirt spots and other objects. The models achieved a mean average precision of 0.874 in detecting dirt on real datasets, showcasing the effectiveness of using synthetic data for training in this application, which has not been reported in previous works.
Article
Computer Science, Artificial Intelligence
Hamdan Abdellatef, Lina J. Karam
Summary: This paper proposes performing the learning and inference processes in the compressed domain to reduce computational complexity and improve speed of neural networks. Experimental results show that modified ResNet-50 in the compressed domain is 70% faster than traditional spatial-based ResNet-50 while maintaining similar accuracy. Additionally, a preprocessing step with partial encoding is suggested to improve resilience to distortions caused by low-quality encoded images. Training a network with highly compressed data can achieve good classification accuracy with significantly reduced storage requirements.
Article
Computer Science, Artificial Intelligence
Victor R. Barradas, Yasuharu Koike, Nicolas Schweighofer
Summary: Inverse models are essential for human motor learning as they map desired actions to motor commands. The shape of the error surface and the distribution of targets in a task play a crucial role in determining the speed of learning.
Article
Computer Science, Artificial Intelligence
Ting Zhou, Hanshu Yan, Jingfeng Zhang, Lei Liu, Bo Han
Summary: We propose a defense strategy that reduces the success rate of data poisoning attacks in downstream tasks by pre-training a robust foundation model.
Article
Computer Science, Artificial Intelligence
Hao Sun, Li Shen, Qihuang Zhong, Liang Ding, Shixiang Chen, Jingwei Sun, Jing Li, Guangzhong Sun, Dacheng Tao
Summary: In this paper, the convergence rate of AdaSAM in the stochastic non-convex setting is analyzed. Theoretical proof shows that AdaSAM has a linear speedup property and decouples the stochastic gradient steps with the adaptive learning rate and perturbed gradient. Experimental results demonstrate that AdaSAM outperforms other optimizers in terms of performance.
Article
Computer Science, Artificial Intelligence
Juntong Yun, Du Jiang, Li Huang, Bo Tao, Shangchun Liao, Ying Liu, Xin Liu, Gongfa Li, Disi Chen, Baojia Chen
Summary: In this study, a dual manipulator grasping detection model based on the Markov decision process is proposed. By parameterizing the grasping detection model of dual manipulators using a cross entropy convolutional neural network and a full convolutional neural network, stable grasping of complex multiple objects is achieved. Robot grasping experiments were conducted to verify the feasibility and superiority of this method.
Article
Computer Science, Artificial Intelligence
Miaohui Zhang, Kaifang Li, Jianxin Ma, Xile Wang
Summary: This paper proposes an unsupervised person re-identification (Re-ID) method that uses two asymmetric networks to generate pseudo-labels for each other by clustering and updates and optimizes the pseudo-labels through alternate training. It also designs similarity compensation and similarity suppression based on the camera ID of pedestrian images to optimize the similarity measure. Extensive experiments show that the proposed method achieves superior performance compared to state-of-the-art unsupervised person re-identification methods.
Article
Computer Science, Artificial Intelligence
Florian Bacho, Dominique Chu
Summary: This paper proposes a new approach called the Forward Direct Feedback Alignment algorithm for supervised learning in deep neural networks. By combining activity-perturbed forward gradients, direct feedback alignment, and momentum, this method achieves better performance and convergence speed compared to other local alternatives to backpropagation.
Article
Computer Science, Artificial Intelligence
Xiaojian Ding, Yi Li, Shilin Chen
Summary: This research paper addresses the limitations of recursive feature elimination (RFE) and its variants in high-dimensional feature selection tasks. The proposed algorithms, which introduce a novel feature ranking criterion and an optimal feature subset evaluation algorithm, outperform current state-of-the-art methods.
Article
Computer Science, Artificial Intelligence
Naoko Koide-Majima, Shinji Nishimoto, Kei Majima
Summary: Visual images observed by humans can be reconstructed from brain activity, and the visualization of arbitrary natural images from mental imagery has been achieved through an improved method. This study provides a unique tool for directly investigating the subjective contents of the brain.
Article
Computer Science, Artificial Intelligence
Huanjie Tao, Qianyue Duan
Summary: In this paper, a hierarchical attention network with progressive feature fusion is proposed for facial expression recognition (FER), addressing the challenges posed by pose variation, occlusions, and illumination variation. The model achieves enhanced performance by aggregating diverse features and progressively enhancing discriminative features.
Article
Computer Science, Artificial Intelligence
Zhenyi Wang, Pengfei Yang, Linwei Hu, Bowen Zhang, Chengmin Lin, Wenkai Lv, Quan Wang
Summary: In the face of the complex landscape of deep learning, we propose a novel subgraph-level performance prediction method called SLAPP, which combines graph and operator features through an innovative graph neural network called EAGAT, providing accurate performance predictions. In addition, we introduce a mixed loss design with dynamic weight adjustment to improve predictive accuracy.
Article
Computer Science, Artificial Intelligence
Yiyang Yin, Shuangling Luo, Jun Zhou, Liang Kang, Calvin Yu-Chian Chen
Summary: Medical image segmentation is crucial for modern healthcare systems, especially in reducing surgical risks and planning treatments. Transanal total mesorectal excision (TaTME) has become an important method for treating colon and rectum cancers. Real-time instance segmentation during TaTME surgeries can assist surgeons in minimizing risks. However, the dynamic variations in TaTME images pose challenges for accurate instance segmentation.
Article
Computer Science, Artificial Intelligence
Teng Cheng, Lei Sun, Junning Zhang, Jinling Wang, Zhanyang Wei
Summary: This study proposes a scheme that combines the start-stop point signal features for wideband multi-signal detection, called Fast Spectrum-Size Self-Training network (FSSNet). By utilizing start-stop points to build the signal model, this method successfully solves the difficulty of existing deep learning methods in detecting discontinuous signals and achieves satisfactory detection speed.
Article
Computer Science, Artificial Intelligence
Wenming Wu, Xiaoke Ma, Quan Wang, Maoguo Gong, Quanxue Gao
Summary: The layer-specific modules in multi-layer networks are critical for understanding the structure and function of the system. However, existing methods fail to accurately characterize and balance the connectivity and specificity of these modules. To address this issue, a joint learning graph clustering algorithm (DRDF) is proposed, which learns the deep representation and discriminative features of the multi-layer network, and balances the connectivity and specificity of the layer-specific modules through joint learning.
Article
Computer Science, Artificial Intelligence
Guanghui Yue, Guibin Zhuo, Weiqing Yan, Tianwei Zhou, Chang Tang, Peng Yang, Tianfu Wang
Summary: This paper proposes a novel boundary uncertainty aware network (BUNet) for precise and robust colorectal polyp segmentation. BUNet utilizes a pyramid vision transformer encoder to learn multi-scale features and incorporates a boundary exploration module (BEM) and a boundary uncertainty aware module (BUM) to handle boundary areas. Experimental results demonstrate that BUNet outperforms other methods in terms of performance and generalization ability.