User Model Research Articles

Enhancing ID-based Recommendation with Large Language Models

Large Language Models (LLMs) have recently garnered significant attention in various domains, including recommendation systems. Recent research leverages the capabilities of LLMs to improve the performance and user modeling aspects of recommender systems. These studies primarily focus on utilizing LLMs to interpret textual data in recommendation tasks. However, it's worth noting that in ID-based recommendations, textual data is absent, and only ID data is available. The untapped potential of LLMs for ID data within the ID-based recommendation paradigm remains relatively unexplored. To this end, we introduce a pioneering approach called “LLM for ID-based Recommendation” (LLM4IDRec). This innovative approach integrates the capabilities of LLMs while exclusively relying on ID data, thus diverging from the previous reliance on textual data. The basic idea of LLM4IDRec is that by employing LLM to augment ID data, if augmented ID data can improve recommendation performance, it demonstrates the ability of LLM to interpret ID data effectively, exploring an innovative way for the integration of LLM in ID-based recommendation. Specifically, we first define a prompt template to enhance LLM's ability to comprehend ID data and the ID-based recommendation task. Next, during the process of generating training data using this prompt template, we develop two efficient methods to capture both the local and global structure of ID data. We feed this generated training data into the LLM and employ LoRA for fine-tuning LLM. Following the fine-tuning phase, we utilize the fine-tuned LLM to generate ID data that aligns with users’ preferences. We design two filtering strategies to eliminate invalid generated data. Thirdly, we can merge the original ID data with the generated ID data, creating augmented data. Finally, we input this augmented data into the existing ID-based recommendation models without any modifications to the recommendation model itself. We evaluate the effectiveness of our LLM4IDRec approach using three widely-used datasets. Our results demonstrate a notable improvement in recommendation performance, with our approach consistently outperforming existing methods in ID-based recommendation by solely augmenting input data.

A Discourse on the Potency of Social Media in Promoting Local Talents in Nigeria

This paper Titled: Discourse on the Potency of Social Media in Promoting Local Talents in Nigeria stems from the impact of social media as a tool for local talents hunt in Nigeria. Nigerians all over the world are highly gifted and talented people; this is seen in their disposition and overwhelming accomplishment over the past decades round the world. Yet, the Nigerian society has not provided an enabling environment to hone home grown and plethora of gifted entrepreneurs, singers, innovators and models. The above situation stimulated the researcher’s interest to inundate readers and Nigerians on the efficacy of the platform of the social media and the advent of other new media as scout platform for local talent in Nigeria. To thoroughly explore the subject matter, the researcher adopted the qualitative research method using the documentary review design. The study adopted Technological Determinism Theory to explicate how the advent of the new media has changed the model of media usage among people. Findings from the study unraveled that the advent of social media and other new media platforms catalyzed the recognition and honing of local talents in Nigeria. Furthermore, it was discovered that in Nigeria, new media (social media) has helped in plethora of ways by bringing to the spotlight some gifted local talents in Nigeria. The study however found that, media illiteracy is a bane to social media usage in Nigeria. Therefore, the study suggests that, if the medium of social media is properly utilized by Nigerians can chart a way forward for new innovations and in local talent discovery in Nigeria.

"You Can either Blame Technology or Blame a Person..." --- A Conceptual Model of Users' AI-Risk Perception as a Tool for HCI

AI-powered systems pose unknown challenges for designers, policymakers, and users, making it more difficult to assess potential harms and outcomes. Although understanding risks is a requirement for building trust in technology, users are often excluded from risk assessments and explanations in policy and design. To address this issue, we conducted three workshops with 18 participants and discussed the EU AI Act, which is the European proposal for a legal framework for AI regulation. Based on results of these workshops, we propose a user-centered conceptual model with five risk dimensions (Design and Development, Operational, Distributive, Individual, and Societal) that includes 17 key risks. We further identify six criteria for categorizing use cases. Our conceptual model (1) contributes to responsible design discourses by connecting the risk assessment theories with user-centered approaches, and (2) supports designers and policymakers in more strongly considering a user perspective that complements their own expert views.

