Social biases in generative models have gained increasing attention. This paper proposes an automatic evaluation protocol for text-to-image generation, examining how gender bias originates and perpetuates in the generation process of Stable Diffusion. Using triplet prompts that vary by gender indicators, we trace presentations at several stages of the generation process and explore dependencies between prompts and images. Our findings reveal the bias persists throughout all internal stages of the generating process and manifests in the entire images. For instance, differences in object presence, such as different instruments and outfit preferences, are observed across genders and extend to overall image layouts. Moreover, our experiments demonstrate that neutral prompts tend to produce images more closely aligned with those from masculine prompts than with their female counterparts. We also investigate prompt-image dependencies to further understand how bias is embedded in the generated content. Finally, we offer recommendations for developers and users to mitigate this effect in text-to-image generation.
Photography, Computer applications to medicine. Medical informatics
Joko Siswanto, Sri Yulianto Joko Prasetyo, Sutarto Wijono
et al.
Accurate predictions of the number of public transport passengers on buses in each region are crucial for operations. They are required by the planning and management authority for bus public transport. A deep learning-based LSTM prediction model is proposed to predict the number of passengers in 4 bus public transportation areas (central, north, south, and west), evaluated by MSLE, MAPE, and SMAPE with dropout, neuron, and train-test variations. The CSV dataset obtained from Auckland Transport(AT) New Zealand metro patronage report on bus performance(1/01/2019-31/07/2023) is used for evaluation. The best prediction model was obtained from the lowest evaluation value and relatively fast time with a dropout of 0.2, 32 neurons, and train-test 80-20. The prediction model on training and testing data improves with the suitability of tuning for four predictions for the next 12 months with mutual fluctuations. The strong negative correlation is central-south, while the strong positive correlation is north-west. Predictions are less closely interconnected and dependent, namely central-south. With its potential to significantly impact policy-making, this prediction model can increase public transport mobility in each region, leading to a more efficient and accessible public transport system and ultimately enhancing the public's daily lives. This research has practical implications for public transport authorities, as it can guide them in making informed decisions about service planning and resource allocation.
Muhammadjon Tursunov, Khabibullo Sabirov, Ramazon Alikulov
et al.
The results obtained on the photoelectric battery (PVB) and photothermal battery (PTB) based on a new type of autonomous moving cooling system photothermal devices (PTD) are presented in this research work. This new type of device (PTD) is self-cooling and has the ability to provide hot water for the village residents while increasing the efficiency of the PTB. There was created a new experimental copy of the device with a power of 300 W based on the design of the PTD with a new type of cooling system with a power of 60 W mentioned in previous scientific research. It consists of a 180 W PVB, a 60A•h battery, a 2kW inverter, a 50A controller, a radiator for cooling hot water, 5 cooling fans, a pump, and a cart-shaped structure for their installation. It is possible to get results in two different situations in the experimental copy with a new type of cooling system. It is possible to increase the efficiency of PTD by fully using the cooling system, and to obtain hot water for the household without a sharp decrease in efficiency by partially using the cooling system. There is mentioned a study of hot water regimes with a temperature of 40-50 0C for the agricultural sector, depending on the intensity of solar radiation and ambient temperature. Preliminary tests showed that the power of the PTD differs from the power of the PVB by up to 70 W. This new PTD showed that it is possible to use it in many other cases, such as water supply, lighting, watching TV, listening to the radio, using a refrigerator, charging computers and phones in rural areas.
Abstract Plant diseases significantly threaten global agriculture, impacting crop yield and food security. Nearly 30% of the crop yield is lost due to plant diseases. Efficient identification and classification of plant diseases through computer vision techniques have become imperative for timely intervention. However, popular plant disease datasets often suffer from data imbalance, with certain classes underrepresented, hindering the performance of machine learning models. Traditional data augmentation methods, such as rotation and flipping, are limited in their effectiveness, especially when faced with imbalanced datasets. To address this limitation, we explore advanced data augmentation techniques, including Generative Adversarial Networks (GANs) such as CycleGAN and LeafGAN, which have shown promise in generating synthetic images. However, we propose an innovative approach of Object-based single Style Transfer on a single neural network for augmenting the plant disease dataset. This technique focuses on mitigating data imbalance issues within datasets, which can adversely affect the model’s ability to generalize across diverse classes. The proposed method is compared with state-of-the-art data augmentation techniques, highlighting its superiority in addressing data imbalance issues. Our approach aims to produce more realistic and diverse synthetic images, leading to improved model generalization and accuracy in plant disease classification tasks validated using different classifiers. The efficiency of our approach is validated through extensive experimentation and benchmarking against existing methods.
