Pub Date : 2022-07-18DOI: 10.1109/IJCNN55064.2022.9892953
Arturo de Souza, M. Melchiades, S. Rigo, G. D. O. Ramos
Mental disorders affect a large number of people worldwide. In response to the increasing number of people affected by such illnesses, there has been an increased interest in the use of state-of-the-art technologies to mitigate its effects. This paper presents a Sequence Model for Stress Classification (MoStress), which is a novel pipeline for pre-processing physio-logical data collected from wearable devices and for identifying stress sequences using a recurrent neural network (RNN). Using the WESAD dataset, the RNN model achieved accuracy of 86% in a three-class classification problem (baseline vs. stress vs. amusement). When only considering the presence of stress or not, we achieved an accuracy of 96.5% as well as precision, recall, and f'1-score of 96%, 93%, and 94%, respectively. Those results are close to other papers using the same dataset, however, the neural network used on MoStress, is considerable simpler.
{"title":"MoStress: a Sequence Model for Stress Classification","authors":"Arturo de Souza, M. Melchiades, S. Rigo, G. D. O. Ramos","doi":"10.1109/IJCNN55064.2022.9892953","DOIUrl":"https://doi.org/10.1109/IJCNN55064.2022.9892953","url":null,"abstract":"Mental disorders affect a large number of people worldwide. In response to the increasing number of people affected by such illnesses, there has been an increased interest in the use of state-of-the-art technologies to mitigate its effects. This paper presents a Sequence Model for Stress Classification (MoStress), which is a novel pipeline for pre-processing physio-logical data collected from wearable devices and for identifying stress sequences using a recurrent neural network (RNN). Using the WESAD dataset, the RNN model achieved accuracy of 86% in a three-class classification problem (baseline vs. stress vs. amusement). When only considering the presence of stress or not, we achieved an accuracy of 96.5% as well as precision, recall, and f'1-score of 96%, 93%, and 94%, respectively. Those results are close to other papers using the same dataset, however, the neural network used on MoStress, is considerable simpler.","PeriodicalId":106974,"journal":{"name":"2022 International Joint Conference on Neural Networks (IJCNN)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114570826","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}
Pub Date : 2022-07-18DOI: 10.1109/IJCNN55064.2022.9892628
R. S. Jaisurya, Snehasis Mukherjee
Single image dehazing is a popular research topic among the researchers in computer vision, machine learning, image processing, and graphics. Most of the recent methods for single image dehazing are based upon supervised learning set up. However, supervised methods require annotation of the data, which often makes the dehazing methods biased towards the manual annotation errors. Unsupervised methods are more likely to produce realistic, clear images. However, fewer efforts are found in the literature for single image dehazing in unsupervised set up. We propose an enhanced CycleGAN architecture for Unpaired single image dehazing, with an attention-based transformer architecture embedded in the generator. The proposed transformer comprises three components: 1) A Feature Attention (FA) block combining channel attention and pixel attention mechanism, 2) A Dynamic feature enhancement block for dynamically capturing the spatial structured features and 3) An adaptive mix-up module to preserve the flow of shallow features from downsampling. Experiments on the benchmark datasets show the efficacy of the proposed method. Codes for this work are available in the link: https://github.com/rsjai47/Attention-Based-CycleDehaze.
{"title":"Attention-based Single Image Dehazing Using Improved CycleGAN","authors":"R. S. Jaisurya, Snehasis Mukherjee","doi":"10.1109/IJCNN55064.2022.9892628","DOIUrl":"https://doi.org/10.1109/IJCNN55064.2022.9892628","url":null,"abstract":"Single image dehazing is a popular research topic among the researchers in computer vision, machine learning, image processing, and graphics. Most of the recent methods for single image dehazing are based upon supervised learning set up. However, supervised methods require annotation of the data, which often makes the dehazing methods biased towards the manual annotation errors. Unsupervised methods are more likely to produce realistic, clear images. However, fewer efforts are found in the literature for single image dehazing in unsupervised set up. We propose an enhanced CycleGAN architecture for Unpaired single image dehazing, with an attention-based transformer architecture embedded in the generator. The proposed transformer comprises three components: 1) A Feature Attention (FA) block combining channel attention and pixel attention mechanism, 2) A Dynamic feature enhancement block for dynamically capturing the spatial structured features and 3) An adaptive mix-up module to preserve the flow of shallow features from downsampling. Experiments on the benchmark datasets show the efficacy of the proposed method. Codes for this work are available in the link: https://github.com/rsjai47/Attention-Based-CycleDehaze.","PeriodicalId":106974,"journal":{"name":"2022 International Joint Conference on Neural Networks (IJCNN)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114605207","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}
This paper studies adversarial attacks and defences against deep learning models trained on infrared data to classify the presence of humans and detect their bounding boxes, which differently from the standard RGB case is an open research problem with multiple consequences related to safety and secure artificial intelligence applications. The paper has two major contributions. Firstly, we study the effectiveness of the Projected Gradient Descent (PGD) adversarial attack against Convolutional Neural Networks (CNNs) trained exclusively on infrared data, and the effectiveness of adversarial training as a possible defense against the attack. Secondly, we study the response of an object detection model trained on infrared images under adversarial attacks. In particular, we propose and empirically evaluate two attacks: one classical attack from the literature on object detection, and a new hybrid attack which exploits a common CNN base architecture of the classifier and the object detector. We show for the first time that adversarial attacks weaken the performance of classification and detection models trained on infrared images only. We also prove that the defense adversarial training optimized for the infinity norm increases the robustness of different classification models trained on infrared data.
