User Interaction Information Research Articles

Clinical, Research, and Educational Applications of ChatGPT in Dentistry: A Narrative Review

Artificial intelligence (AI), specifically Generative Pre-trained Transformer (GPT) technology, has revolutionized various fields, including medicine and dentistry. The AI model ChatGPT, developed by OpenAI, mimics human language on a large scale, generating coherent and contextually appropriate responses. ChatGPT serves as an auxiliary resource for diagnosis and decision-making across various medical disciplines. This comprehensive narrative review aims to explore how ChatGPT can assist the dental sector, highlighting its potential to enhance various aspects of the discipline. This review includes a literature search on the application of ChatGPT in dentistry, with a focus on the differences between the free version, ChatGPT 3.5, and the more advanced subscription-based version, ChatGPT 4. Specifically, ChatGPT has proven to be effective in enhancing user interaction, providing fast and accurate information and improving the accessibility of knowledge. However, despite these advantages, several limitations are identified, including concerns regarding the accuracy of responses in complex scenarios, ethical considerations surrounding its use, and the need for improved training to handle highly specialized queries. In conclusion, while ChatGPT offers numerous benefits in terms of efficiency and scalability, further research and development are needed to address these limitations, particularly in areas requiring greater precision, ethical oversight, and specialized expertise.

EnNet: Enhanced Interactive Information Network with Zero-Order Optimization.

Interactive image segmentation extremely accelerates the generation of high-quality annotation image datasets, which are the pillars of the applications of deep learning. However, these methods suffer from the insignificance of interaction information and excessively high optimization costs, resulting in unexpected segmentation outcomes and increased computational burden. To address these issues, this paper focuses on interactive information mining from the network architecture and optimization procedure. In terms of network architecture, the issue mentioned above arises from two perspectives: the less representative feature of interactive regions in each layer and the interactive information weakened by the network hierarchy structure. Therefore, the paper proposes a network called EnNet. The network addresses the two aforementioned issues by employing attention mechanisms to integrate user interaction information across the entire image and incorporating interaction information twice in a design that progresses from coarse to fine. In terms of optimization, this paper proposes a method of using zero-order optimization during the first four iterations of training. This approach can reduce computational overhead with only a minimal reduction in accuracy. The experimental results on GrabCut, Berkeley, DAVIS, and SBD datasets validate the effectiveness of the proposed method, with our approach achieving an average NOC@90 that surpasses RITM by 0.35.

KGCF: Social relationship-aware graph collaborative filtering for recommendation

With the increasing popularity of recommendation techniques and social networks, social network recommendation has become a significant research field, i.e., predicting a user's preferences based on her or his historical interaction data, because the social relationships of users can not only enrich their interaction information, but also imply the accurate characterization of their preferences. Although there are some existing studies that predict users' preferences by using the characteristics of social relationships, their time complexity and hardware resource consumption are often very expensive, which limits the feasibility in large-scale real-world scenarios. To address these issues, in this paper, we propose a simple and effective K-core Graph Collaborative Filtering (KGCF) model to incorporate the user's social features into a recommendation framework. The user's interaction information is auto-encoded linearly by incorporating the social relationship graph into the user's interaction information and constructing a multi-layer graph filter. The linear autoencoder graph filter alleviates the highly sparse data problem and dramatically reduces the training time and hardware resource consumption. Experimental results on several real-world datasets show the superior effectiveness and spatiotemporal efficiency compared with the existing baselines, especially the training speed is significantly accelerated for large-scale datasets.

The 4Fs model of digital marketing: Its impact on brand positioning

Introduction/Objective: The 4Fs of digital marketing — flow, functionality, feedback, and loyalty — enhance effective and appealing communication across digital platforms. This study aims to evaluate the application of the 4Fs model in a Mexico-based ecotourism hotel, validating its measurement framework and analyzing the relationship between these variables and brand positioning. Methodology: A quantitative, non-experimental, cross-sectional, descriptive, and correlational approach was employed. The study surveyed 280 followers of the company’s Facebook page, using a 21-item questionnaire. Confirmatory factor analysis with maximum likelihood estimation assessed construct validity, while Cronbach’s alpha coefficient measured reliability. Results: Pearson correlation analysis revealed a strong, significant relationship between the 4Fs of digital marketing and brand positioning: flow (0.808), functionality (0.829), feedback (0.885), and loyalty (0.864). All correlations were statistically significant at the 0.01 level (two-tailed). Conclusions: The study affirmed that the company’s efforts on its websites to foster pleasant user interaction and effective information retrieval significantly contribute to creating memorable brand associations.

