H.3. Artificial Intelligence
Vahideh Monemizadeh; Kourosh Kiani
Abstract
Anomaly detection is becoming increasingly crucial across various fields, including cybersecurity, financial risk management, and health monitoring. However, it faces significant challenges when dealing with large-scale, high-dimensional, and unlabeled datasets. This study focuses on decision tree-based ...
Read More
Anomaly detection is becoming increasingly crucial across various fields, including cybersecurity, financial risk management, and health monitoring. However, it faces significant challenges when dealing with large-scale, high-dimensional, and unlabeled datasets. This study focuses on decision tree-based methods for anomaly detection due to their scalability, interpretability, and effectiveness in managing high-dimensional data. Although Isolation Forest (iForest) and its extended variant, Extended Isolation Forest (EIF), are widely used, they exhibit limitations in identifying anomalies, particularly in handling normal data distributions and preventing the formation of ghost clusters. The Rotated Isolation Forest (RIF) was developed to address these challenges, enhancing the model's ability to discern true anomalies from normal variations by employing randomized rotations in feature space. Building on this approach, we proposed the Discrete Rotated Isolation Forest (DRIF) model, which integrates an Autoencoder for dimensionality reduction. Using a discrete probability distribution and an Autoencoder enhance computational efficiency. Experimental evaluations on synthetic and real-world datasets demonstrate that proposed model outperforms iForest, EIF, and RIF. And also achieving higher Receiver Operating Characteristic-Area Under the Curve (ROC-AUC) scores and significantly faster execution times. These findings establish the proposed model as a robust, scalable, and efficient approach for unsupervised anomaly detection in high-dimensional datasets.
H.3.8. Natural Language Processing
Nura Esfandiari; Kourosh Kiani; Razieh Rastgoo
Abstract
A chatbot is a computer program system designed to simulate human-like conversations and interact with users. It is a form of conversational agent that utilizes Natural Language Processing (NLP) and sequential models to understand user input, interpret their intent, and generate appropriate answer. This ...
Read More
A chatbot is a computer program system designed to simulate human-like conversations and interact with users. It is a form of conversational agent that utilizes Natural Language Processing (NLP) and sequential models to understand user input, interpret their intent, and generate appropriate answer. This approach aims to generate word sequences in the form of coherent phrases. A notable challenge associated with previous models lies in their sequential training process, which can result in less accurate outcomes. To address this limitation, a novel generative chatbot is proposed, integrating the power of Reinforcement Learning (RL) and transformer models. The proposed chatbot aims to overcome the challenges associated with sequential training by combining these two approaches. The proposed approach employs a Double Deep Q-Network (DDQN) architecture with utilizing a transformer model as the agent. This agent takes the human question as an input state and generates the bot answer as an action. To the best of our knowledge, this is the first time that a generative chatbot is proposed using a DDQN architecture with the embedded transformer as an agent. Results on two public datasets, Daily Dialog and Chit-Chat, validate the superiority of the proposed approach over state-of-the-art models involves employing various evaluation metrics.
Oladosu Oladimeji; Olayanju Oladimeji
Abstract
Breast cancer is the second major cause of death and accounts for 16% of all cancer deaths worldwide. Most of the methods of detecting breast cancer are very expensive and difficult to interpret such as mammography. There are also limitations such as cumulative radiation exposure, over-diagnosis, false ...
Read More
Breast cancer is the second major cause of death and accounts for 16% of all cancer deaths worldwide. Most of the methods of detecting breast cancer are very expensive and difficult to interpret such as mammography. There are also limitations such as cumulative radiation exposure, over-diagnosis, false positives and negatives in women with a dense breast which pose certain uncertainties in high-risk population. The objective of this study is Detecting Breast Cancer Through Blood Analysis Data Using Classification Algorithms. This will serve as a complement to these expensive methods. High ranking features were extracted from the dataset. The KNN, SVM and J48 algorithms were used as the training platform to classify 116 instances. Furthermore, 10-fold cross validation and holdout procedures were used coupled with changing of random seed. The result showed that KNN algorithm has the highest and best accuracy of 89.99% and 85.21% for cross validation and holdout procedure respectively. This is followed by the J48 with 84.65% and 75.65% for the two procedures respectively. SVM had 77.58% and 68.69% respectively. Although it was also discovered that Blood Glucose level is a major determinant in detecting breast cancer, it has to be combined with other attributes to make decision as a result of other health issues like diabetes. With the result obtained, women are advised to do regular check-ups including blood analysis in order to know which of the blood components need to be worked on to prevent breast cancer based on the model generated in this study.