{"title":"Evaluating Adversarial Attacks and Defences in Infrared Deep Learning Monitoring Systems","authors":"Flaminia Spasiano, Gabriele Gennaro, Simone Scardapane","doi":"10.1109/IJCNN55064.2022.9891997","DOIUrl":"https://doi.org/10.1109/IJCNN55064.2022.9891997","url":null,"abstract":"This paper studies adversarial attacks and defences against deep learning models trained on infrared data to classify the presence of humans and detect their bounding boxes, which differently from the standard RGB case is an open research problem with multiple consequences related to safety and secure artificial intelligence applications. The paper has two major contributions. Firstly, we study the effectiveness of the Projected Gradient Descent (PGD) adversarial attack against Convolutional Neural Networks (CNNs) trained exclusively on infrared data, and the effectiveness of adversarial training as a possible defense against the attack. Secondly, we study the response of an object detection model trained on infrared images under adversarial attacks. In particular, we propose and empirically evaluate two attacks: one classical attack from the literature on object detection, and a new hybrid attack which exploits a common CNN base architecture of the classifier and the object detector. We show for the first time that adversarial attacks weaken the performance of classification and detection models trained on infrared images only. We also prove that the defense adversarial training optimized for the infinity norm increases the robustness of different classification models trained on infrared data.","PeriodicalId":106974,"journal":{"name":"2022 International Joint Conference on Neural Networks (IJCNN)","volume":"2010 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122016207","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}
Pub Date : 2022-07-18DOI: 10.1109/IJCNN55064.2022.9892114
Bradley Walters, S. Ortega-Martorell, I. Olier, Paulo J. G. Lisboa
A major challenge in delivering reliable and trustworthy computational intelligence for practical applications in clinical medicine is interpretability. This aspect of machine learning is a major distinguishing factor compared with traditional statistical models for the stratification of patients, which typically use rules or a risk score identified by logistic regression. We show how functions of one and two variables can be extracted from pre-trained machine learning models using anchored Analysis of Variance (ANOVA) decompositions. This enables complex interaction terms to be filtered out by aggressive regularisation using the Least Absolute Shrinkage and Selection Operator (LASSO) resulting in a sparse model with comparable or even better performance than the original pre-trained black-box. Besides being theoretically well-founded, the decomposition of a black-box multivariate probabilistic binary classifier into a General Additive Model (GAM) comprising a linear combination of non-linear functions of one or two variables provides full interpretability. In effect this extends logistic regression into non-linear modelling without the need for manual intervention by way of variable transformations, using the pre-trained model as a seed. The application of the proposed methodology to existing machine learning models is demonstrated using the Multi-Layer Perceptron (MLP), Support Vector Machine (SVM), Random Forests (RF) and Gradient Boosting Machines (GBM), to model a data frame from a well-known benchmark dataset available from Physionet, the Medical Information Mart for Intensive Care (MIMIC-III). Both the classification performance and plausibility of clinical interpretation compare favourably with other state-of-the-art sparse models namely Sparse Additive Models (SAM) and the Explainable Boosting Machine (EBM).
