Quantum theory-inspired inter-sentence semantic interaction model for textual adversarial defense DOI Creative Commons
Jiacheng Huang, Long Chen,

Xiaoyin Yi

et al.

Complex & Intelligent Systems, Journal Year: 2024, Volume and Issue: 11(1)

Published: Dec. 30, 2024

Abstract Deep neural networks have a recognized susceptibility to diverse forms of adversarial attacks in the field natural language processing and such security issue poses substantial risks erodes trust artificial intelligence applications among people who use them. Meanwhile, quantum theory-inspired models that represent word composition as mixture words modeled non-linear semantic interaction. However, modeling without considering interaction between sentences current literature does not exploit potential probabilistic description for improving robustness settings. In present study, novel inter-sentence model is proposed enhancing via fusing contextual semantics. More specifically, it analyzed why humans are able understand textual examples, crucial point observed adept at associating information from context comprehend paragraph. Guided by this insight, input text segmented into subsentences, with simulating comprehension representing each subsentence particle within system, utilizing density matrix interactions. A loss function integrating cross-entropy orthogonality losses employed encourage measurement states. Comprehensive experiments conducted validate efficacy methodology, results underscore its superiority over baseline even commercial based on large terms accuracy across attack scenarios, showing approach under attacks.

Language: Английский

Effective text classification using BERT, MTM LSTM, and DT DOI
Saman Jamshidi,

Mahin Mohammadi,

Saeed Bagheri

et al.

Data & Knowledge Engineering, Journal Year: 2024, Volume and Issue: 151, P. 102306 - 102306

Published: April 21, 2024

Language: Английский

Citations

15
Evolving techniques in sentiment analysis: a comprehensive review DOI Creative Commons

M. R. Pavan Kumar,

Lal Khan, Hsien-Tsung Chang

et al.

PeerJ Computer Science, Journal Year: 2025, Volume and Issue: 11, P. e2592 - e2592

Published: Jan. 28, 2025

With the rapid expansion of social media and e-commerce platforms, an unprecedented volume user-generated content has emerged, offering organizations, governments, researchers invaluable insights into public sentiment. Yet, vast unstructured nature this data challenges traditional analysis methods. Sentiment analysis, a specialized field within natural language processing, evolved to meet these by automating detection categorization opinions emotions in text. This review comprehensively examines evolving techniques sentiment detailing foundational processes such as gathering feature extraction. It explores spectrum methodologies, from classical word embedding machine learning algorithms recent contextual advanced transformer models like Generative Pre-trained Transformer (GPT), Bidirectional Encoder Representations Transformers (BERT), T5. critical comparison methods, article highlights their appropriate uses limitations. Additionally, provides thorough overview current trends, future directions, exploration unresolved challenges. By synthesizing developments, equips with solid foundation for assessing state guiding advancements dynamic field.

Language: Английский

Citations

1
Sentiment Analysis in Employee Experience Using Natural Language Processing and Machine Learning DOI
Tuğçe Şimşek, Ahmet Bahadır Şimşek

Advances in human resources management and organizational development book series, Journal Year: 2025, Volume and Issue: unknown, P. 309 - 346

Published: Jan. 24, 2025

This chapter focuses on the use of sentiment analysis in handling employee experience. When organisations start thinking about experience their employees as a factor that influences performance and turnover, provides quantitative way measuring emotions, satisfaction engagement employees. attentions several NLP techniques can be employed feedback which include tokenization, stop-word removal vectorization. It also looks at how other machine learning models such Naive Bayes, Support Vector Machines, Long Short-Term Memory used to categorize emotions positive, negative, or neutral. In addition, problem language, culture, data bias is described, ways solve them are described. The future potential real-time emotion along with organizational KPIs for improving management employees' depicted.

Language: Английский

Citations

0
A joint-training topic model for social media texts DOI Creative Commons
Simeng Qin, Mingli Zhang,

Haiju Hu

et al.