H.3. Artificial Intelligence
Saheb Ghanbari Motlagh; Fateme Razi Astaraei; Mojtaba Hajihosseini; Saeed Madani
Abstract
This study explores the potential use of Machine Learning (ML) techniques to enhance three types of nano-based solar cells. Perovskites of methylammonium-free formamidinium (FA) and mixed cation-based cells exhibit a boosted efficiency when employing ML techniques. Moreover, ML methods are utilized to ...
Read More
This study explores the potential use of Machine Learning (ML) techniques to enhance three types of nano-based solar cells. Perovskites of methylammonium-free formamidinium (FA) and mixed cation-based cells exhibit a boosted efficiency when employing ML techniques. Moreover, ML methods are utilized to identify optimal donor complexes, high blind temperature materials, and to advance the thermodynamic stability of perovskites. Another significant application of ML in dye-sensitized solar cells (DSSCs) is the detection of novel dyes, solvents, and molecules for improving the efficiency and performance of solar cells. Some of these materials have increased cell efficiency, short-circuit current, and light absorption by more than 20%. ML algorithms to fine-tune network and plasmonic field bandwidths improve the efficiency and light absorption of surface plasmonic resonance (SPR) solar cells. This study outlines the potential of ML techniques to optimize and improve the development of nano-based solar cells, leading to promising results for the field of solar energy generation and supporting the demand for sustainable and dependable energy.
H.3. Artificial Intelligence
Pouria Rabiei; Nosratali Ashrafi-Payaman
Abstract
Today, the amount of data with graph structure has increased dramatically. Detecting structural anomalies in the graph, such as nodes and edges whose behavior deviates from the expected behavior of the network, is important in real-world applications. Thus, in our research work, we extract the structural ...
Read More
Today, the amount of data with graph structure has increased dramatically. Detecting structural anomalies in the graph, such as nodes and edges whose behavior deviates from the expected behavior of the network, is important in real-world applications. Thus, in our research work, we extract the structural characteristics of the dynamic graph by using graph convolutional neural networks, then by using temporal neural network Like GRU, we extract the short-term temporalcharacteristics of the dynamic graph and by using the attention mechanism integrated with GRU, long-term temporal dependencies are considered. Finally, by using the neural network classifier, the abnormal edge is detected in each timestamp. Conducted experiments on the two datasets, UC Irvine messages and Digg with three baselines, including Goutlier, Netwalk and CMSketch illustrate our model outperform existing methods in a dynamic graph by 10 and 15% onaverage on the UCI and Digg datasets respectively. We also measured the model with AUC and confusion matrix for 1, 5, and 10 percent anomaly injection.
H.3. Artificial Intelligence
Rasoul Hosseinzadeh; Mahdi Sadeghzadeh
Abstract
The attention mechanisms have significantly advanced the field of machine learning and deep learning across various domains, including natural language processing, computer vision, and multimodal systems. This paper presents a comprehensive survey of attention mechanisms in Transformer architectures, ...
Read More
The attention mechanisms have significantly advanced the field of machine learning and deep learning across various domains, including natural language processing, computer vision, and multimodal systems. This paper presents a comprehensive survey of attention mechanisms in Transformer architectures, emphasizing their evolution, design variants, and domain-specific applications in NLP, computer vision, and multimodal learning. We categorize attention types by their goals like efficiency, scalability, and interpretability, and provide a comparative analysis of their strengths, limitations, and suitable use cases. This survey also addresses the lack of visual intuitions, offering a clearer taxonomy and discussion of hybrid approaches, such as sparse-hierarchical combinations. In addition to foundational mechanisms, we highlight hybrid approaches, theoretical underpinnings, and practical trade-offs. The paper identifies current challenges in computation, robustness, and transparency, offering a structured classification and proposing future directions. By comparing state-of-the-art techniques, this survey aims to guide researchers in selecting and designing attention mechanisms best suited for specific AI applications, ultimately fostering the development of more efficient, interpretable, and adaptable Transformer-based models.