A Large-Scale Reviews-Driven Multi-Criteria Product Ranking Approach Based on User Credibility and Division Mechanism

Massive online reviews provide consumers with the convenience of obtaining product information, but it is still worth exploring how to provide consumers with useful and reliable product rankings. The existing ranking methods do not fully mine user information, rating, and text comment information to obtain scientific and reasonable information aggregation methods. Therefore, this study constructs a user credibility model and proposes a large-scale user information aggregation method to obtain a new product ranking method. First, in order to obtain the aggregate weight of large-scale users, this paper proposes a consistency modeling method of text comments and star ratings by mining the associated information of user comments, including user interaction information and user personalized characteristics information, combined with sentiment analysis technology, and then constructs a user credibility model. Second, a double-layer group division mechanism considering user regions and comment time is designed to develop the large-scale group ratings aggregation approach. Third, based on the user credibility model and the large-scale ratings aggregation approach, a product ranking method is developed. Finally, the feasibility and effectiveness of the proposed method are verified through a case study for automobile ranking and a comparative analysis is furnished. The analysis results of the application case of automobile ranking show that there is a significant difference between the ranking results obtained by the ratings aggregation method based on the arithmetic mean and the ranking results obtained by this method. The method in this study comprehensively considers user credibility and group division, which can be reflected in user aggregation weights and the group aggregation process, and can also obtain more scientific and reasonable decision results.

Hybrid Graph Neural Network Recommendation Based on Multi-Behavior Interaction and Time Sequence Awareness

In recent years, mining user multi-behavior information for prediction has become a hot topic in recommendation systems. Usually, researchers only use graph networks to capture the relationship between multiple types of user-interaction information and target items, while ignoring the order of interactions. This makes multi-behavior information underutilized. In response to the above problem, we propose a new hybrid graph network recommendation model called the User Multi-Behavior Graph Network (UMBGN). The model uses a joint learning mechanism to integrate user–item multi-behavior interaction sequences. We designed a user multi-behavior information-aware layer to focus on the long-term multi-behavior features of users and learn temporally ordered user–item interaction information through BiGRU units and AUGRU units. Furthermore, we also defined the propagation weights between the user–item interaction graph and the item–item relationship graph according to user behavior preferences to capture more valuable dependencies. Extensive experiments on three public datasets, namely MovieLens, Yelp2018, and Online Mall, show that our model outperforms the best baselines by 2.04%, 3.82%, and 3.23%.

Predicting Violence-Induced Stress in an Arabic Social Media Forum

Social Media such as Facebook plays a substantial role in virtual communities by sharing ideas and ideologies among different populations over time. Social interaction analysis aids in defining people’s emotions and aids in assessing public attitudes, towards different issues such as violence against women and children. In this paper, we proposed an Arabic language prediction model to identify the issue of Violence-Induced Stress in social media. We searched for Arabic posts of many countries through Facebook application programming interface (API). We discovered that the stress state of a battered woman is usually related to her friend’s stress states on Facebook. We applied a large real database from Facebook platforms to analytically investigate the correlation of violence-induced stress states and the victim interactions on social media. We extracted a set of textual, spatial, and interaction attributes from various features. Therefore, we are proposing a hybrid model–an interaction graph model incorporated in a deep learning neural model to leverage post content and interaction data for violence-induced stress detection. Experiments depict that our proposed hybrid model can enhance the prediction performance by 10% in F1-measure. Also, considering the user interaction information can learn an interesting phenomenon, where, the sparse social interactions of violence-induced stress stressed victims is higher by around 15% percent non-battered users, signifying that the structure of the friends of such victims is less connected than non-stressed users.