Humanities and Social Sciences Communications, Journal Year: 2025, Volume and Issue: 12(1)

Published: March 1, 2025

Language: Английский

Citations

0
Arduino-Based Alcohol Detection Device: Enhancing Safety in Vehicle Operation through Sensor Technology DOI Open Access
Shanmugasundaram Hariharan,

S. Barath,

R. Monica

et al.

Asian Journal of Applied Science and Technology, Journal Year: 2024, Volume and Issue: 08(01), P. 133 - 150

Published: Jan. 1, 2024

In this context, an Arduino-based alcohol detector is example of a device that has the potential to detect presence in surrounding environment. It possible use tool check results individuals who have drank while operating motor vehicle. A MQ-3 sensor utilized by order ascertain whether or not readily available. The component constitutes heats layer conducting material simultaneously measuring resistance substrate. There change whenever it subjected scents vapors alcohol. Signals both digital and analogue types can be obtained from sensor. distinction made between two very plain way. are only conceivable states output take communicating with microcontroller. These high low, which means they represent values 1 0, respectively. An analog signal, on other hand, received microcontroller, provides indication amount present environment utilizing wide range values, ranging 0 1023. LED, sensor, Arduino Uno components required construct device. confined areas for showing straightforward applications small scale, performs admirably. process installing gadget vehicles yet another approach may used lessen number accidents caused drunk driving. addition being user-friendly easy repair, level sensitivity

Language: Английский

Citations

0
Harnessing Linguistic Analysis for ADHD Diagnosis Support: A Stylometric Approach to Self-Defining Memories DOI Open Access
Florian Cafiero, Juan Barrios Rudloff, Simon Gabay

et al.

Published: April 24, 2024

This study explores the potential of stylometric analysis in identifying Self-Defining Memories (SDMs) authored by individuals with Attention-Deficit/Hyperactivity Disorder (ADHD) versus a control group. A sample 198 SDMs were written 66 adolescents and then analysed using Support Vector Classifiers (SVC). The included variety linguistic features such as character 3-grams, function words, sentence length, or lexical richness among others. It also metadata about participants (gender, age) their (self-reported sentiment after recalling memories). results reveal promising ability to accurately classify SDMs, perfect prediction (F1=1.0) contextually simpler setup text-by-text prediction, satisfactory levels precision (F1 = 0.77) when predicting individual individual. Such highlight significant role that characteristics play reflecting distinctive cognitive patterns associated ADHD. While not substitute for professional diagnosis, textual offers supportive avenue early detection deeper understanding

Language: Английский

Citations

0
Quantum theory-inspired inter-sentence semantic interaction model for textual adversarial defense DOI Creative Commons
Jiacheng Huang, Long Chen,

Xiaoyin Yi

et al.

Complex & Intelligent Systems, Journal Year: 2024, Volume and Issue: 11(1)

Published: Dec. 30, 2024

Abstract Deep neural networks have a recognized susceptibility to diverse forms of adversarial attacks in the field natural language processing and such security issue poses substantial risks erodes trust artificial intelligence applications among people who use them. Meanwhile, quantum theory-inspired models that represent word composition as mixture words modeled non-linear semantic interaction. However, modeling without considering interaction between sentences current literature does not exploit potential probabilistic description for improving robustness settings. In present study, novel inter-sentence model is proposed enhancing via fusing contextual semantics. More specifically, it analyzed why humans are able understand textual examples, crucial point observed adept at associating information from context comprehend paragraph. Guided by this insight, input text segmented into subsentences, with simulating comprehension representing each subsentence particle within system, utilizing density matrix interactions. A loss function integrating cross-entropy orthogonality losses employed encourage measurement states. Comprehensive experiments conducted validate efficacy methodology, results underscore its superiority over baseline even commercial based on large terms accuracy across attack scenarios, showing approach under attacks.

Language: Английский

Citations

0