Predicting Electric Vehicle Charging Demand in Residential Areas Using POI Data and Decision‐Making Model

Transport electrification is a crucial element of the ongoing energy transition, essential for achieving carbon peaking and carbon neutrality goals. The proliferation of electric vehicles (EVs) introduces significant challenges to power distribution network stability due to their aggregated charging load in residential areas, particularly during peak electricity consumption periods. This paper proposes a method to predict the spatiotemporal distribution of EV charging demand in residential areas using geographic information points of interest (POI) data features and a decision‐making model. Utilizing real historical data, probability distribution models for EV users' arrival times and charging characteristics were constructed using Gaussian Mixture Models (GMM). The spatiotemporal characteristics of EV travel and charging behaviors were analyzed, and a comprehensive charging decision model incorporating both emergency and stochastic scenarios was developed. The model's efficacy in capturing the probability distributions of characteristic variables was validated through a case study. The results demonstrate the model's potential for accurately predicting EV charging demand, providing valuable insights for infrastructure planning and resource allocation. © 2024 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.

Enhancing and Explaining Energy Management in Smart Manufacturing Production via ML Within the Context of Industry 4.0

Objectives: This study aims to explore the integration of big data and machine learning to enhance energy efficiency in manufacturing, focusing on the opportunities presented by Industry 4.0 and cyber-physical production systems. Theoretical Framework: The research leverages supervised learning techniques to analyze and predict machine-specific energy consumption patterns, utilizing energy disaggregation as a foundational concept. Method: Various machine learning algorithms, including Lasso regression, Linear regression, and Decision Tree, will be employed to develop predictive models for energy usage. Additionally, Explainable Machine Learning (XML) techniques will be utilized to ensure interpretability and clarity in the prediction outcomes. Results and Discussion: The proposed framework aims to elucidate the relationship between equipment utility and energy consumption, providing comprehensible explanations that enhance decision-making processes within intelligent industrial environments. Research Implications: This work highlights the significant potential of XML in transforming machine learning applications in manufacturing, paving the way for improved energy efficiency and operational effectiveness. Originality/Value: The introduction of an innovative framework that combines machine learning with explainability in the context of energy consumption marks a valuable contribution to the fields of manufacturing and sustainable industry practices.

The Short Video Platform Recommendation Mechanism Based on the Improved Neural Network Algorithm to the Mainstream Media

Currently, the capacity of short videos continues to increase. Video manufacturers hope to enhance user experience and stickiness through recommendation algorithms, while users seek personalized videos to save time and money. Therefore, in order to address the data sparsity and high-dimensional feature extraction, this study proposes a novel short video platform recommendation model. The proposed method utilizes the term frequency inverse document frequency algorithm for text mining, and combines error back propagation neural network for learning to explore the potential connection between users and videos. This method combines natural language processing and image analysis in deep learning to construct accurate user and video models, deeply explore user interests, and improve the accuracy and effectiveness of recommendation systems for user preferences. The research results showed that the recommendation accuracy of this method was 66% and 70% respectively, and the prediction accuracy was 73.50% and 88% respectively. When Num=128, 200 data points were recommended within 0.3678s. The proposed algorithm outperforms the other three methods in terms of recommendation accuracy compared with the ItemCF and UserCF algorithms. This is because the method uses an approach based on image and user vector, combined with relevant features of the video and user. The proposed method can deeply explore the relevant features of videos and users, overcoming the data scarcity in previous collaborative screening, and guiding video recommendation on practical media platforms.

Integrated Analysis of Utilization Efficiency of Intelligent Sports Venues Based on Data Mining Algorithms

With the rapid development of information technology and the increasing prosperity of the sports industry, the construction and management of intelligent sports venues have become an important way to improve the efficiency of sports resource utilization and meet diverse fitness needs. This paper focuses on the integrated analysis of the utilization efficiency of intelligent sports venues based on data mining algorithms. Aiming to reveal the inherent laws and potential value in the operation of sports venues through advanced data processing and analysis techniques, and provide a scientific basis for optimizing resource allocation and improving service quality. Intelligent sports venues utilize modern information technologies such as the Internet of Things, big data, and cloud computing to achieve comprehensive and intelligent monitoring and management of venue facilities, personnel activities, environmental conditions, and more. This new management model not only improves the operational efficiency of the venue but also enhances the user experience and meets personalized and diverse fitness needs. This paper focuses on the current situation of the opening of university sports venues to the outside world, aiming to reveal the pain points and bottlenecks in their operation process through detailed investigation and analysis. This study introduces advanced concepts of middleware architecture and constructs a multi-level data integration framework that includes local models, global models, user models, and mapping rules. This framework can effectively integrate distributed heterogeneous data sources, achieving seamless integration and efficient interaction of data. At the same time, with the help of distributed programming technology in cloud computing, it is possible to conduct in-depth mining and analysis of integrated massive data, discover potential value, and provide a scientific basis for decision support.