MAMOM: Multicriteria Attribution Model for Online Marketing

One of the problems facing online marketing in omni-channel environments is efficient budget allocation to each channel. The complexity of this problem is greater in omni-channel environments due to the greater number of channels and the need to analyze data sources containing user interaction information. To solve this problem, different attribution models have been proposed to assign the weight that each channel has in the acquisition of a product, also known as conversion. Each of these attribution models adopts a strategy to define the weighting of channels. The decision-making strategy is established using the company’s expert knowledge, which can contain different criteria depending on the department to which the expert belongs. The aim of this research is to present a new multicriteria attribution model for online marketing (MAMOM) based on Analytical Hierarchical Process, that resolves this type of problem. MAMOM is a meta-model that takes as input information related to channel features, user interactions and even decisions from other strategies. This information is integrated to enable the integration of interdepartmental strategies as well as to obtain a dynamic attribution model based on expert interviews. The expert assessment procedure is carried out subjectively and simply with a pairwise assessment of the criteria, finally MAMOM obtains a final formulation to calculate the investment to be made in each channel if the experts’ opinions are considered. Results show that first five channels selected by the MAMOM are nearly identical to those that would be obtained with the traditional models but with a different sequence caused by experts’ knowledge. This result shows the capacity of MAMOM to factor in expert opinion and experience for making the investment outcome more aligned with the tactical and strategic objectives of a given online marketing campaign.

Research on the Spectral Domain Graph Convolution Collaborative Filtering Algorithm Based on Reinforcement Learning and Chebyshev

To solve the abovementioned problem, we propose a collaborative filtering recommendation algorithm that incorporates singular value decomposition (SVD) and Chebyshev truncation in spectral domain convolution. Firstly, the SVD algorithm is used to optimize the adjacency matrix, mine the potential association information between users and items, and expand the user-item adjacency matrix. Finally, based on the MovieLens-1M public dataset, the proposed algorithm (CBSVD-SCF) is compared with other commonly used algorithms. The results show that the article optimizes the recommendation effect of the algorithm based on the traditional collaborative filtering algorithm by combining the temporal order and sequence of user interaction information, as well as the popularity of items and the activity of users; the experimental results on MovieLens show that the optimized collaborative filtering recommendation algorithm can effectively improve the recommendation effect.

Categorizing Touch-Input Locations from Touchscreen Device Interfaces via On-Board Mechano-Acoustic Transducers

Many mobile electronics devices, including smartphones and tablets, require the user to interact physically with the device via tapping the touchscreen. Conveniently, these compact devices are also equipped with high-precision transducers such as accelerometers and microphones, integrated mechanically and designed on-board to support a range of user functionalities. However, unintended access to these transducer signals (bypassing normal on-board data access controls) may allow sensitive user interaction information to be detected and thereby exploited. In this study, we show that acoustic features extracted from the on-board microphone signals, supported with accelerometer and gyroscope signals, may be used together with machine learning techniques to successfully determine the user’s touch input location on a touchscreen: our ensemble model, namely the random forest model, predicts touch input location with up to 86% accuracy in a realistic scenario. Accordingly, we present the approach and techniques used, the performance of the model developed, and also discuss limitations and possible mitigation methods to thwart possible exploitation of such unintended signal channels.

GIF Thumbnails: Attract More Clicks to Your Videos

With the rapid increase of mobile devices and online media, more and more people prefer posting/viewing videos online. Generally, these videos are presented on video streaming sites with image thumbnails and text titles. While facing huge amounts of videos, a viewer clicks through a certain video with high probability because of its eye-catching thumbnail. However, current video thumbnails are created manually, which is time-consuming and quality-unguaranteed. And static image thumbnails contain very limited information of the corresponding videos, which prevents users from successfully clicking what they really want to view. In this paper, we address a novel problem, namely GIF thumbnail generation, which aims to automatically generate GIF thumbnails for videos and consequently boost their Click-Through-Rate (CTR). Here, a GIF thumbnail is an animated GIF file consisting of multiple segments from the video, containing more information of the target video than a static image thumbnail. To support this study, we build the first GIF thumbnails benchmark dataset that consists of 1070 videos covering a total duration of 69.1 hours, and 5394 corresponding manually-annotated GIFs. To solve this problem, we propose a learning-based automatic GIF thumbnail generation model, which is called Generative Variational Dual-Encoder (GEVADEN). As not relying on any user interaction information (e.g. time-sync comments and real-time view counts), this model is applicable to newly-uploaded/rarely-viewed videos. Experiments on our built dataset show that GEVADEN significantly outperforms several baselines, including video-summarization and highlight-detection based ones. Furthermore, we develop a pilot application of the proposed model on an online video platform with 9814 videos covering 1231 hours, which shows that our model achieves a 37.5% CTR improvement over traditional image thumbnails. This further validates the effectiveness of the proposed model and the promising application prospect of GIF thumbnails.