{"title":"Towards interpretable machine learning for clinical decision support","authors":"Bradley Walters, S. Ortega-Martorell, I. Olier, Paulo J. G. Lisboa","doi":"10.1109/IJCNN55064.2022.9892114","DOIUrl":"https://doi.org/10.1109/IJCNN55064.2022.9892114","url":null,"abstract":"A major challenge in delivering reliable and trustworthy computational intelligence for practical applications in clinical medicine is interpretability. This aspect of machine learning is a major distinguishing factor compared with traditional statistical models for the stratification of patients, which typically use rules or a risk score identified by logistic regression. We show how functions of one and two variables can be extracted from pre-trained machine learning models using anchored Analysis of Variance (ANOVA) decompositions. This enables complex interaction terms to be filtered out by aggressive regularisation using the Least Absolute Shrinkage and Selection Operator (LASSO) resulting in a sparse model with comparable or even better performance than the original pre-trained black-box. Besides being theoretically well-founded, the decomposition of a black-box multivariate probabilistic binary classifier into a General Additive Model (GAM) comprising a linear combination of non-linear functions of one or two variables provides full interpretability. In effect this extends logistic regression into non-linear modelling without the need for manual intervention by way of variable transformations, using the pre-trained model as a seed. The application of the proposed methodology to existing machine learning models is demonstrated using the Multi-Layer Perceptron (MLP), Support Vector Machine (SVM), Random Forests (RF) and Gradient Boosting Machines (GBM), to model a data frame from a well-known benchmark dataset available from Physionet, the Medical Information Mart for Intensive Care (MIMIC-III). Both the classification performance and plausibility of clinical interpretation compare favourably with other state-of-the-art sparse models namely Sparse Additive Models (SAM) and the Explainable Boosting Machine (EBM).","PeriodicalId":106974,"journal":{"name":"2022 International Joint Conference on Neural Networks (IJCNN)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116735577","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}
Pub Date : 2022-07-18DOI: 10.1109/IJCNN55064.2022.9892881
T. Iinuma, S. Nobukawa, S. Yamaguchi
An echo state network (ESN), consisting of an input layer, reservoir, and output layer, provides a higher learning-efficient approach than other recurrent neural networks (RNNs). In the design of ESNs, a sufficiently large number of reservoir neurons is required compared to the dimension of the input signal. Thus, the number of neurons must be increased for high-dimensional input to achieve good performance. However, an increase in the number of neurons increases the computational load. To solve this problem, we propose an assembly ESN (AESN) architecture comprising a feature extraction part that uses multiple sub-ESNs with segregated components of high-dimensional input and a feature integration part. To validate the effectiveness of the proposed AESN, we investigated and compared the conventional ESN with the AESN under high-dimensional input. The results show that the AESN is possibly superior to the conventional ESN in accuracy, memory performance, and computational load. We believe that the AESN also has a correct integration function. Therefore, the proposed method is expected to solve high-dimensional problems with improved accuracy.
{"title":"Assembly of Echo State Networks Driven by Segregated Low Dimensional Signals","authors":"T. Iinuma, S. Nobukawa, S. Yamaguchi","doi":"10.1109/IJCNN55064.2022.9892881","DOIUrl":"https://doi.org/10.1109/IJCNN55064.2022.9892881","url":null,"abstract":"An echo state network (ESN), consisting of an input layer, reservoir, and output layer, provides a higher learning-efficient approach than other recurrent neural networks (RNNs). In the design of ESNs, a sufficiently large number of reservoir neurons is required compared to the dimension of the input signal. Thus, the number of neurons must be increased for high-dimensional input to achieve good performance. However, an increase in the number of neurons increases the computational load. To solve this problem, we propose an assembly ESN (AESN) architecture comprising a feature extraction part that uses multiple sub-ESNs with segregated components of high-dimensional input and a feature integration part. To validate the effectiveness of the proposed AESN, we investigated and compared the conventional ESN with the AESN under high-dimensional input. The results show that the AESN is possibly superior to the conventional ESN in accuracy, memory performance, and computational load. We believe that the AESN also has a correct integration function. Therefore, the proposed method is expected to solve high-dimensional problems with improved accuracy.","PeriodicalId":106974,"journal":{"name":"2022 International Joint Conference on Neural Networks (IJCNN)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129507351","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}
Pub Date : 2022-07-18DOI: 10.1109/IJCNN55064.2022.9892381
Hongying Ma, Wuyang Xue, R. Ying, Peilin Liu
Model-based reinforcement learning algorithms can alleviate the low sample efficiency problem compared with modelfree methods for control tasks. However, the learned policy's performance often lags behind the best model-free algorithms since its weak exploration ability. Existing model-based reinforcement learning algorithms learn policy by interacting with the learned world model and then use the learned policy to guide a new round of world model learning. Due to weak policy exploration ability, the learned world model has a large bias. As a result, it fails to learn the globally optimal policy on such a world model. This paper improves the learned world model by maximizing both the reward and the corresponding policy entropy in the framework of maximum entropy reinforcement learning. The effectiveness of applying the maximum entropy approach to model-based reinforcement learning is supported by the better performance of our algorithm on several complex mujoco and deepmind control suite tasks.