PERCEPTIONS OF THE APPLICATION OF THE CARE MODEL FOR CHRONIC CONDITIONS IN PARANÁ, BRAZIL

The current healthcare model for users with chronic conditions in the Unified Health System needs to be transformed. In Paraná, there are strategies for continuing education and co-financing of services focused on the Chronic Conditions Care Model. This is a descriptive and exploratory study with a qualitative approach that analyzed the perceptions of Primary Health Care and Specialized Outpatient Care professionals about the implementation of the Chronic Conditions Care Model. Four focus groups were held, with the participation of 37 professionals, and the discussions permeated the work processes from the perspective of thematic analysis. The statements identified elements such as risk stratification of the health condition, organization of care, and team rapprochement. The development of the care plan is incipient due to the lack of understanding of professionals and users about the tool. In health units, access occurs according to the logic of acute care, and care sharing through regulatory centers. Continuing education strategies, structured public policies and specific financing are drivers for the consolidation of the Model.

Evaluating functional ability in older adults’ object retrieval behavior from kitchen furniture using OpenPose and REBA

The purpose of this study is to evaluate, through analysis, the ability of older persons to retrieve items from kitchen cabinets. To achieve this goal, data were collected from 128 valid questionnaires and supplemented with field research and user interviews. The study revealed that the elderly’s behavior in retrieving items from kitchen spaces is characterized by both high frequency and difficulty. For this experiment, a total of 42 participants, comprising 21 males and 21 females from the self-care elderly population in the Yangtze River Delta region, were recruited. Two different experimental settings were arranged: one with kitchen utensils arranged in a straight line and another with a purpose-made chest of drawers with varying heights. Video recordings using the Logitech C930C were utilized to capture the gestures and behaviors of the elderly while retrieving objects from the kitchen cupboards (cabinets). By employing a combination of the OpenPose model and the Rapid Entire Body Assessment (REBA) method, which involves calculating human posture angles, REBA scores, and determining the risk level of Work-Related Musculoskeletal Disorders (WMSDs), a risk assessment framework for manual operations associated with WMSDs was developed. Using the angle data acquired from the user operation experiment as parameters, a gradient model of the elderly user’s operational capability was established. The findings indicated a significant impact of neck, trunk, and knee movements on the subjects (P < 0.001). The participants were able to distinguish between different levels of exertion, categorizing movements as ‘easy’, ‘moderate,’ or ‘strenuous.’ These results form the basis for a comfort gradient model for leaning over and retrieving items. Given the prevalent conditions of bone and joint degeneration and osteoporosis among the elderly population, it is evident that they face challenges when accessing items in the kitchen. Therefore, investigating the elderly’s execution abilities during the retrieval process becomes crucial. Understanding how different cabinet heights impact the joint angles of the elderly can be instrumental in optimizing cabinet designs for elderly users, thereby reducing their physical exertion in the kitchen and enhancing their comfort levels. This research holds significant value in improving the quality of life for the elderly population at home and fostering the advancement of elderly-friendly design principles.

Mitigating power grid impact from proactive data center workload shifts: A coordinated scheduling strategy integrating synergistic traffic - data - power networks

The widespread adoption of cloud computing has markedly escalated the demand for Internet services, rendering Internet Data Centers (IDCs) a critical component in the consideration for demand response (DR) initiatives and grid energy management. This work proposes a flexible scheduling strategy integrating data, traffic, and power networks (DN-TN-PN) to mitigate the impact of proactive IDC workload transfers on power system balance. The spatial-temporal transfer characteristics of EV charging and discharging loads are utilized to track the proactive spatial-temporal transfer of IDC workloads. Considering the influence of weather factors and traffic congestion on EV travel time, an incentive pricing model for EV users is introduced to alleviate bus load fluctuation and line congestion in the distribution network. Numerical simulations demonstrate that the proposed EV pricing strategy, which accounts for the coupled DN-TN-PN, reduces IDC-induced load fluctuations by over 20 % and decreases transmission line occupancy by approximately 15 %, opening up more opportunities for the implementation of DR programs.