A. Hadian; M. Bagherian; B. Fathi Vajargah
Abstract
Background: One of the most important concepts in cloud computing is modeling the problem as a multi-layer optimization problem which leads to cost savings in designing and operating the networks. Previous researchers have modeled the two-layer network operating problem as an Integer Linear Programming ...
Read More
Background: One of the most important concepts in cloud computing is modeling the problem as a multi-layer optimization problem which leads to cost savings in designing and operating the networks. Previous researchers have modeled the two-layer network operating problem as an Integer Linear Programming (ILP) problem, and due to the computational complexity of solving it jointly, they suggested a two-stage procedure for solving it by considering one layer at each stage.Aim: In this paper, considering the ILP model and using some properties of it, we propose a heuristic algorithm for solving the model jointly, considering unicast, multicast, and anycast flows simultaneously. Method: We first sort demands in decreasing order and use a greedy method to realize demands in order. Due to the high computational complexity of ILP model, the proposed heuristic algorithm is suitable for networks with a large number of nodes; In this regard, various examples are solved by CPLEX and MATLAB soft wares. Results: Our simulation results show that for small values of M and N CPLEX fails to find the optimal solution, while AGA finds a near-optimal solution quickly.Conclusion: The proposed greedy algorithm could solve the large-scale networks approximately in polynomial time and its approximation is reasonable.
Seyedeh R. Mahmudi Nezhad Dezfouli; Y. Kyani; Seyed A. Mahmoudinejad Dezfouli
Abstract
Due to the small size, low contrast, variable position, shape, and texture of multiple sclerosis lesions, one of the challenges of medical image processing is the automatic diagnosis and segmentation of multiple sclerosis lesions in Magnetic resonance images. Early diagnosis of these lesions in the first ...
Read More
Due to the small size, low contrast, variable position, shape, and texture of multiple sclerosis lesions, one of the challenges of medical image processing is the automatic diagnosis and segmentation of multiple sclerosis lesions in Magnetic resonance images. Early diagnosis of these lesions in the first stages of the disease can effectively diagnose and evaluate treatment. Also, automated segmentation is a powerful tool to assist professionals in improving the accuracy of disease diagnosis. This study uses modified adaptive multi-level conditional random fields and the artificial neural network to segment and diagnose multiple sclerosis lesions. Instead of assuming model coefficients as constant, they are considered variables in multi-level statistical models. This study aimed to evaluate the probability of lesions based on the severity, texture, and adjacent areas. The proposed method is applied to 130 MR images of multiple sclerosis patients in two test stages and resulted in 98% precision. Also, the proposed method has reduced the error detection rate by correcting the lesion boundaries using the average intensity of neighborhoods, rotation invariant, and texture for very small voxels with a size of 3-5 voxels, and it has shown very few false-positive lesions. The proposed model resulted in a high sensitivity of 91% with a false positive average of 0.5.
L. Falahiazar; V. Seydi; M. Mirzarezaee
Abstract
Many of the real-world issues have multiple conflicting objectives that the optimization between contradictory objectives is very difficult. In recent years, the Multi-objective Evolutionary Algorithms (MOEAs) have shown great performance to optimize such problems. So, the development of MOEAs will always ...
Read More
Many of the real-world issues have multiple conflicting objectives that the optimization between contradictory objectives is very difficult. In recent years, the Multi-objective Evolutionary Algorithms (MOEAs) have shown great performance to optimize such problems. So, the development of MOEAs will always lead to the advancement of science. The Non-dominated Sorting Genetic Algorithm II (NSGAII) is considered as one of the most used evolutionary algorithms, and many MOEAs have emerged to resolve NSGAII problems, such as the Sequential Multi-Objective Algorithm (SEQ-MOGA). SEQ-MOGA presents a new survival selection that arranges individuals systematically, and the chromosomes can cover the entire Pareto Front region. In this study, the Archive Sequential Multi-Objective Algorithm (ASMOGA) is proposed to develop and improve SEQ-MOGA. ASMOGA uses the archive technique to save the history of the search procedure, so that the maintenance of the diversity in the decision space is satisfied adequately. To demonstrate the performance of ASMOGA, it is used and compared with several state-of-the-art MOEAs for optimizing benchmark functions and designing the I-Beam problem. The optimization results are evaluated by Performance Metrics such as hypervolume, Generational Distance, Spacing, and the t-test (a statistical test); based on the results, the superiority of the proposed algorithm is identified clearly.
