Article
Education & Educational Research
Mohammed A. Al-Sharafi, Mostafa Al-Emran, Mohammad Iranmanesh, Noor Al-Qaysi, Noorminshah A. Iahad, Ibrahim Arpaci
Summary: This research develops a theoretical model to understand the sustainable use of AI-based chatbots for educational purposes. Through analyzing data from an online survey of 448 university students, using a hybrid SEM-ANN approach, the study confirms the hypotheses in the developed model. The results highlight the significant impact of knowledge application on the sustainable use of chatbots.
INTERACTIVE LEARNING ENVIRONMENTS
(2022)
Article
Computer Science, Theory & Methods
Kimiya Keyvan, Jimmy Xiangji Huang
Summary: The recent advancements in Natural Language Processing technology have facilitated more conversation-based human and machine interactions. Conversational Search Systems, such as chatbots and Virtual Personal Assistants, have provided a limited platform for users and devices to communicate. In this context, users often struggle to express their information needs in a way that machines can understand. This survey explores different approaches, evaluation methods, and future prospects for resolving ambiguous queries in conversational search technology. The study highlights the significance of understanding user queries for retrieving relevant documents and meeting their needs by predicting potential requests.
ACM COMPUTING SURVEYS
(2023)
Article
Automation & Control Systems
Baoxun Wang, Zhen Xu, Huan Zhang, Kexin Qiu, Deyuan Zhang, Chengjie Sun
Summary: This paper presents a new methodology for modeling the local semantic distribution of responses to a given query in the human-conversation corpus, and explores a specified adversarial learning mechanism for training Neural Response Generation (NRG) models. Experimental results demonstrate that modeling the reasonable local distribution of the query-response corpus is crucial for adversarial NRG, and the proposed LocalGAN shows promise in improving both training stability and the quality of generated results.
JOURNAL OF MACHINE LEARNING RESEARCH
(2021)
Article
Green & Sustainable Science & Technology
Dominik Siemon, Rangina Ahmad, Henrik Harms, Triparna de Vreede
Summary: Artificial intelligence technologies can assist individuals in managing anxiety and improving their mental health. Personality-Adaptive Conversational Agents (PACAs) are able to infer users' personality traits and adapt accordingly, offering potential benefits for mental health care but raising concerns about trust and privacy.
Article
Computer Science, Cybernetics
Mengyao Li, Amudha V. Kamaraj, John D. Lee
Summary: This paper models dynamic trust evolution in conversations using a novel method called trajectory epistemic network analysis (T-ENA). The findings show that agent reliability significantly affects people's conversations in the analytic process of trust, and trust dynamics manifest through conversation topic diversity and flow.
INTERNATIONAL JOURNAL OF HUMAN-COMPUTER INTERACTION
(2023)
Article
Computer Science, Information Systems
Yaqiong Qiao, Xiangyang Luo, Jiangtao Ma, Meng Zhang, Chenliang Li
Summary: This paper proposes a novel Twitter user geolocation method based on heterogeneous relationship modeling and representation learning. By constructing two heterogeneous graphs and capturing the complex topological structure, the method generates natural language-like node sequences. Experimental results show that this method outperforms state-of-the-art methods.
INFORMATION SCIENCES
(2023)
Article
Thermodynamics
Ali Sohani, Siamak Hoseinzadeh, Saman Samiezadeh, Ivan Verhaert
Summary: An enhanced design for a solar still desalination system was employed to develop artificial neural network (ANN) models, with FF and RBF types identified as the best structures for predicting distillate production and water temperature. Error analysis on data not used for ANN model development showed varying errors in different months.
JOURNAL OF THERMAL ANALYSIS AND CALORIMETRY
(2022)
Article
Energy & Fuels
Rasool Amiri Kolajoobi, Mohammad Emami Niri, Shahram Amini, Yousof Haghshenas
Summary: Well placement optimization is crucial for field management and economy, but it requires significant computational time and resources. This article presents a data-driven proxy model in an artificial intelligence framework to address this issue. The model uses a sequence of static proxies trained in a time-dependent manner to optimize the oil recovery for different well configurations, and the results show successful achievement of the objectives.
JOURNAL OF ENERGY RESOURCES TECHNOLOGY-TRANSACTIONS OF THE ASME
(2023)
Article
Computer Science, Information Systems
Edona Elshan, Philipp Ebel, Matthias Soellner, Jan Marco Leimeister
Summary: Smart personal assistants (SPAs) like Alexa promise individualized user interactions, but using low code environments alone does not guarantee a good conversation. In this study, we propose the SPADE method to support domain experts in developing effective SPAs with conversational and anthropomorphic design steps. Our proof of concept and proof of value results show that SPADE is useful in different domains beyond private set-ups.