Computational linguistics. Natural language processing, Electronic computers. Computer science
Tahereh Vasei, Harshil Gediya, Maryam Ravan
et al.
This study investigates the neurophysiological effects of transcutaneous electroacupuncture stimulation (TEAS) on brain activity, using advanced machine learning techniques. This work analyzed the electroencephalograms (EEG) of 48 study participants, in order to analyze the brain’s response to different TEAS frequencies (2.5, 10, 80, and sham at 160 pulses per second (pps)) across 48 participants through pre-stimulation, during-stimulation, and post-stimulation phases. Our approach introduced several novel aspects. EEGNet, a convolutional neural network specifically designed for EEG signal processing, was utilized in this work, achieving over 95% classification accuracy in detecting brain responses to various TEAS frequencies. Additionally, the classification accuracies across the pre-stimulation, during-stimulation, and post-stimulation phases remained consistently high (above 92%), indicating that EEGNet effectively captured the different time-based brain responses across different stimulation phases. Saliency maps were applied to identify the most critical EEG electrodes, potentially reducing the number needed without sacrificing accuracy. A phase-based analysis was conducted to capture time-based brain responses throughout different stimulation phases. The robustness of EEGNet was assessed across demographic and clinical factors, including sex, age, and psychological states. Additionally, the responsiveness of different EEG frequency bands to TEAS was investigated. The results demonstrated that EEGNet excels in classifying EEG signals with high accuracy, underscoring its effectiveness in reliably classifying EEG responses to TEAS and enhancing its applicability in clinical and therapeutic settings. Notably, gamma band activity showed the highest sensitivity to TEAS, suggesting significant effects on higher cognitive functions. Saliency mapping revealed that a subset of electrodes (Fp1, Fp2, Fz, F7, F8, T3, T4) could achieve accurate classification, indicating potential for more efficient EEG setups.
The overcomplete convolutional structure for biological images and volume segmentation is an excellent solution to the problem in which traditional codec methods cannot accurately segment the boundary regions. Although such methods perform well, the drawback that convolutional operations do not effectively learn global and remote semantic information interactions must be addressed. Accordingly, a new image segmentation network, KTU-Net, is proposed for the medical image segmentation of liver tumors. The network structure constitutes three branches: 1)Kite-Net, which is an overcomplete convolutional network that learns to capture input details and precise edges; 2)U-Net, which learns high-level features; 3)Transformer, which learns sequential representations of input bodies and efficiently captures global multiscale information. KTU-Net is designed for both early and late fusion, and a hybrid loss function is adopted to guide network training to achieve increased stability. From extensive experimental results regarding the LiTS liver tumor segmentation dataset, KTU-Net achieves higher or similar segmentation accuracy than other advanced 3D medical image segmentation methods such as KiU-Net, TransBTS, and UNETR. Fusing the three branching features, the average Dice scores of liver tumors are improved by 0.7% and 2.1%, achieving increased quality of features learned by the network and more accurate segmentation results of liver tumors, thus providing a reliable basis for doctors to perform precise liver tumor cell assessments and treatment plans.
Abstract Deep neural networks have achieved great success in both computer vision and natural language processing tasks. How to improve the gradient flows is crucial in training very deep neural networks. To address this challenge, a gradient enhancement approach is proposed through constructing the short circuit neural connections. The proposed short circuit is a unidirectional neural connection that back propagates the sensitivities rather than gradients in neural networks from the deep layers to the shallow layers. Moreover, the short circuit is further formulated as a gradient truncation operation in its connecting layers, which can be plugged into the backbone models without introducing extra training parameters. Extensive experiments demonstrate that the deep neural networks, with the help of short circuit connection, gain a large margin of improvement over the baselines on both computer vision and natural language processing tasks. The work provides the promising solution to the low‐resource scenarios, such as, intelligence transport systems of computer vision, question answering of natural language processing.
Deep learning has been instrumental in solving difficult problems by automatically learning, from sample data, the rules (algorithms) that map an input to its respective output. Purpose: Perform land use landcover (LULC) classification using the training data of satellite imagery for Moscow region and compare the accuracy attained from different models. Methods: The accuracy attained for LULC classification using deep learning algorithm and satellite imagery data is dependent on both the model and the training dataset used. We have used state-of-the-art deep learning models and transfer learning, together with dataset appropriate for the models. Different methods were applied to fine tuning the models with different parameters and preparing the right dataset for training, including using data augmentation. Results: Four models of deep learning from Residual Network (ResNet) and Visual Geometry Group (VGG) namely: ResNet50, ResNet152, VGG16 and VGG19 has been used with transfer learning. Further training of the models is performed with training data collected from Sentinel-2 for the Moscow region and it is found that ResNet50 has given the highest accuracy for LULC classification for this region. Practical relevance: We have developed code that train the 4 models and make classification of the input image patches into one of the 10 classes (Annual Crop, Forest, Herbaceous Vegetation, Highway, Industrial, Pasture, Permanent Crop, Residential, River, and Sea&Lake).