J.10.5. Industrial
Arezoo Zamany; Abbas Khamseh; Sayedjavad Iranbanfard
Abstract
The international transfer of high technologies plays a pivotal role in the transformation of industries and the transition to Industry 5.0 - a paradigm emphasizing human-centric, sustainable, and resilient industrial development. However, this process faces numerous challenges and complexities, necessitating ...
Read More
The international transfer of high technologies plays a pivotal role in the transformation of industries and the transition to Industry 5.0 - a paradigm emphasizing human-centric, sustainable, and resilient industrial development. However, this process faces numerous challenges and complexities, necessitating a profound understanding of its key variables and concepts. The present research aimed to identify and analyze these variables in the realm of high technology transfer in Industry 5.0. Following a systematic literature review protocol, 84 relevant articles published between 2017 and 2024 were selected based on predefined criteria including relevance to the research topic, publication quality, and citation impact. These articles were analyzed using a comprehensive text mining approach incorporating keyword extraction, sentiment analysis, topic modeling, and concept clustering techniques implemented through Python libraries including NLTK, SpaCy, TextBlob, and Scikit-learn. The results categorize the key variables and concepts into five main clusters: high technologies (including AI, IoT, and robotics), technology transfer mechanisms, Industry 5.0 characteristics, implementation challenges (such as cybersecurity risks and high adoption costs) and opportunities (including increased productivity and innovation potential), and regulatory frameworks. These findings unveil various aspects of the technology transfer process, providing insights for stakeholders while highlighting the critical role of human-technology collaboration in Industry 5.0. The study's limitations include potential bias from focusing primarily on English-language literature and the inherent constraints of computational text analysis in capturing context-dependent nuances. This research contributes to a deeper understanding of technology transfer dynamics in Industry 5.0, offering practical implications for policymaking and implementation strategies.
H.6.5.13. Signal processing
Samira Moghani; Hossein Marvi; Zeynab Mohammadpoory
Abstract
This study introduces a novel classification framework based on Deep Orthogonal Non-Negative Matrix Factorization (Deep ONMF), which leverages scalogram representations of phonocardiogram (PCG) signals to hierarchically extract structural features crucial for detecting valvular heart diseases (VHDs). ...
Read More
This study introduces a novel classification framework based on Deep Orthogonal Non-Negative Matrix Factorization (Deep ONMF), which leverages scalogram representations of phonocardiogram (PCG) signals to hierarchically extract structural features crucial for detecting valvular heart diseases (VHDs). Scalograms, generated via the Continuous Wavelet Transform (CWT), serve as the foundational input to the proposed feature extraction pipeline, which integrates them with Deep ONMF in a unified and segmentation-free architecture. The resulting scalogram–Deep ONMF framework is designed to hierarchically extract features through two complementary perspectives: Scale-Domain Analysis (SDA) and Temporal-Domain Analysis (TDA). These extracted features are then classified using shallow classifiers, with Random Forest (RF) achieving the best results, particularly when paired with SDA features based on the Bump wavelet. Experimental evaluations on two public PCG datasets—one with five heart sound classes and another with binary classification—demonstrate the effectiveness of the proposed method, achieving high classification accuracies of up to 98.40% and 97.23%, respectively, thereby confirming its competitiveness with state-of-the-art techniques. The results suggest that the proposed approach offers a practical and powerful solution for automated heart sound analysis, with potential applications beyond VHD detection.
H.3.8. Natural Language Processing
P. Kavehzadeh; M. M. Abdollah Pour; S. Momtazi
Abstract
Over the last few years, text chunking has taken a significant part in sequence labeling tasks. Although a large variety of methods have been proposed for shallow parsing in English, most proposed approaches for text chunking in Persian language are based on simple and traditional concepts. In this paper, ...
Read More
Over the last few years, text chunking has taken a significant part in sequence labeling tasks. Although a large variety of methods have been proposed for shallow parsing in English, most proposed approaches for text chunking in Persian language are based on simple and traditional concepts. In this paper, we propose using the state-of-the-art transformer-based contextualized models, namely BERT and XLM-RoBERTa, as the major structure of our models. Conditional Random Field (CRF), the combination of Bidirectional Long Short-Term Memory (BiLSTM) and CRF, and a simple dense layer are employed after the transformer-based models to enhance the model's performance in predicting chunk labels. Moreover, we provide a new dataset for noun phrase chunking in Persian which includes annotated data of Persian news text. Our experiments reveal that XLM-RoBERTa achieves the best performance between all the architectures tried on the proposed dataset. The results also show that using a single CRF layer would yield better results than a dense layer and even the combination of BiLSTM and CRF.
