Computing and Informatics (E-Journal - Institute of Informatics, SAS, Bratislava)
Not a member yet
1506 research outputs found
Sort by
Large Scale Fine-Tuned Transformers Models Application for Business Names Generation
Get PDFNatural language processing (NLP) involves the computer analysis and processing of human languages using a variety of techniques aimed at adapting various tasks or computer programs to linguistically process natural language. Currently, NLP is increasingly applied to a wide range of real-world problems. These tasks can vary from extracting meaningful information from unstructured data, analyzing sentiment, translating text between languages, to generating human-level text autonomously. The goal of this study is to employ transformer-based natural language models to generate high-quality business names. Specifically, this work investigates whether larger models, which require more training time, yield better results for generating relatively short texts, such as business names. To achieve this, we utilize different transformer architectures, including both freely available and proprietary models, and compare their performance. Our dataset comprises 250 928 observations of business names. Based on the perplexity metric, the top-performing model in our study is the GPT2-Medium model. However, our findings reveal a discrepancy between human evaluation and perplexity-based assessment. According to human evaluation, the best results are obtained using the GPT-Neo-1.3B model. Interestingly, the larger GPT-Neo-2.7B model yields poorer results, with its performance not being statistically different from that of the GPT-Neo-125M model, which is 20 times smaller
Temperature Matrix-Based Data Placement Optimization in Edge Computing Environment
Get PDFThe scale of data shows an explosive growth trend, with wide use of cloud storage. However, there are challenges such as network latency and energy consumption. The emergence of edge computing brings data close to the edge of the network, making it a good supplement to cloud computing. The spatiotemporal characteristics of data have been largely ignored in studies of data placement and storage optimization. To this end, a temperature matrix-based data placement method using an improved Hungarian algorithm (TEMPLIH) is proposed in this work. A temperature matrix is used to reflect the influence of data characteristics on its placement. A data replica matrix selection algorithm based on temperature matrix (RSA-TM) is proposed to meet latency requirements. Then, an improved Hungarian algorithm based on replica matrix (IHA-RM) is proposed, which satisfies the balance among the multiple goals of latency, cost, and load balancing. Compared with other data placement strategies, experiments show that the proposed method can effectively reduce the cost of data placement while meeting user access latency requirements and maintaining a reasonable load balance between edge servers. Further improvement is discussed and the idea of regional value is proposed
CTransNet: Convolutional Neural Network Combined with Transformer for Medical Image Segmentation
Get PDFThe Transformer has been widely used for many tasks in NLP before, but there is still much room to explore the application of the Transformer to the image domain. In this paper, we propose a simple and efficient hybrid Transformer framework, CTransNet, which combines self-attention and CNN to improve medical image segmentation performance. Capturing long-range dependencies at different scales. To this end, this paper proposes an effective self-attention mechanism incorporating relative position information encoding, which can reduce the time complexity of self-attention from O(n2) to O(n), and a new self-attention decoder that can recover fine-grained features in encoder from skip connection. This paper aims to address the current dilemma of Transformer applications: i.e., the need to learn induction bias from large amounts of training data. The hybrid layer in CTransNet allows the Transformer to be initialized as a CNN without pre-training. We have evaluated the performance of CTransNet on several medical segmentation datasets. CTransNet shows superior segmentation performance, robustness, and great promise for generalization to other medical image segmentation tasks
Computing Aspects of Simulation Based on Conservation Laws Conducted on HPC Cluster
Get PDFThe large amount of computing resources required for the simulation of complex natural processes demands a thorough analysis of the efficiency of the calculations and the conditions that influence it. This study investigates computing aspects of fire simulation conducted on a compute cluster. Current fire simulators based on principles of computational fluid dynamics are capable to realistically model majority of complex phenomena related to fire. Fire simulations are highly computationally demanding itself, however, they often lead to extensive parametrical studies requiring high performance computing systems. Smoke stratification and visibility during fire in a road tunnel with two emergency lay-bys are investigated by parametrical study comprising of 24 fire scenarios with the tunnel geometry modifications and various heat release rates and fire locations. Main tendencies of smoke spread in the downstream lay-by are identified and their mutual interactions are analysed. The simulation efficiency of particular simulations is analysed and the reasons of their varied elapsed times are investigated. The analysis indicates that the main reason of this variability are different jet fans velocities influenced by simulation scenario settings