Rhizomer helps researchers and practitioners explore knowledge graphs available as Semantic Web data by performing the three data analysis tasks: overview, zoom and filter, and details-on-demand. This approach makes it easier for users to get an idea about the overall structure and intricacies of a dataset, when compared to existing approaches and even without prior knowledge. Rhizomer is helpful for data reusers, who want to know about the reuse opportunities of a given dataset, and for knowledge graph creators, who can check if the generated data follow their expectations. Rhizomer has been applied in many scenarios, from research and commercial projects to teaching.
We created a hand texture resource (with different skin tone versions as well as non-human hands) for use in virtual reality studies. This makes it easier to run lab and remote studies where the hand representation is matched to the participants’ own skin tone. We validate that the virtual hands with our textures align with participants’ view of their own real hands and allow to create VR applications where participants have an increased sense of body ownership. These properties are critical for a range of VR studies, such as of immersion.
Sharaf Alhomdy, Fursan Thabit, Fua'ad Hasan Abdulrazzak
et al.
The contagion of the Coronavirus (COVID-19) led to a global lockdown that put governments in emergency mode. With the total number of positive cases worldwide exceeding the 97.46 million mark, social distancing appears to be the only effective strategy to contain the virus at the moment. As a result, companies face obstacles and find it difficult to respond to this current challenge of remote working. The impact of the novel COVID-19 has created many new challenges, and many of us have had to adopt new ways of working. With the need for accessing to critical applications and the scalability of the infrastructure, cloud computing is emerging as an underlying technology. The cloud technology had a major role in fighting the epidemic; it becomes a salvation for governments and organizations in numerous fields of life, education, health, industry, communication, remote surveillance, and more information. Therefore, this study presents the benefits, characteristics and applications of cloud computing and explains how the cloud contributes to improving life in all regions of the world during COVID-19. It shows that the cloud computing helps countries in combating COVID 19, in education and health sectors, also in the economic and commercial aspects. It investigates the current state by distributing an online questionnaire to various people of academic and non-academic backgrounds in different places over the world in the ICT and education sectors. The results showed that there is an effective role for cloud computing during the epidemic.
In this paper, we conduct an in-depth study of Japanese keyword extraction from news reports, train external computer document word sets from text preprocessing into word vectors using the Ship-gram model in the deep learning tool Word2Vec, and calculate the cosine distance between word vectors. In this paper, the sliding window in TextRank is designed to connect internal document information to improve the in-text semantic coherence. The main idea is to use not only the statistical and structural features of words but also the semantic features of words extracted through word-embedding techniques, i.e., multifeature fusion, to obtain the importance weights of words themselves and the attraction weights between words and then iteratively calculate the final weight of each word through the graph model algorithm to determine the extracted keywords. To verify the performance of the algorithm, extensive simulation experimental studies were conducted on three different types of datasets. The experimental results show that the proposed keyword extraction algorithm can improve the performance by a maximum of 6.45% and 20.36% compared with the existing word frequency statistics and graph model methods, respectively; MF-Rank can achieve a maximum performance improvement of 1.76% compared with PW-TF.
Pantelis Koutroumpis, Aija Leiponen, Llewellyn D. W. Thomas
An analysis of patenting history from 1850 to 2010 to detect long-term patterns of knowledge spillovers via prior-art citations of patented inventions.
Tushar Rajvanshi, Maria Antonia Maisto, Angela Dell’Aversano
et al.
This paper deals with the problem of estimating the RCS from near-field data by image-based approaches. In particular, a rigorous focusing procedure based on a weighted adjoint scheme, which is also applicable to an arbitrary measurement curve, is developed. The developed formalism allows us to address the important question concerning the need to employ a multi-frequency configuration to estimate the RCS. Accordingly, it is shown that if RCS is required at a given frequency, then the target image obtained solely at such a frequency can be exploited provided that the spatial truncation arising from the size of the investigated area is properly taken into account.