{"title":"MaxEnt Dreamer: Maximum Entropy Reinforcement Learning with World Model","authors":"Hongying Ma, Wuyang Xue, R. Ying, Peilin Liu","doi":"10.1109/IJCNN55064.2022.9892381","DOIUrl":"https://doi.org/10.1109/IJCNN55064.2022.9892381","url":null,"abstract":"Model-based reinforcement learning algorithms can alleviate the low sample efficiency problem compared with modelfree methods for control tasks. However, the learned policy's performance often lags behind the best model-free algorithms since its weak exploration ability. Existing model-based reinforcement learning algorithms learn policy by interacting with the learned world model and then use the learned policy to guide a new round of world model learning. Due to weak policy exploration ability, the learned world model has a large bias. As a result, it fails to learn the globally optimal policy on such a world model. This paper improves the learned world model by maximizing both the reward and the corresponding policy entropy in the framework of maximum entropy reinforcement learning. The effectiveness of applying the maximum entropy approach to model-based reinforcement learning is supported by the better performance of our algorithm on several complex mujoco and deepmind control suite tasks.","PeriodicalId":106974,"journal":{"name":"2022 International Joint Conference on Neural Networks (IJCNN)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129602656","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}
Pub Date : 2022-07-18DOI: 10.1109/IJCNN55064.2022.9892022
Ammar Ahmed, Archi Yadav, Avinash Sharma, R. Bapi
An active area of research in cognitive science is characterizing the relationship between brain structure and the observed functional activations. Recent graph diffusion models have had great success in mapping whole-brain, resting-state dynamics measured using functional Magnetic Resonance Imaging (fMRI) to the brain structure derived using diffusion and T1 brain imaging. Here we test the application of one such graph diffusion method called the Multiple Kernel Learning (MKL) model. MKL model, formulated as a reaction-diffusion system using Wilson-Cowan equations, combines multiple diffusion kernels at different scales to predict functional connectome (FC) arising from a fixed structural connectome (SC). Our simulation results demonstrate that the MKL model successfully mapped the relationship between SC and FC from five different Electroen-cephalogram (EEG) bands (delta, theta, alpha, beta, and gamma). We used simultaneously acquired EEG-fMRI and NODDI dataset of 17 participants. The correlation between predicted FC and ground truth FC was higher for EEG bands than for fMRI data. The prediction accuracy peaked for the alpha band, and the highest frequency band, gamma had the lowest prediction accuracy. To the best of our knowledge, this is the first such end-to-end application of multiple kernel graph diffusion framework for modeling EEG data. One of the important features of MKL model is its ability to incorporate structural connectivity features into the generative model that predicts the EEG functional connectivity.
{"title":"Multiple Kernel Learning for Modeling Resting State EEG Connectomes using Structural Connectivity of the Brain","authors":"Ammar Ahmed, Archi Yadav, Avinash Sharma, R. Bapi","doi":"10.1109/IJCNN55064.2022.9892022","DOIUrl":"https://doi.org/10.1109/IJCNN55064.2022.9892022","url":null,"abstract":"An active area of research in cognitive science is characterizing the relationship between brain structure and the observed functional activations. Recent graph diffusion models have had great success in mapping whole-brain, resting-state dynamics measured using functional Magnetic Resonance Imaging (fMRI) to the brain structure derived using diffusion and T1 brain imaging. Here we test the application of one such graph diffusion method called the Multiple Kernel Learning (MKL) model. MKL model, formulated as a reaction-diffusion system using Wilson-Cowan equations, combines multiple diffusion kernels at different scales to predict functional connectome (FC) arising from a fixed structural connectome (SC). Our simulation results demonstrate that the MKL model successfully mapped the relationship between SC and FC from five different Electroen-cephalogram (EEG) bands (delta, theta, alpha, beta, and gamma). We used simultaneously acquired EEG-fMRI and NODDI dataset of 17 participants. The correlation between predicted FC and ground truth FC was higher for EEG bands than for fMRI data. The prediction accuracy peaked for the alpha band, and the highest frequency band, gamma had the lowest prediction accuracy. To the best of our knowledge, this is the first such end-to-end application of multiple kernel graph diffusion framework for modeling EEG data. One of the important features of MKL model is its ability to incorporate structural connectivity features into the generative model that predicts the EEG functional connectivity.","PeriodicalId":106974,"journal":{"name":"2022 International Joint Conference on Neural Networks (IJCNN)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129806487","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}
Pub Date : 2022-07-18DOI: 10.1109/IJCNN55064.2022.9892755
Maciej Golgowski, S. Osowski
The paper analyzes and compares two forms of wavelet transformation: discrete (DWT) and continuous (CWT) combined with machine learning in the analysis of the bearing failure. It presents the automatic system to detect the anomaly in the rolling bearing based on wavelet analysis of vibration waveforms combined with the set of classical and deep classifiers. The wavelet transformation is used in the stage of pre-processing of the signal for generating the input attributes in the final classification system. The considered structures of the classifiers include 6 classical machine learning tools integrated into an ensemble and a combination of a few deep Convolutional Neural Networks (CNN) to develop the most accurate diagnostics of the bearing. The calculations have been done in Python and Matlab. The results of both approaches DWT and CWT are discussed and compared. They show the high effectiveness of the approach based on the cooperation of wavelet transform and machine learning methods.