H.3. Artificial Intelligence
Amirhossein Khabbaz; Mansoor Fateh; Ali Pouyan; Mohsen Rezvani
Abstract
Autism spectrum disorder (ASD) is a collection of inconstant characteristics. Anomalies in reciprocal social communications and disabilities in perceiving communication patterns characterize These features. Also, exclusive repeated interests and actions identify ASD. Computer games have affirmative effects ...
Read More
Autism spectrum disorder (ASD) is a collection of inconstant characteristics. Anomalies in reciprocal social communications and disabilities in perceiving communication patterns characterize These features. Also, exclusive repeated interests and actions identify ASD. Computer games have affirmative effects on autistic children. Serious games have been widely used to elevate the ability to communicate with other individuals in these children. In this paper, we propose an adaptive serious game to rate the social skills of autistic children. The proposed serious game employs a reinforcement learning mechanism to learn such ratings adaptively for the players. It uses fuzzy logic to estimate the communication skills of autistic children. The game adapts itself to the level of the child with autism. For that matter, it uses an intelligent agent to tune the challenges through playtime. To dynamically evaluate the communication skills of these children, the game challenges may grow harder based on the development of a child's skills through playtime. We also employ fuzzy logic to estimate the playing abilities of the player periodically. Fifteen autistic children participated in experiments to evaluate the presented serious game. The experimental results show that the proposed method is effective in the communication skill of autistic children.
H.6.5.2. Computer vision
Kourosh Kiani; Razieh Rastgoo; Alireza Chaji; Sergio Escalera
Abstract
Image inpainting, the process of restoring missing or corrupted regions of an image by reconstructing pixel information, has recently seen considerable advancements through deep learning-based approaches. Aiming to tackle the complex spatial relationships within an image, in this paper, we introduce ...
Read More
Image inpainting, the process of restoring missing or corrupted regions of an image by reconstructing pixel information, has recently seen considerable advancements through deep learning-based approaches. Aiming to tackle the complex spatial relationships within an image, in this paper, we introduce a novel deep learning-based pre-processing methodology for image inpainting utilizing the Vision Transformer (ViT). Unlike CNN-based methods, our approach leverages the self-attention mechanism of ViT to model global contextual dependencies, improving the quality of inpainted regions. Specifically, we replace masked pixel values with those generated by the ViT, utilizing the attention mechanism to extract diverse visual patches and capture discriminative spatial features. To the best of our knowledge, this is the first instance of such a pre-processing model being proposed for image inpainting tasks. Furthermore, we demonstrate that our methodology can be effectively applied using a pre-trained ViT model with a pre-defined patch size, reducing computational overhead while maintaining high reconstruction fidelity. To assess the generalization capability of the proposed methodology, we conduct extensive experiments comparing our approach with four standard inpainting models across four public datasets. The results validate the efficacy of our pre-processing technique in enhancing inpainting performance, particularly in scenarios involving complex textures and large missing regions.
N. Nowrozian; F. Tashtarian
Abstract
Battery power limitation of sensor nodes (SNs) is a major challenge for wireless sensor networks (WSNs) which affects network survival. Thus, optimizing the energy consumption of the SNs as well as increasing the lifetime of the SNs and thus, extending the lifetime of WSNs are of crucial importance in ...
Read More
Battery power limitation of sensor nodes (SNs) is a major challenge for wireless sensor networks (WSNs) which affects network survival. Thus, optimizing the energy consumption of the SNs as well as increasing the lifetime of the SNs and thus, extending the lifetime of WSNs are of crucial importance in these types of networks. Mobile chargers (MCs) and wireless power transfer (WPT) technologies have played an important long role in WSNs, and much research has been done on how to use the MC to enhance the performance of WSNs in recent decades. In this paper, we first review the application of MCs and WPT technologies in WSNs. Then, forwarding issues the MC has been considered in the role of power transmitter in WSNs and the existing approaches are categorized, with the purposes and limitations of MC dispatching studied. Then an overview of the existing articles is presented and to better understand the contents, tables and figures are offered that summarize the existing methods. We examine them in different dimensions such as advantages and disadvantages etc. Finally, the future prospects of MC are discussed.