Photography, Computer applications to medicine. Medical informatics
The study investigated the relationship between employees’ expectation (EE) and organizational silence (OS). Design/Methodology: A survey of one hundred and eight (108) working class Professional Master Students was carried out eliciting responses through a self- constructed instrument that has Cronbach alpha of 0.864 and 0.825 reliability values for employees’ expectation and organizational silence respectively. Both Pearson product moment correlation and multiple regressions statistics were used to test the stated research hypotheses. Findings: It was found that employees’ expectations have a strong positive and statistically significant relationship with organization silence. The indices of employee expectations: employee control (EC), employee ownership (EO) and employee appreciation (EA) separately correlate positively and significantly with organization silence. The study also showed that they are strong predictors of organizational silence except employee control that is somewhat a weak predictor. In combination however, employee expectation is a strong, positive and significant predictor of organizational silence. The influence of employees’ expectations of control, appreciation and ownership explained 30.5% of organization silence. Conclusion/recommendation: The study established that the types of organization silence are not limited to the three existing one of quiescent, acquiescent and pro-social but also include accrual benefits. Also, the motives for organization silence of resignation, fear and other-oriented, were extended to include self-oriented benefits. Research implications: Organization scholars, business owners and researchers should seek for the expectations of the employees as they contribute to change and work related improvement in the quest to stem the tide of silence behaviour climate in business organization.
Green technology has drawn a huge amount of attention with the development of the modern world. Similarly with the development in communication technology the industries and researchers are focusing to make this communication as green as possible. In cellular technology the evolution of 5G is the next step to fulfil the user demands and it will be available to the users in 2020. This will increase the energy consumption by which will result in excess emission of co2. In this paper different techniques for the green communication technology and some challenges are discussed. These techniques include device-to-device communication (D2D), massive Multiple-Input Multiple-Output (MIMO) systems, heterogeneous networks (HetNets) and Green Internet of Things (IoT).
Nowadays, in the after-treatment of diesel exhaust gas, a diesel particulate filter (DPF) has been used to trap nano-particles of the diesel soot. However, as there are more particles inside the filter, the pressure which corresponds to the filter backpressure increases, which worsens the fuel consumption rate, together with the abatement of the available torque. Thus, a filter with lower backpressure would be needed. To achieve this, it is necessary to utilize the information on the phenomena including both the soot transport and its removal inside the DPF, and optimize the filter substrate structure. In this paper, to obtain useful information for optimization of the filter structure, we tested seven filters with different porosities and pore sizes. The porosity and pore size were changed systematically. To consider the soot filtration, the particle-laden flow was simulated by a lattice Boltzmann method (LBM). Then, the flow field and the pressure change were discussed during the filtration process.
Parmi les grands réservoirs publics de la fin de la période médiévale, que compte encore la ville d'Alexandrie, l'un d'entre eux fait l'objet d'une étude architecturale et archéologique approfondie. Entreprise en 2003 par le Centre d'Études Alexan-
drines (CEAlex), en collaboration avec le Service égyptien des Antiquités, l'analyse de la citerne el-Nabih a fait appel, dans un premier temps, à la mise en œuvre des outils traditionnellement utilisés par les archéologues et les architectes sur les chantiers de fouilles. Ce n'est qu'à partir du printemps 2008, à la suite de notre rencontre avec Yves Egels, Ingénieur général à l'Institut Géographique National (IGN), et à la dotation par le Centre National de la Recherche Scientifique (CNRS) d'un scanner laser 3D (Trimble R GXTM), que notre équipe a pu envisager d'utiliser la photogrammétrie et la
lasergrammétrie en complément des levés manuels précédemment réalisés. L'irruption récente de ces technologies dans le domaine de l'archéologie nous a permis d'entrevoir, avec l'exemple du chantier de la citerne el-Nabih, l'étendue des potentialités que nous apportaient ces nouveaux outils dans nos travaux de recherche. Une fois nos besoins clairement définis, c'est grâce à l'échange entre les utilisateurs que nous étions devenus et les «producteurs de méthodes», que nous avons pu optimiser et adapter l'outil afin qu'il réponde au mieux à nos attentes. Cette communication porte sur l'utilisation
de la photogrammétrie et de la lasergrammétrie dans l'étude architecturale et archéologique de la citerne el-Nabih. Nous abordons, en premier lieu, les caractéristiques du site, les outils et méthodes mis en œuvre et les difficultés rencontrées.
Dans un second temps, nous faisons le point sur l'apport de ces techniques avant de conclure sur les perspectives qu'elles offrent dans notre domaine de recherche.
Instruments and machines, Applied optics. Photonics
Two goodness-of-fit tests for copulas are being investigated. The first one deals with the case of elliptical copulas and the second one deals with independent copulas. These tests result from the expansion of the projection pursuit methodology that we will introduce in the present article. This method enables us to determine on which axis system these copulas lie as well as the exact value of these very copulas in the basis formed by the axes previously determined irrespective of their value in their canonical basis. Simulations are also presented as well as an application to real datasets.