Consumers’ Continuance Intention Toward Online Shopping Apps: An Integrated Model

ABSTRACT Understanding the continuance intention to use online shopping apps is crucial, as it impacts both the continuance adoption and ongoing motivation for shopping. This study, based on the Theory of Planned Behavior and the Technology Acceptance Model, employed SEM to analyze responses from 343 participants. It aimed to construct and compare a model of users’ continuance intentions toward online shopping apps. The findings reveal that continuance intentions are positively driven by the marketing mix, usage motivation, perceived behavioral control, and attitude, and negatively by perceived risk. Interestingly, perceived behavioral control does not significantly affect continuance intentions. Attitudes are boosted by usage motivation, perceived usefulness, and ease of use, whereas perceived behavioral control is shaped by motivation and negatively by risk. Both ease of use and usefulness benefit from marketing and motivation but are hindered by risk. Marketing mix positively impacts motivation, underscoring the interplay of factors in adopting the apps.

Mobility Prediction via Rule-Enhanced Knowledge Graph

With the rapid development of location acquisition technologies, massive mobile trajectories have been collected and made available to us, which support a fantastic way of understanding and modeling individuals’ mobility. However, existing data-driven methods either fail to capture the long-range dependency or suffer from a high computational cost. To overcome these issues, we propose a knowledge-driven framework for mobility prediction, which leverages knowledge graphs (KG) to formulate the mobility prediction task into the KG completion problem through integrating the structured “knowledge” from the mobility data. However, most related mobility prediction works only focus on the structured information encoded in existing triples, which ignores the rich semantic information of relation paths composed of multiple relation triples. In this article, we apply a dedicated module to extract the supplementary semantic structure of paths in KG, which contributes to the interpretability and accuracy of our model. Specifically, the extracted rules are applied to capture the dependencies between relational facts. Moreover, by incorporating user information in the entity-relation space with the corresponding hyperplane, our method could capture diverse user mobility patterns and model the personal characteristics of users to improve the accuracy of mobility prediction. Extensive evaluations illustrate that our proposed model beats state-of-the-art mobility prediction algorithms, which verifies the superiority of utilizing logical rules and user hyperplanes. Our implementation code is available at https://github.com/tsinghua-fib-lab/RulekG-MobiPre.git

Potential assessment of coordinated regulation of power load of emerging industrial users based on extreme scenarios of electric vehicle aggregators

AbstractWith the advancement of industrial low carbonization and electrification, emerging industrial production technologies have the characteristics of high‐power consumption and significant impact load. In the context of increasing global climate change, frequent extreme weather events have brought serious challenges to the balance of power and electricity in the power system, and have a significant impact on the production scheduling of industrial users, especially industrial users using electrified production technology. First, based on the representative high‐power industrial users of hydrogen reduction steel plants and internet datacentres (IDC), this paper establishes a flexible resource scheduling optimization model for single industrial users, and considers the collaborative relationship between multi‐user flexible resources to establish a collaborative scheduling optimization model for industrial users. Then, considering the charging and discharging characteristics of electric vehicles (EV) in extreme scenarios, the redundant capacity of EVs is aggregated by EV aggregators and sold to industrial users, and a collaborative scheduling optimization model of EV aggregators and industrial users is established. Finally, the effectiveness of the proposed model and algorithm is verified by simulation analysis. Compared with the traditional industrial user production optimization, the proposed model can tap the potential of multi‐agent scheduling operation in extreme scenarios.