Smart healthcare in smart cities: wireless patient monitoring system using IoT

Recently, the development of building a smart city has been inclined rapidly with the help of emerging technologies in 5G and internet of things (IoT). In addition with smart city, concepts of smart traffic systems and smart health are also being developed. Combining the above two, smart navigation system for ambulances is being coined. When critical patients are referred by an ambulance, there is a lot of time wastage in facilitation of information and the hospital is completely unaware of the patient’s parameters who is arriving beforehand. Therefore, patient monitor and ambulance tracking system is an efficient system used to carry out a quick thirty-second diagnosis using heartbeat, temperature, breath rate sensors to record vital patient parameters required initially by the doctors to start any treatment and remotely transmit these parameters over wireless medium to the hospital even before the ambulance is deployed. The patient also gets an instant way to request an ambulance on touch of a button without having to call up the hospital and also instantly send a short message service (SMS) to an emergency contact giving both the hospital and contact necessary emergency details a lot earlier. This saves a lot of time, each second of which is important to the patient at life risk. The patient can keep a track of the ambulance’s location, which gives them an idea of its arrival and can also get instant navigation toward the nearest hospital for themselves if need be. The device can be kept with the patient and also installed inside the ambulance. The device enables the use of IoT sensors and android applications for better user interaction and efficient information transmission. The model dictated here is potentially outperformed with easiness and in a better way toward smart healthcare monitoring services.

Joint representation learning with ratings and reviews for recommendation

Recommender system is an important technique to find the information that the users may be interested by their feedbacks. However, it is still a challenge to model the preference of users due to the sparsity of user feedbacks. To alleviate this problem, many methods are developed by extracting information from various kinds of auxiliary information that are related to the users. In the auxiliary information, review is the popular one, since it can reflect both user preferences and item characteristics. Moreover, the review can generate plausible recommendation explanations in the recommendation results. In this paper, we propose a hybrid deep collaborative filtering model that jointly learns rating embedding and textural feature from ratings and reviews respectively. Specifically, two embedding layers are employed to learn rating embedding for users and items based on the interactions, and two attention-based GRU networks attempt to learn context-aware representation as textural feature for users and items from reviews. To leverage the contribution between rating embedding and textual feature and obtain the fused features for users and items, a proposed gating mechanism is used. Then an interaction-learning layer is adopted to learn the user and item interaction information based on the fused user and item features. The prediction score is obtained with the factorization machine. Experimental results on six real-world datasets demonstrate the superior performance of the proposed method over several state-of-the-art methods.

Joint user mention behavior modeling for mentionee recommendation

As an emerging online interaction service in Twitter-like social media systems, mention serves to significantly improve both user interaction experience and information propagation. In recent years, the problem of mentionee recommendation, i.e., recommending mentionees (mentioned users) when mentioners (mentioning users) mention others, has received considerable attention. However, the extreme sparsity of mentioner-mentionee matrix creates a severe challenge. While an increasing line of work has exploited diverse effects such as the textual content and spatio-temporal context influences to cope with this challenge, there lacks a comprehensive study of the joint effect of all these influencing factors. In light of this, we propose a joint latent-class probabilistic model, named Joint Topic-Area Model (JTAM), to tackle the mentionee recommendation problem by simultaneously learning and modeling users’ semantic interests, the spatio-temporal mentioning patterns of mentioners, the geographical distribution of mentionees, and their joint effects on users’ mention behaviors in a unified way. Moreover, to facilitate online query performance, we design an efficient query answering approach that enables fast top-k mentionee recommendation. To evaluate the performance of our method, we conduct extensive experiments on a large real-world dataset. The results demonstrate the superiority of our method in recommending mentionees in terms of both effectiveness and efficiency compared with other state-of-the-art methods.