S. Ghandibidgoli; H. Mokhtari
Abstract
In many applications of the robotics, the mobile robot should be guided from a source to a specific destination. The automatic control and guidance of a mobile robot is a challenge in the context of robotics. So, in current paper, this problem is studied using various machine learning methods. Controlling ...
Read More
In many applications of the robotics, the mobile robot should be guided from a source to a specific destination. The automatic control and guidance of a mobile robot is a challenge in the context of robotics. So, in current paper, this problem is studied using various machine learning methods. Controlling a mobile robot is to help it to make the right decision about changing direction according to the information read by the sensors mounted around waist of the robot. Machine learning methods are trained using 3 large datasets read by the sensors and obtained from machine learning database of UCI. The employed methods include (i) discriminators: greedy hypercube classifier and support vector machines, (ii) parametric approaches: Naive Bayes’ classifier with and without dimensionality reduction methods, (iii) semiparametric algorithms: Expectation-Maximization algorithm (EM), C-means, K-means, agglomerative clustering, (iv) nonparametric approaches for defining the density function: histogram and kernel estimators, (v) nonparametric approaches for learning: k-nearest neighbors and decision tree and (vi) Combining Multiple Learners: Boosting and Bagging. These methods are compared based on various metrics. Computational results indicate superior performance of the implemented methods compared to the previous methods using the mentioned dataset. In general, Boosting, Bagging, Unpruned Tree and Pruned Tree (θ = 10-7) have given better results compared to the existing results. Also the efficiency of the implemented decision tree is better than the other employed methods and this method improves the classification precision, TP-rate, FP- rate and MSE of the classes by 0.1%, 0.1%, 0.001% and 0.001%.
H.3.7. Learning
Laleh Armi; Elham Abbasi
Abstract
In this paper, we propose an innovative classification method for tree bark classification and tree species identification. The proposed method consists of two steps. In the first step, we take the advantages of ILQP, a rotationally invariant, noise-resistant, and fully descriptive color texture feature ...
Read More
In this paper, we propose an innovative classification method for tree bark classification and tree species identification. The proposed method consists of two steps. In the first step, we take the advantages of ILQP, a rotationally invariant, noise-resistant, and fully descriptive color texture feature extraction method. Then, in the second step, a new classification method called stacked mixture of ELM-based experts with a trainable gating network (stacked MEETG) is proposed. The proposed method is evaluated using the Trunk12, BarkTex, and AFF datasets. The performance of the proposed method on these three bark datasets shows that our approach provides better accuracy than other state-of-the-art methods.Our proposed method achieves an average classification accuracy of 92.79% (Trunk12), 92.54% (BarkTex), and 91.68% (AFF), respectively. Additionally, the results demonstrate that ILQP has better texture feature extraction capabilities than similar methods such as ILTP. Furthermore, stacked MEETG has shown a great influence on the classification accuracy.
F.4.18. Time series analysis
Fatemeh Moodi; Amir Jahangard Rafsanjani; Sajjad Zarifzadeh; Mohammad Ali Zare Chahooki
Abstract
This article proposes a novel hybrid network integrating three distinct architectures -CNN, GRU, and LSTM- to predict stock price movements. Here with Combining Feature Extraction and Sequence Learning and Complementary Strengths can Improved Predictive Performance. CNNs can effectively identify short-term ...
Read More
This article proposes a novel hybrid network integrating three distinct architectures -CNN, GRU, and LSTM- to predict stock price movements. Here with Combining Feature Extraction and Sequence Learning and Complementary Strengths can Improved Predictive Performance. CNNs can effectively identify short-term dependencies and relevant features in time series, such as trends or spikes in stock prices. GRUs designed to handle sequential data. They are particularly useful for capturing dependencies over time while being computationally less expensive than LSTMs. In the hybrid model, GRUs help maintain relevant historical information in the sequence without suffering from vanishing gradient problems, making them more efficient for long sequences. LSTMs excel at learning long-term dependencies in sequential data, thanks to their memory cell structure. By retaining information over longer periods, LSTMs in the hybrid model ensure that important trends over time are not lost, providing a deeper understanding of the time series data. The novelty of the 1D-CNN-GRU-LSTM hybrid model lies in its ability to simultaneously capture short-term patterns and long-term dependencies in time series data, offering a more nuanced and accurate prediction of stock prices. The data set comprises technical indicators, sentiment analysis, and various aspects derived from pertinent tweets. Stock price movement is categorized into three categories: Rise, Fall, and Stable. Evaluation of this model on five years of transaction data demonstrates its capability to forecast stock price movements with an accuracy of 0.93717. The improvement of proposed hybrid model for stock movement prediction over existing models is 12% for accuracy and F1-score metrics.