{"title":"Detection of bearing failures using wavelet transformation and machine learning approach","authors":"Maciej Golgowski, S. Osowski","doi":"10.1109/IJCNN55064.2022.9892755","DOIUrl":"https://doi.org/10.1109/IJCNN55064.2022.9892755","url":null,"abstract":"The paper analyzes and compares two forms of wavelet transformation: discrete (DWT) and continuous (CWT) combined with machine learning in the analysis of the bearing failure. It presents the automatic system to detect the anomaly in the rolling bearing based on wavelet analysis of vibration waveforms combined with the set of classical and deep classifiers. The wavelet transformation is used in the stage of pre-processing of the signal for generating the input attributes in the final classification system. The considered structures of the classifiers include 6 classical machine learning tools integrated into an ensemble and a combination of a few deep Convolutional Neural Networks (CNN) to develop the most accurate diagnostics of the bearing. The calculations have been done in Python and Matlab. The results of both approaches DWT and CWT are discussed and compared. They show the high effectiveness of the approach based on the cooperation of wavelet transform and machine learning methods.","PeriodicalId":106974,"journal":{"name":"2022 International Joint Conference on Neural Networks (IJCNN)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129815476","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}
Pub Date : 2022-07-18DOI: 10.1109/IJCNN55064.2022.9892380
Swapna Sasi, Taher Yunus Lilywala, B. Bhattacharya
We have made a comparative study of three optimisation algorithms viz. Random Search (RS), Grid Search (GS) and Bayesian Optimization (BO) to find optimal hyperparameter combinations in an existing brain-inspired thalamocortical model that can simulate brain signals such as local field potentials (lfp) and electroencephalogram (eeg). The layout and parameters for the model are sourced from anatomical and physiological data. However, there is a lot of missing data in such sources due to obvious constraints in wet-lab experimental studies. In our previous work, the missing data are set by trial and error. As the scale of the model gets larger though, the combinatorics of the hyperparameters explode and manual parameter tuning gets non-trivial. The goal of this study is to identify the optimisation algorithm (among the three abovementioned) that gives the best performance at minimal computational costs; performance is evaluated by setting an objective, which is to search for hyperparameter combinations that can simulate theta (4 – 8 Hz), alpha (8 – 13 Hz) and beta (13 – 30 Hz) rhythms, which are typically observed in eeg and lfp. Each optimisation algorithm is tested on a small model (thalamus only) with eight hyperparameters and a large model (thalamocortical) with maximum of fifteen hyperparameters. The performance metric for each algorithm is measured by the number of times the objective is achieved during a fixed number of trials. Our results demonstrate that BO performs the best in reaching the objective with a 30.5% better performance compared to GS and 13% better than RS. In comparison, GS performance is lower with an exponential increase in time with increasing grid size. Overall, our study demonstrates the suitability of using the BO for optimising hyperparameter search in our thalamocortical network model of the visual pathway.