Screen-Cam Robust and Blind Watermarking for Tile Satellite Images

To achieve watermark synchronization from an image with perspective distortion is a challenge. This paper proposes a screen-cam robust and blind watermarking scheme for tile satellite images, which means we do not need any user interaction or additional information in watermark detection. To achieve this, at the watermark embedding side, we divide tiles into synchronization tiles and message tiles, and propose a discrete Fourier transform (DFT) based embedding method to cope with the quality degradation caused by screen-cam process. At the extraction side, based on the idea of template matching, we first use a synchronization response index to estimate an appropriate scale level and positions of embedded synchronization watermarks from the noise component, which is estimated by Wiener filter. Then, we propose verification, selection, and precise locations methods to ensure the effectiveness and accuracy of synchronization detection results. After that, we can extract the regions of message tiles. Finally, we extract watermark message based on the local max value from the DFT domain of the noise component. Experimental results demonstrate the validity of the proposed scheme against common attacks and screen-cam attack with a tripod as well as handhold shooting.

OA user behavior analysis with the heterogeneous information network model

Analysis of network users’ behaviors is an important part for improving network security. This paper analyzes users’ group behavior via their interactions within the Office Application (OA) system. Specifically, we construct a heterogeneous information network model based on the interactive messages among users in the OA system. The model contains two types of nodes: user and topic nodes, and relationships between users and topics that are encoded in matrices. We then elicit several meta paths in the model, which integrates both user interaction and semantic information, and propose a community division method for group user behavior analysis. To verify the proposed method, we conduct experiments on five-year data collected from a real OA system. The results evaluated by an improved indicator named Normalized Mutual Information show the effectiveness of our work.

Virtual reality with the integrated automatic presentation scrip for improving concepts understanding of Urban Design - a case study in Tatsumi region of Tokyo Bay zone, Japan

Urban environment is the place that human depends on to survive. The quality of urban design concepts will directly impact the change of urban environment. However, due to the lack of user interaction and clear information exchange between users and designers, it is difficult for users to understand urban design concepts, and sometimes even misunderstanding will be caused. Thus, the present paper seeks to improve users’ understanding of urban design concepts via available communication media and using useful presentation technology (VR with the integrated APS). In this research, we recruited 60 participants to view the script video (the running result of APS) or/and free navigate in the virtual environment of Tatsumi region, Tokyo Bay Zone, Japan. By studying the data of subjective feedbacks, we investigated how APS impacts participants’ cognition and evaluation of Sustainable Taches and Reactivate Space (STARS, the design concepts). Furthermore, qualitative interview was developed to further understand participants’ design concepts understanding and to verify whether VR with the integrated APS can improve users’ understanding of urban design concepts. Results from this research explains the future use of VR with the integrated APS technology as an efficient tool in urban design schemes presentation.

Tag-aware dynamic music recommendation

We present a tag-aware dynamic music recommendation framework that achieves personalized and accurate music recommendations to users. The proposed framework leverages the available semantic labels (in terms of tags) of music tracks to complement a highly sparse user-item interaction matrix, which effectively addresses the data sparsity issue faced by most music recommendation systems. Music tracks are more accurately represented by aggregating the latent factors derived from both the tag space and the user interaction information. The proposed framework further employs a Gaussian state-space model to capture the evolving nature of users’ preferences over time, which helps achieve time-sensitive recommendation of music. A variational approximation is developed to achieve fast inference and learning of model parameters. Experiments conducted using actual music data and comparison with state-of-the-art competitive recommendation algorithms help demonstrate the effectiveness of the proposed framework.

Static Scene Illumination Estimation from Videos with Applications

We present a system that automatically recovers scene geometry and illumination from a video, providing a basis for various applications. Previous image based illumination estimation methods require either user interaction or external information in the form of a database. We adopt structure-from-motion and multi-view stereo for initial scene reconstruction, and then estimate an environment map represented by spherical harmonics (as these perform better than other bases). We also demonstrate several video editing applications that exploit the recovered geometry and illumination, including object insertion (e.g., for augmented reality), shadow detection, and video relighting.

Research on Collaborative Filtering Algorithm in Microblogging Recommendation

The personalized demand of users in social networks is an important factor of the user satisfaction. Integrated user microblog content and user interaction information, introduce the concept of interest similarity and interaction trust. This paper proposes an improved collaborative filtering recommendation algorithm. Computing user interest similarity and user interaction trust, linear fitting of the 2 sub similarity. It can calculate the total similarity between the users of micro-blog personalized recommendation. The experimental result shows that the new algorithm can effectively improve the accuracy, quality and user satisfaction of recommendation system in social networks.