N. Esfandian; F. Jahani bahnamiri; S. Mavaddati
Abstract
This paper proposes a novel method for voice activity detection based on clustering in spectro-temporal domain. In the proposed algorithms, auditory model is used to extract the spectro-temporal features. Gaussian Mixture Model and WK-means clustering methods are used to decrease dimensions of the spectro-temporal ...
Read More
This paper proposes a novel method for voice activity detection based on clustering in spectro-temporal domain. In the proposed algorithms, auditory model is used to extract the spectro-temporal features. Gaussian Mixture Model and WK-means clustering methods are used to decrease dimensions of the spectro-temporal space. Moreover, the energy and positions of clusters are used for voice activity detection. Silence/speech is recognized using the attributes of clusters and the updated threshold value in each frame. Having higher energy, the first cluster is used as the main speech section in computation. The efficiency of the proposed method was evaluated for silence/speech discrimination in different noisy conditions. Displacement of clusters in spectro-temporal domain was considered as the criteria to determine robustness of features. According to the results, the proposed method improved the speech/non-speech segmentation rate in comparison to temporal and spectral features in low signal to noise ratios (SNRs).
F. Rismanian Yazdi; M. Hosseinzadeh; S. Jabbehdari
Abstract
Wireless body area networks (WBAN) are innovative technologies that have been the anticipation greatly promote healthcare monitoring systems. All WBAN included biomedical sensors that can be worn on or implanted in the body. Sensors are monitoring vital signs and then processing the data and transmitting ...
Read More
Wireless body area networks (WBAN) are innovative technologies that have been the anticipation greatly promote healthcare monitoring systems. All WBAN included biomedical sensors that can be worn on or implanted in the body. Sensors are monitoring vital signs and then processing the data and transmitting to the central server. Biomedical sensors are limited in energy resources and need an improved design for managing energy consumption. Therefore, DTEC-MAC (Diverse Traffic with Energy Consumption-MAC) is proposed based on the priority of data classification in the cluster nodes and provides medical data based on energy management. The proposed method uses fuzzy logic based on the distance to sink and the remaining energy and length of data to select the cluster head. MATLAB software was used to simulate the method. This method compared with similar methods called iM-SIMPLE and M-ATTEMPT, ERP. Results of the simulations indicate that it works better to extend the lifetime and guarantee minimum energy and packet delivery rates, maximizing the throughput.
H.3. Artificial Intelligence
Hassan Haji Mohammadi; Alireza Talebpour; Ahamd Mahmoudi Aznaveh; Samaneh Yazdani
Abstract
Coreference resolution is one of the essential tasks of natural languageprocessing. This task identifies all in-text expressions that refer to thesame entity in the real world. Coreference resolution is used in otherfields of natural language processing, such as information extraction,machine translation, ...
Read More
Coreference resolution is one of the essential tasks of natural languageprocessing. This task identifies all in-text expressions that refer to thesame entity in the real world. Coreference resolution is used in otherfields of natural language processing, such as information extraction,machine translation, and question-answering.This article presents a new coreference resolution corpus in Persiannamed Mehr corpus. The article's primary goal is to develop a Persiancoreference corpus that resolves some of the previous Persian corpus'sshortcomings while maintaining a high inter-annotator agreement. Thiscorpus annotates coreference relations for noun phrases, namedentities, pronouns, and nested named entities. Two baseline pronounresolution systems are developed, and the results are reported. Thecorpus size includes 400 documents and about 170k tokens. Corpusannotation is done by WebAnno preprocessing tool.
H.3. Artificial Intelligence
Fariba Taghinezhad; Mohammad Ghasemzadeh
Abstract
Artificial neural networks are among the most significant models in machine learning that use numeric inputs. This study presents a new single-layer perceptron model based on categorical inputs. In the proposed model, every quality value in the training dataset receives a trainable weight. Input data ...