Unraveling the dynamics of ChatGPT adoption and utilization through Structural Equation Modeling

ChatGPT, an advanced Artificial Intelligence tool, is getting considerable attention in higher education. ChatGPT significantly changes the student learning experience through its AI-aided support, personalized study assistance and effective educational experiences, and it has become an object of particular interest in this context. This research aimed to build a technology acceptance and usage model that encapsulates the elements influencing students’ adoption and utilization of ChatGPT, drawing on constructs from the ‘Unified Theory of Acceptance and Use of Technology’ and ‘Flow Theory’. The proposed model was found valid and prolific, with the credibility of the results relying on the self-reported surveys of 505 students from three universities in Pakistan. Structural Equation Modelling (SEM) was used to analyze data that confirmed the robustness and validity of the proposed model of the study. The study findings supported nine out of the ten proposed hypotheses. Perceived playfulness was declared the paramount predictor of behavioral intention, while perceived values and performance expectancy were the next-level predictors. Additionally, behavioral attention was a high and inspiring determinant of ChatGPT usage behavior, followed by attention focus. This analysis demonstrates a need for a thorough investigation of AI tools like ChatGPT in higher education.

IPSRM: An intent perceived sequential recommendation model

Objectives:Sequential recommendation aims to recommend items that are relevant to users’ interests based on their existing interaction sequences. Current models lack in capturing users’ latent intentions and do not sufficiently consider sequence information during the modeling of users and items. Additionally, noise in user interaction sequences can affect the model’s optimization process. Methods:This paper introduces an intent perceived sequential recommendation model (IPSRM). IPSRM employs the generalized expectation–maximization (EM) framework, alternating between learning sequence representations and optimizing the model to better capture the underlying intentions of user interactions. Specifically, IPSRM maps unlabeled behavioral sequences into frequency domain filtering and random Gaussian distribution space. These mappings reduce the impact of noise and improve the learning of user behavior representations. Through clustering process, IPSRM captures users’ potential interaction intentions and incorporates them as one of the supervisions into the contrastive self-supervised learning process to guide the optimization process. Results:Experimental results on four standard datasets demonstrate the superiority of IPSRM. Comparative experiments also verify that IPSRM exhibits strong robustness under cold start and noisy interaction conditions. Conclusions:Capturing latent user intentions, integrating intention-based supervision into model optimization, and mitigating noise in sequential modeling significantly enhance the performance of sequential recommendation systems.

Exploring Cross-site User Modeling Without Cross-site User Identity Linkage: A Case Study of Content Preference Prediction

Performing user modeling on two or more social media platforms collaboratively and complementing each other (cross-site user modeling) has been a significant problem in the area of social media mining in recent years. The core of this problem is to get to know a person’s identities on multiple platforms and then train user models collaboratively among these platforms. However, for privacy protection, many people do not want their identities on different platforms to be linked and disclosed. For this problem, we set cross-site Content Preference Prediction as a task and propose a cross-site user modeling method without cross-site User Identity Linkage (UIL). The core thought borrowed from privacy-preserving recommender system research is to organize social media identities into groups to hide the identity linkage among platforms. Experiments on real-world datasets suggest that our method outperforms the existing cross-site user modeling methods with cross-site UIL regarding several metrics.

An Infrastructure’s emissions assessment tool: AMICO - Account Method of Infrastructures embodied CarbOn

Abstract In 2013, the UK Infrastructure Carbon Review report highlighted how the infrastructure sector contributed to the national total emissions and, therefore, had a key role in contributing to the national reduction, adopting targeted strategies to reduce them. The main objective of AMICO’s research project: “Account Method of Infrastructures embodied CarbOn”, is the development of a tool for the analysis and parametric evaluation of carbon emissions (Embodied Carbon, EC) and energy resource requirements (Embodied Energy, EE) related to the design and construction of infrastructures (railway tunnels, bridges, highways …) AMICO tool is designed to be used for tendering, contract start-up and construction of infrastructure by Webuild Group (Italian Joint Stock Company) as part of a research project funded by MIMIT - Ministry of Business and Made in Italy, with the scientific support of the Politecnico di Torino (Department of Architecture and Design). AMICO aims to assess potential impacts in terms of greenhouse gas emissions (GHG) and potential primary energy resource needs over the infrastructure life cycle and at different design process stages, allowing designers and constructors to evaluate decarbonisation in their choices. AMICO is currently organised into two calculation models: 1. An Excel spreadsheet file, divided into worksheets containing algorithms for accounting of EC and EE, organised into two usage models: Basic, which will be used for tenders and the early design stage, and Advanced, which will be used in the detailed design or construction stage. 2. A code for the BIM environment that accounts for impacts (EE and EC) related to infrastructure systems (as defined by UNICLASS) combined and concurrently with design in the BIM environment (this section is not the focus of this paper). An innovation aspect is sharing a database of environmental data about materials and processes (transportation, construction, demolition) directly related to the Bill of Quantity without using Life Cycle Assessment software. The database will be integrated into the Datalake to speed up and optimise infrastructure design and construction work. You can set the database to the location of the work. Through data digitisation, AMICO is designed to be as automated, easy to fill in, and intuitive as possible.

