Text Summarization might seem like a trivial problem, but in reality, is essential for data acquisition and understanding by compressing large amounts of text into specific a nd to-the-point summaries. This research i...
详细信息
This paper discusses the use of federated learning as a method for optimizing decision-making in communication systems. Federated learning is a machine learning technique that enables the training of models on decentr...
详细信息
Data Augmentation (DA) is an effective strategy to increase model generalisation. In Natural Language Processing (NLP), DA remains in its early stages, primarily due to the inherent sensitivity of textual data, which ...
详细信息
ISBN:
(数字)9798331530983
ISBN:
(纸本)9798331530990
Data Augmentation (DA) is an effective strategy to increase model generalisation. In Natural Language Processing (NLP), DA remains in its early stages, primarily due to the inherent sensitivity of textual data, which complicates the augmentation process. These complexities are further increased in multilingual contexts, as DA techniques effective in one language may not yield comparable results in others. As a result, manually identifying optimal DA techniques that preserve semantic integrity and are language agnostic remains a significant challenge. It emphasises the necessity for automating the text DA process with Explainable Artificial Intelligence (XAI), which offers a promising approach by calculating word importance scores to effectively guide the augmentation. In this paper, a comprehensive review of existing approaches in auto-text DA is presented, beginning with an exploration of the concept of auto-text DA, followed by a discussion on its cross-lingual applications, and concluding with the integration of XAI for semantic preservation. The paper highlights the critical need for further research to enhance the effectiveness and applicability of XAI in auto-text DA, enabling its use across diverse languages.
Biomedical signals are extremely difficult to analyze, mainly due to the non-stationary nature of these signals. Filtering does not always bring the desired results, because often the desired information is filtered o...
详细信息
With the ever-increasing popularity of pretrained Video-Language Models (VidLMs), there is a pressing need to develop robust evaluation methodologies that delve deeper into their visio-linguistic capabilities. To addr...
详细信息
The current paper proposes a new approach for peripheral speech emotion analysis and gender estimation incorporating the best machine learning architectures such as CNNs and LSTMs. Its correct depiction of emotions an...
详细信息
ISBN:
(数字)9798331523923
ISBN:
(纸本)9798331523930
The current paper proposes a new approach for peripheral speech emotion analysis and gender estimation incorporating the best machine learning architectures such as CNNs and LSTMs. Its correct depiction of emotions and the variation of the gendered speech to indicate manifold signal-to-noise ratio make the approach accommodate the difficulty inherent in recognizing emotions from the recording of voices as well as predicting the gender of a speaker. It is then able to ally with preprocessing and collection of a vast range of different speech samples, feature extraction of pitch, tone and spectral data. These features serve as input to the CNNs and LSTMs: LSTMs work well at temporal regular features and the CNNs work well in feature extraction by location (spatio feature). The approach is tested on the RAVDESS and IEMOCAP datasets and establishes a near perfect match to prior approaches with reference to recognition accuracy of emotion analysis and gender prediction. Thus, applying the identified dual method approach, which could be described by such characteristics as highly accurate data pre-preprocessing, clear indication of the features of interest, a different CNN-LSTM model joint training, it offers a reasonably good platform to gain certain insights into such characteristics of the speech as our experiment demonstrates. This work offers satisfactory configurations on concerns such as human computer relation, customer relation and mental health surveillance to suggest that other applications could be achieved in future.
Training summarization models requires substantial amounts of training data. However for less resourceful languages like Hungarian, openly available models and datasets are notably scarce. To address this gap our pape...
Electroencephalography (EEG) signals, reflecting human brain activity, hold potential beyond medical diagnosis, particularly in emotion recognition. Despite the development of machine learning models utilizing EEG dat...
Electroencephalography (EEG) signals, reflecting human brain activity, hold potential beyond medical diagnosis, particularly in emotion recognition. Despite the development of machine learning models utilizing EEG data for this purpose, achieving good enough accuracy remains a challenge due to signals complexity and non-stationary nature, especially in extracting effective features that encapsulate temporal and frequency information. This paper introduces a novel hand-crafted feature extraction technique that avoids conventional signal segmentation and analyzes the entire length of EEG signals. This method builds a convolutional network utilizing Wavelet Scattering Transform (WST) blocks, followed by deriving a comprehensive 17-feature set from the raw EEG data and WST scattering coefficients. This integrative set takes advantage of the WST’s ability to produce a signal representation that is stable against noise, invariant to time shifts, and captures both temporal and frequency components while also leveraging the intrinsic properties of the raw data, offering an alternative to the computational deep models. The integration of Linear Discriminant Analysis for dimensionality reduction and the K-Nearest Neighbors algorithm for classification, further refined by a majority voting mechanism across all channels, results in a robust classification framework. The proposed method is evaluated across GAMEEMO and DEAP datasets with two and four emotional classes, using Leave-One-Subject-Out validation, achieving classification accuracy exceeding 97%. The findings support the effectiveness of this approach in EEG-based emotion recognition. Furthermore, an ablation study on the two datasets is implemented to assess each component’s impact, revealing insights into the model’s effectiveness and improvement areas.
The article the efficiency of the Microgrid network when transitioning to a transactive power system that uses control algorithms called to optimize the distribution of power between sources of distributed generation ...
详细信息
The advent of the Internet has significantly stream-lined daily tasks through the rapid increase of online services. Everyday activities, such as purchasing goods and scheduling appointments with healthcare profession...
The advent of the Internet has significantly stream-lined daily tasks through the rapid increase of online services. Everyday activities, such as purchasing goods and scheduling appointments with healthcare professionals, have become more speedy, efficient and user-friendly with the integration of the Internet. The continuous improvement of online services has led to many people moving towards digital activities. As a result, it has heightened the recording of personal and payment transaction data across various storage mediums, including databases and log files. The protection and regulation of this sensitive data are imperative, aligning with the guidelines outlined in GDPR and PCI-DSS compliances. Recognizing exposed personal data poses a considerable challenge. This research introduces a novel approach to identifying payment card industry data (PCI) and personally identifiable information (PII). The research project proposes a machine learning-based text classification model utilizing the Convolutional Neural Network (CNN) model to discern PII and PCI data within a given text. The CNN model has been constructed and compared against Naive Bayes, Gradient Boost, Random Forest, and Support Vector Machine (SVM) models. The CNN model achieved the highest accuracy at 0.96 (96%). Additionally, the F1 scores for each class were significant, with PII scoring 0.94, PCI scoring 0.95, and Normal scoring 0.99. Following the model's construction and training, it was employed with the saved tokenizer's word indexes and label encoders in the developed classification tool. This tool successfully delivered the promised results, identifying exposed PII and PCI data.
暂无评论