EEG-EMG Analysis Method in Hybrid Brain Computer Interface for Hand Rehabilitation Training
Get PDFBrain-computer interfaces (BCIs) have demonstrated immense potential in aiding stroke patients during their physical rehabilitation journey. By reshaping the neural circuits connecting the patient’s brain and limbs, these interfaces contribute to the restoration of motor functions, ultimately leading to a significant improvement in the patient’s overall quality of life. However, the current BCI primarily relies on Electroencephalogram (EEG) motor imagery (MI), which has relatively coarse recognition granularity and struggles to accurately recognize specific hand movements. To address this limitation, this paper proposes a hybrid BCI framework based on Electroencephalogram and Electromyography (EEG-EMG). The framework utilizes a combination of techniques: decoding EEG by using Graph Convolutional LSTM Networks (GCN-LSTM) to recognize the subject’s motion intention, and decoding EMG by using a convolutional neural network (CNN) to accurately identify hand movements. In EEG decoding, the correlation between channels is calculated using Standardized Permutation Mutual Information (SPMI), and the decoding process is further explained by analyzing the correlation matrix. In EMG decoding, experiments are conducted on two task paradigms, both achieving promising results. The proposed framework is validated using the publicly available WAL-EEG-GAL (Wearable interfaces for hand function recovery Electroencephalography Grasp-And-Lift) dataset, where the average classification accuracies of EEG and EMG are 0.892 and 0.954, respectively. This research aims to establish an efficient and user-friendly EEG-EMG hybrid BCI, thereby facilitating the hand rehabilitation training of stroke patients
Experimental Evaluation of Cloud-Based Facial Emotion Recognition Services
Get PDFThe main goal of this paper is to perform an extensive analysis of the accuracy of six selected cloud-based facial emotion recognition services on three facial images datasets. The evaluation was performed on more than 33 000 images depicting eight different emotions. Results show that emotion recognition services show a varying level of accuracy over different types of datasets, having a lower accuracy for images of lower quality, but performing considerably better for images taken in ideal conditions. Based on these results we believe that cloud-based facial emotional recognition services do not have the expected accuracy for some use cases and therefore must be selected with care when developing a system that relies on emotion-based interactions
The Performance Analysis of the Thermal Discrete Element Method Computations on the GPU
Get PDFThe paper presents a GPU implementation of the thermal discrete element method (TDEM) and the comparative analysis of its performance. Several discrete element models for granular flows, the bonded particle model and the TDEM are considered for quantitative comparison of computational performance. The performance measured on NVIDIA(R) Tesla™ P100 GPU is compared with that attained by running the same OpenCL code on Intel(R) Xeon™ E5-2630 CPU with 20 cores. The presented GPU implementation of the TDEM increases the computing time of the bonded particle model only up to 30.6 % of the computing time of the simplest DEM model, which is an acceptable decrease in the performance required for solving coupled thermomechanical problems
Privacy Issue: From Static to Dynamic Online Social Networks
Get PDFToday's societies have become more dependent on social networks in terms of communications and interactions. These networks contain most of the people's activities, which can be public or even personal events. In the last decade, social networks have turned into more prominent platforms in managing and organizing public events. The Egyptian revolution in 2011 and the Ukrainian revolution in 2014 are good reflections of such events. However, it is not known how much the privacy issue of users is revealed in the reality as a consequence of their online interactions. In this work, we investigate the privacy issue in online social networks and its reflection on real life. Our dataset was extracted from the Facebook groups/pages that were involved in the 2019 Iraqi October revolution. Our approach generates a static network using the collected dataset. Then, we investigate the generated static network in terms of detecting potential anomalies. After that, we project the static network (including its characteristics) into a dynamic environment and generate a dynamic network for investigating the privacy issue in the real life. The contribution of this work lies in projecting a real-world static network into a dynamic environment aiming at investigating users' privacy in the real world. Finally, this kind of approach has not been given enough attention in the literature and it is therefore deeply investigated in this article
Portability of Interfaces in Healthcare EAI Environments
Get PDFEnterprise Application Integration (EAI) and HL7 (Health Level Seven) messaging are well established technologies in healthcare environments. Due to the age and longevity of HL7 standards (especially HL7 V2.x) and their widespread use, many interfaces outlive the middleware on which they run and must be ported to new systems. This often requires the entire code of the interface to be rewritten, which is associated with great effort and costs. This paper shows a generic EAI framework based on configuration and dependency injection for implementing reusable interfaces upfront and the results when applied to a real production EAI environment of an Austrian healthcare provider