JOURNAL OF MANAGEMENT INFORMATION SYSTEMS
(2023)
Article
Chemistry, Multidisciplinary
S. K. Safdar Hossain, Bamidele Victor Ayodele, Zaid Abdulhamid Alhulaybi, Muhammad Mudassir Ahmad Alwi
Summary: This study explores the feasibility of using machine learning to model biohydrogen production from waste glycerol. The findings show that the multilayer perceptron neural network has better predictive performance, and the combination of activation functions in the hidden and outer layers and the optimization algorithm type significantly affect the model's performance. Waste glycerol is the most significant input variable in predicting biohydrogen production.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Artificial Intelligence
Ruiqin Wang, Jungang Lou, Yunliang Jiang
Summary: This paper introduces a time-aware Lightweight Graph Convolutional Attention Network (LightGCAN) that can effectively capture static and dynamic user preferences using different GNN strategies. The static and dynamic user preferences are combined and fed into a dual-channel Deep Neural Network (DNN) model for feature interaction learning and matching score prediction. Experimental results show that LightGCAN outperforms the state-of-the-art recommendation methods.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Engineering, Chemical
Zhan Zhao, Mingzhi Jin, Fang Qin, Simon X. Yang
Summary: In this study, a biological neural network (BNN) approach was proposed to model the distribution of particles on a vibrating screen surface, improving the similarity with DEM simulation results by optimizing the BNN model coefficients. The results demonstrate that the BNN approach has clear advantages in solving the modeling problem when particles are fed to screens with multiple areas and non-uniform rates.
Article
Computer Science, Artificial Intelligence
Angelo Cafaro, Brian Ravenet, Catherine Pelachaud
Summary: This paper examines the nonverbal reactions of the interruptee during conversational interruptions and proposes a novel technique driven by an evolutionary algorithm to build a computational model for ECAs to manage user's interruptions. It also presents a methodology for building an ECA behavioral model through an interactive study driven by an evolutionary algorithm.
IEEE TRANSACTIONS ON AFFECTIVE COMPUTING
(2022)
Article
Engineering, Chemical
Amin Mahdavi-Meymand, Wojciech Sulisz
Summary: In this study, ARIA models were developed to enhance the prediction of boiling point rise in a multi-stage flash desalination system. The ARIA models showed greater accuracy and increased prediction efficiency compared to regular models. The ARIA-ANFIS model performed the best, reducing the error in RF predictions by 69.66%.
Article
Energy & Fuels
Kuang-Yow Lian, Yong-Jie Hong, Che-Wei Chang, Yu-Wei Su
Summary: This paper proposes a new method called the backward modeling approach (BMA) to achieve optimal chiller loading (OCL) for reducing energy consumption in industries with multiple-chillers of different efficiency. The developed OCL regulator (OCLR) based on the novel BMA approach consists of a conditional generative network (cGAN) and a deep neural network (DNN). By using the developed OCLR, the chilled water supply temperature can be set to achieve the desired energy saving. The experimental results showed that the data-driven OCLR based on BMA has high performance and was able to save significant energy.
Article
Computer Science, Interdisciplinary Applications
Ines M. Fernandez-Guerrero, Zoraida Callejas, David Griol, Antonio Fernandez-Cano
Article
Computer Science, Artificial Intelligence
David Griol, Zoraida Callejas, Jose Manuel Molina, Araceli Sanchis
Summary: Conversational systems have become indispensable for billions of users, and designing the dialogue model is one of the most complex tasks in system development. Rule-based systems struggle to adapt to phenomena not considered during design.
Article
Computer Science, Artificial Intelligence
Lukas Mateju, David Griol, Zoraida Callejas, Jose Manuel Molina, Araceli Sanchis
Summary: This paper evaluates different configurations of deep-learned dialog management using three dialog corpora from different application domains, identifying key factors that can impact accuracy in the results.
Article
Chemistry, Analytical
Cristina Luna-Jimenez, David Griol, Zoraida Callejas, Ricardo Kleinlein, Juan M. Montero, Fernando Fernandez-Martinez
Summary: This paper proposes a multimodal emotion recognition system based on speech and facial information, achieving high accuracy. The study demonstrates that combining different modalities and employing fusion strategies can improve system performance.
Article
Computer Science, Interdisciplinary Applications
Pablo Canas, David Griol, Zoraida Callejas
Summary: Conversational interfaces have become ubiquitous in both personal and industrial environments, improving access to services and saving costs through automation and enhanced customer support. However, designing the dialog model for system responses remains a challenging task, especially for complex conversational interactions.
JOURNAL OF COMPUTATIONAL SCIENCE
(2021)
Article
Chemistry, Multidisciplinary
Cristina Luna-Jimenez, Ricardo Kleinlein, David Griol, Zoraida Callejas, Juan M. Montero, Fernando Fernandez-Martinez
Summary: This paper proposes an automatic emotion recognizer system consisting of a speech emotion recognizer and a facial emotion recognizer. The pre-trained xlsr-Wav2Vec2.0 transformer is evaluated for the speech emotion recognizer, and the best results are achieved when fine-tuning the whole model. The facial emotion recognizer shows slightly better performance with sequential models compared to static models. Error analysis suggests that visual systems can be improved with a detector of high-emotional load frames. Finally, combining these two modalities with a late fusion strategy improves the system's performance.