Read More
Artificial neural networks are among the most significant models in machine learning that use numeric inputs. This study presents a new single-layer perceptron model based on categorical inputs. In the proposed model, every quality value in the training dataset receives a trainable weight. Input data is classified by determining the weight vector that corresponds to the categorical values in it. To evaluate the performance of the proposed algorithm, we have used 10 datasets. We have compared the performance of the proposed method to that of other machine learning models, including neural networks, support vector machines, naïve Bayes classifiers, and random forests. According to the results, the proposed model resulted in a 36% reduction in memory usage when compared to baseline models across all datasets. Moreover, it demonstrated a training speed enhancement of 54.5% for datasets that contained more than 1000 samples. The accuracy of the proposed model is also comparable to other machine learning models.
Z. Imanimehr
Abstract
Peer-to-peer video streaming has reached great attention during recent years. Video streaming in peer-to-peer networks is a good way to stream video on the Internet due to the high scalability, high video quality, and low bandwidth requirements. In this paper the issue of live video streaming in peer-to-peer ...
Read More
Peer-to-peer video streaming has reached great attention during recent years. Video streaming in peer-to-peer networks is a good way to stream video on the Internet due to the high scalability, high video quality, and low bandwidth requirements. In this paper the issue of live video streaming in peer-to-peer networks which contain selfish peers is addressed. To encourage peers to cooperate in video distribution, tokens are used as an internal currency. Tokens are gained by peers when they accept requests from other peers to upload video chunks to them, and tokens are spent when sending requests to other peers to download video chunks from them. To handle the heterogeneity in the bandwidth of peers, the assumption has been made that the video is coded as multi-layered. For each layer the same token has been used, but priced differently per layer. Based on the available token pools, peers can request various qualities. A new token-based incentive mechanism has been proposed, which adapts the admission control policy of peers according to the dynamics of the request submission, request arrival, time to send requests, and bandwidth availability processes. Peer-to-peer requests could arrive at any time, so the continuous Markov Decision Process has been used.
M. Gordan; Saeed R. Sabbagh-Yazdi; Z. Ismail; Kh. Ghaedi; H. Hamad Ghayeb
Abstract
A structural health monitoring system contains two components, i.e. a data collection approach comprising a network of sensors for recording the structural responses as well as an extraction methodology in order to achieve beneficial information on the structural health condition. In this regard, data ...
Read More
A structural health monitoring system contains two components, i.e. a data collection approach comprising a network of sensors for recording the structural responses as well as an extraction methodology in order to achieve beneficial information on the structural health condition. In this regard, data mining which is one of the emerging computer-based technologies, can be employed for extraction of valuable information from obtained sensor databases. On the other hand, data inverse analysis scheme as a problem-based procedure has been developing rapidly. Therefore, the aforesaid scheme and data mining should be combined in order to satisfy increasing demand of data analysis, especially in complex systems such as bridges. Consequently, this study develops a damage detection methodology based on these strategies. To this end, an inverse analysis approach using data mining is applied for a composite bridge. To aid the aim, the support vector machine (SVM) algorithm is utilized to generate the patterns by means of vibration characteristics dataset. To compare the robustness and accuracy of the predicted outputs, four kernel functions, including linear, polynomial, sigmoid, and radial basis function (RBF) are applied to build the patterns. The results point out the feasibility of the proposed method for detecting damage in composite slab-on-girder bridges.
Document and Text Processing
Mina Tabatabaei; Hossein Rahmani; Motahareh Nasiri
Abstract
The search for effective treatments for complex diseases, while minimizing toxicity and side effects, has become crucial. However, identifying synergistic combinations of drugs is often a time-consuming and expensive process, relying on trial and error due to the vast search space involved. Addressing ...
Read More
The search for effective treatments for complex diseases, while minimizing toxicity and side effects, has become crucial. However, identifying synergistic combinations of drugs is often a time-consuming and expensive process, relying on trial and error due to the vast search space involved. Addressing this issue, we present a deep learning framework in this study. Our framework utilizes a diverse set of features, including chemical structure, biomedical literature embedding, and biological network interaction data, to predict potential synergistic combinations. Additionally, we employ autoencoders and principal component analysis (PCA) for dimension reduction in sparse data. Through 10-fold cross-validation, we achieved an impressive 98 percent area under the curve (AUC), surpassing the performance of seven previous state-of-the-art approaches by an average of 8%.