A structural model of pre-service teachers’ attitude, acceptance, and continuance intention towards mobile augmented reality

The increasing prevalence of mobile augmented reality (MAR) has garnered attention in contemporary times. However, its application within educational settings remains largely untapped, primarily due to its predominant utilization in entertainment domains. MAR presents opportunities for augmenting learning experiences, fostering enhanced adaptability in educational methodologies, and enriching students’ academic endeavors. Therefore, this study aims to enhance the use of MAR and formulates a model of MAR usage for pre-service teachers. In the context of this study, MAR learning cardiovascular (MARLCardio) serves as an experimental tool for exemplifying the features of MAR. This study integrates the Unified Theory of Acceptance and Use of Technology (UTAUT), Technology Acceptance Model (TAM), and Expectation Confirmation Model (ECM) as underpinning theories. This study was undertaken to determine factors that influence pre-service teachers’ attitude, acceptance, and their intention to continue using MAR in the learning process. Responses from 456 pre-service teachers through a survey were analyzed using structural equation modelling. Findings in this study found that pre-service teachers agree that facilitating condition play a crucial role in shaping pre-service teachers’ positive attitude and influencing their intention to continue using MAR. Meanwhile, performance expectancy plays a major role in ensuring pre-service teachers’ acceptance of using MAR in their learning process. It was anticipated that the formulated MAR structural model would offer insights and support forthcoming research endeavors aimed at fostering the sustainability of MAR utilization.

Confidential outsourced support vector machine learning based on well-separated structure

Support Vector Machine (SVM) has revolutionized various domains and achieved remarkable successes. This progress relies on subtle algorithms and more on large training samples. However, the massive data collection introduces security concerns. To facilitate secure integration of data efficiently for building an accurate SVM classifier, we present a non-interactive protocol for privacy-preserving SVM, named NPSVMT. Specifically, we define a new well-separated structure for computing gradients that can decouple the fusion matter between user data and model parameters, allowing data providers to outsource the collaborative learning task to the cloud. As a result, NPSVMT is capable of removing the multiple communications and eliminating the straggler’s effect (waiting for the last), thereby going beyond those developed with interactive methods, e.g., federated learning. To further decrease the data traffic, we introduce a high-efficient coding method to compress and parse training data. In addition, unlike outsourced schemes based on homomorphic encryption or secret sharing, NPSVMT exploits functional encryption to maintain the data confidentiality, achieving dropout-tolerant secure aggregation. The implementations verify that NPSVMT is faster by orders of magnitude than the existing privacy-preserving SVM schemes on benchmark datasets.

Exploring the Opinion Shifting of Social Media Users Based on Stance Detection

ABSTRACTFine‐grained mining of stance change among social media users during crises contributes significantly to a comprehensive understanding of the development of public opinion online. This study collected detailed information on 227,281 Twitter users who continuously participated in the discussion about the COVID‐19 vaccine to analyze patterns of opinion shifting. Utilizing the COVID‐Twitter‐BERT pre‐trained model, we detect each user's stance on the COVID‐19 vaccine and construct a time‐series dataset reflecting their stance over time. A prediction model for users' opinion shifting is established based on the LightGBM model, and the SHAP explanation method is used to rank the importance of features to identify key features influencing the opinion shifting prediction. Our research findings indicate that nearly half of the users maintained a consistent stance on vaccines, with a relatively low proportion of shifts between pro‐vaccine and anti‐vaccine stances. The F1 value of the prediction model based on the LightGBM model is 0.92. These findings can assist in monitoring social media user opinions online.