APPLIED SCIENCES-BASEL
(2022)
Article
Chemistry, Multidisciplinary
Fernando Fernandez-Martinez, Cristina Luna-Jimenez, Ricardo Kleinlein, David Griol, Zoraida Callejas, Juan Manuel Montero
Summary: Intent recognition is crucial for task-oriented conversational systems, helping to understand user goals and provide appropriate responses. Research shows that transformers model performs the best, and inserting unseen domain words can effectively extend the vocabulary of the model.
APPLIED SCIENCES-BASEL
(2022)
Article
Multidisciplinary Sciences
Ksenia Kharitonova, Zoraida Callejas, David Perez-Fernandez, Asier Gutierrez-Fandino, David Griol
Summary: The ChatSubs dataset is a dialogue dataset containing dialogue data in Spanish and three of Spain's co-official languages (Catalan, Basque, and Galician). It has been obtained from OpenSubtitles and processed to generate segmented dialogues and turns. With over 20 million dialogues and 96 million turns, it is one of the largest dialogue corpus available, making it an ideal resource for research teams interested in training dialogue models in Spanish, Catalan, Basque, and Galician.
Proceedings Paper
Computer Science, Artificial Intelligence
Rafael Pablos-Sarabia, David Griol, Zoraida Callejas
Summary: This paper presents a proposal to enhance mental health organizations' data collection and decision processes through a web application, enabling workers to efficiently store, access, and modify client data. The computerized model is leveraged by an intelligent conversational assistant to provide round-the-clock personalized and ubiquitous interaction with clients, promoting engagement with the organization.
PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON UBIQUITOUS COMPUTING & AMBIENT INTELLIGENCE (UCAMI 2022)
(2023)
Proceedings Paper
Automation & Control Systems
David Griol, Zoraida Callejas, Fernando Fernandez-Martinez, Anna Esposito
Summary: This paper describes a conversational system aimed at promoting healthy lifestyle habits, which provides automation of management tasks, improves user experience, and increases efficiency. The system utilizes Google's Dialogflow platform and various APIs and data repositories in the cloud, and has been integrated into the Facebook Messenger instant messaging platform.
2022 IEEE INTL CONF ON DEPENDABLE, AUTONOMIC AND SECURE COMPUTING, INTL CONF ON PERVASIVE INTELLIGENCE AND COMPUTING, INTL CONF ON CLOUD AND BIG DATA COMPUTING, INTL CONF ON CYBER SCIENCE AND TECHNOLOGY CONGRESS (DASC/PICOM/CBDCOM/CYBERSCITECH)
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Pablo Canas, David Griol, Zoraida Callejas
Summary: This paper introduces a statistical-based dialog manager architecture for designing conversational interfaces, which has been evaluated in a real use case for a train scheduling domain, showing high value in user experience.
HYBRID ARTIFICIAL INTELLIGENT SYSTEMS, HAIS 2022
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
David Griol, Zoraida Callejas
Summary: This paper compares different statistical techniques for training dialogue managers and finds that using generative models and an intent classifier with Neural Networks and Seq2Seq using GRU cells achieve the best accuracy for dialogue management.
16TH INTERNATIONAL CONFERENCE ON SOFT COMPUTING MODELS IN INDUSTRIAL AND ENVIRONMENTAL APPLICATIONS (SOCO 2021)
(2022)
Article
Computer Science, Information Systems
Jhonn Pablo Rodriguez, David Camilo Corrales, David Griol, Zoraida Callejas, Juan Carlos Corrales
Summary: Coffee plays a vital role in rural employment in Colombia, with more than 785,000 workers directly employed in this activity. Colombian coffee growers face various unpredictable factors and have proposed a time series model based on weather and crop management information to improve their agricultural management.
CMC-COMPUTERS MATERIALS & CONTINUA
(2022)
Article
Computer Science, Information Systems
Matthias Kraus, Nicolas Wagner, Zoraida Callejas, Wolfgang Minker
Summary: This article investigates the development and implementation of proactive dialogues for fostering a trustworthy human-computer relationship and providing adequate and timely assistance. The study provides a formalization of proactive dialogue in conversational assistants, identifying its effects on the perceived trustworthiness of a system and user experience. The results contribute significant insights into the human-computer trust relationship and dependencies between proactive dialogue and user specific characteristics.
Review
Computer Science, Information Systems
Claudia Greco, Olimpia Matarazzo, Gennaro Cordasco, Alessandro Vinciarelli, Zoraida Callejas, Anna Esposito
Summary: Existing studies have shown that EEG-based biomarkers are effective in discriminating between depressed and healthy subjects. However, further research is needed to investigate the causal link between EEG measures and depressive subtypes, and to address methodological issues in order to enhance future research in this field.