{"title":"Optimising hyperparameter search in a visual thalamocortical pathway model","authors":"Swapna Sasi, Taher Yunus Lilywala, B. Bhattacharya","doi":"10.1109/IJCNN55064.2022.9892380","DOIUrl":"https://doi.org/10.1109/IJCNN55064.2022.9892380","url":null,"abstract":"We have made a comparative study of three optimisation algorithms viz. Random Search (RS), Grid Search (GS) and Bayesian Optimization (BO) to find optimal hyperparameter combinations in an existing brain-inspired thalamocortical model that can simulate brain signals such as local field potentials (lfp) and electroencephalogram (eeg). The layout and parameters for the model are sourced from anatomical and physiological data. However, there is a lot of missing data in such sources due to obvious constraints in wet-lab experimental studies. In our previous work, the missing data are set by trial and error. As the scale of the model gets larger though, the combinatorics of the hyperparameters explode and manual parameter tuning gets non-trivial. The goal of this study is to identify the optimisation algorithm (among the three abovementioned) that gives the best performance at minimal computational costs; performance is evaluated by setting an objective, which is to search for hyperparameter combinations that can simulate theta (4 – 8 Hz), alpha (8 – 13 Hz) and beta (13 – 30 Hz) rhythms, which are typically observed in eeg and lfp. Each optimisation algorithm is tested on a small model (thalamus only) with eight hyperparameters and a large model (thalamocortical) with maximum of fifteen hyperparameters. The performance metric for each algorithm is measured by the number of times the objective is achieved during a fixed number of trials. Our results demonstrate that BO performs the best in reaching the objective with a 30.5% better performance compared to GS and 13% better than RS. In comparison, GS performance is lower with an exponential increase in time with increasing grid size. Overall, our study demonstrates the suitability of using the BO for optimising hyperparameter search in our thalamocortical network model of the visual pathway.","PeriodicalId":106974,"journal":{"name":"2022 International Joint Conference on Neural Networks (IJCNN)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129859508","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}
Pub Date : 2022-07-18DOI: 10.1109/IJCNN55064.2022.9891917
Lerina Aversano, M. Bernardi, Marta Cimitile, Martina Iammarino, Chiara Verdone
Parkinson's disease is one of the most prevalent neurodegenerative diseases in the world, usually occurring after the age of 50, but in some cases, also affects younger people. It is a disease that affects movement, coordination, and muscle control, all of which cause a range of symptoms that affect patients' writing and drawing skills. Diagnosis is clinical, so it occurs mainly through the evaluation of the patient's movements, coordination, and muscle control. Therefore, the analysis of micrographic models can introduce a new methodology of investigation in the diagnosis and monitoring of Parkinson's disease. This study proposes an approach based on artificial intelligence in combination with the spiral test, which consists in asking the patient to draw a spiral, thanks to which it is possible to make the early diagnosis of Parkinson's disease. The classification is performed with a combination of an Echo State Network and an MLP layer. To validate the approach, several classification algorithms belonging to two macro groups (boosting decision trees based) were used as baseline. The results obtained are very satisfactory with the ESN-based classifier exhibiting an F-Score of 97.8%. The very encouraging results indicate that the proposed approach may be an effective contribution to improving Parkinson's diagnostics.
{"title":"Early Detection of Parkinson's Disease using Spiral Test and Echo State Networks","authors":"Lerina Aversano, M. Bernardi, Marta Cimitile, Martina Iammarino, Chiara Verdone","doi":"10.1109/IJCNN55064.2022.9891917","DOIUrl":"https://doi.org/10.1109/IJCNN55064.2022.9891917","url":null,"abstract":"Parkinson's disease is one of the most prevalent neurodegenerative diseases in the world, usually occurring after the age of 50, but in some cases, also affects younger people. It is a disease that affects movement, coordination, and muscle control, all of which cause a range of symptoms that affect patients' writing and drawing skills. Diagnosis is clinical, so it occurs mainly through the evaluation of the patient's movements, coordination, and muscle control. Therefore, the analysis of micrographic models can introduce a new methodology of investigation in the diagnosis and monitoring of Parkinson's disease. This study proposes an approach based on artificial intelligence in combination with the spiral test, which consists in asking the patient to draw a spiral, thanks to which it is possible to make the early diagnosis of Parkinson's disease. The classification is performed with a combination of an Echo State Network and an MLP layer. To validate the approach, several classification algorithms belonging to two macro groups (boosting decision trees based) were used as baseline. The results obtained are very satisfactory with the ESN-based classifier exhibiting an F-Score of 97.8%. The very encouraging results indicate that the proposed approach may be an effective contribution to improving Parkinson's diagnostics.","PeriodicalId":106974,"journal":{"name":"2022 International Joint Conference on Neural Networks (IJCNN)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128482437","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：