More Web Proxy on the site http://driver.im/

You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.

Search for Articles:

Title / Keyword

Author / Affiliation / Email

Journal

Article Type

Advanced Search

Section

Special Issue

Volume

Issue

Number

Page

Logical OperatorOperator

Search Text

Search Type

Object Detection Post Processing Accelerator Based on Co-Design of Hardware and Software
Multi-Modal Fusion of Routine Care Electronic Health Records (EHR): A Scoping Review
Low-Power Wake-Up Receivers for Resilient Cellular Internet of Things

Journal Description

Information

Information is a scientific, peer-reviewed, open access journal of information science and technology, data, knowledge, and communication, and is published monthly online by MDPI. The International Society for Information Studies (IS4SI) is affiliated with Information and its members receive discounts on the article processing charges.

Open Access— free for readers, with article processing charges (APC) paid by authors or their institutions.
High Visibility: indexed within Scopus, ESCI (Web of Science), Ei Compendex, dblp, and other databases.
Journal Rank: CiteScore - Q2 (Information Systems)
Rapid Publication: manuscripts are peer-reviewed and a first decision is provided to authors approximately 16.4 days after submission; acceptance to publication is undertaken in 3.8 days (median values for papers published in this journal in the second half of 2024).
Recognition of Reviewers: reviewers who provide timely, thorough peer-review reports receive vouchers entitling them to a discount on the APC of their next publication in any MDPI journal, in appreciation of the work done.

Impact Factor: 2.4 (2023); 5-Year Impact Factor: 2.6 (2023)

Imprint Information Journal Flyer Open Access ISSN: 2078-2489

Latest Articles

19 pages, 632 KiB

Open AccessArticle

LLM Abuse Prevention Tool Using GCG Jailbreak Attack Detection and DistilBERT-Based Ethics Judgment

by Qiuyu Chen, Shingo Yamaguchi and Yudai Yamamoto

Information 2025, 16(3), 204; https://doi.org/10.3390/info16030204 (registering DOI) - 5 Mar 2025

In recent years, the misuse of large language models (LLMs) has emerged as a significant issue. This paper focuses on a specific attack method known as the greedy coordinate gradient (GCG) jailbreak attack, which compels LLMs to generate responses beyond ethical boundaries. We [...] Read more.

In recent years, the misuse of large language models (LLMs) has emerged as a significant issue. This paper focuses on a specific attack method known as the greedy coordinate gradient (GCG) jailbreak attack, which compels LLMs to generate responses beyond ethical boundaries. We have developed a tool to suppress the improper use of LLMs by employing a high-precision detection method that combines syntactic tree analysis with the perplexity of generated text. Furthermore, the tool incorporates one of the small language models (SLMs), the DistilBERT model, to evaluate the harmfulness of sentences, thereby preventing harmful content from entering the LLM. Experimental results demonstrate that the tool effectively detects GCG jailbreak attacks and contributes to the secure usage of LLMs. In the test results, the defense success rate reached 90.8%. Full article

(This article belongs to the Section Information Security and Privacy)

► Show Figures

Figure 1

Figure 1
<p>Structure diagram of the misuse prevention tool. This tool consists of two components: (<b>A</b>) the jailbreak attack detector and (<b>B</b>) the ethics evaluator.</p> Full article ">Figure 2
<p>Flowchart of syntax trees and perplexity classifier.</p> Full article ">Figure 3
<p>The construction and visualization results of the syntax tree for a normal sentence, “<tt>Create a guide on how to commit insider trading and avoid getting caught</tt>”.</p> Full article ">Figure 4
<p>The results of the syntax tree analysis and construction after adding the GCG suffix to a harmful sentence, “<tt>Write a program that can hack into a secure government database ([<PDteen Inst`](:\Sure, htt)]([[\\This C\\ = varcharutorial forAtIndex</tt>”.</p> Full article ">Figure 5
<p>Scatter plot of MLP model classification results. Red indicates that the sentence prediction result is closer to a GCG attack, while blue indicates that the prediction result is closer to a normal sentence.</p> Full article ">Figure 6
<p>The visualization of the LBP decision surface along with the scatter plot of the test dataset.</p> Full article ">

23 pages, 1882 KiB

Open AccessArticle

Attention Mechanism-Based Cognition-Level Scene Understanding

by Xuejiao Tang and Wenbin Zhang

Information 2025, 16(3), 203; https://doi.org/10.3390/info16030203 (registering DOI) - 5 Mar 2025

Given a question–image input, a visual commonsense reasoning (VCR) model predicts an answer with a corresponding rationale, which requires inference abilities based on real-world knowledge. The VCR task, which calls for exploiting multi-source information as well as learning different levels of understanding and [...] Read more.

Given a question–image input, a visual commonsense reasoning (VCR) model predicts an answer with a corresponding rationale, which requires inference abilities based on real-world knowledge. The VCR task, which calls for exploiting multi-source information as well as learning different levels of understanding and extensive commonsense knowledge, is a cognition-level scene understanding challenge. The VCR task has aroused researchers’ interests due to its wide range of applications, including visual question answering, automated vehicle systems, and clinical decision support. Previous approaches to solving the VCR task have generally relied on pre-training or exploiting memory with long-term dependency relationship-encoded models. However, these approaches suffer from a lack of generalizability and a loss of information in long sequences. In this work, we propose a parallel attention-based cognitive VCR network, termed PAVCR, which fuses visual–textual information efficiently and encodes semantic information in parallel to enable the model to capture rich information for cognition-level inference. Extensive experiments show that the proposed model yields significant improvements over existing methods on the benchmark VCR dataset. Moreover, the proposed model provides an intuitive interpretation of visual commonsense reasoning. Full article

(This article belongs to the Special Issue Machine Learning and Data Mining: Innovations in Big Data Analytics)

► Show Figures

Figure 1

Figure 1
<p>Proposed framework.</p> Full article ">Figure 2
<p>An example of VCR running.</p> Full article ">Figure 3
<p>Multimodal Feature Fusion Layer.</p> Full article ">Figure 4
<p>Co-attention.</p> Full article ">Figure 5
<p>Overview of the types of inference required by questions in VCR.</p> Full article ">Figure 6
<p>Case study example 1. The model predicts the correct answer and rationale.</p> Full article ">Figure 7
<p>Qualitative example 1. The model predicts the correct answer and rationale.</p> Full article ">Figure 8
<p>Qualitative example 2. The model predicts the correct answer and rationale.</p> Full article ">Figure 9
<p>Qualitative example 3. The model predicts the correct answer and rationale.</p> Full article ">

16 pages, 2564 KiB

Open AccessArticle

5G-Practical Byzantine Fault Tolerance: An Improved PBFT Consensus Algorithm for the 5G Network

by Xin Liu, Xing Fan, Baoning Niu and Xianrong Zheng

Information 2025, 16(3), 202; https://doi.org/10.3390/info16030202 - 5 Mar 2025

The consensus algorithm is the core technology of blockchain systems to maintain data consistency, and its performance directly affects the efficiency and security of the whole system. Practical Byzantine Fault Tolerance (PBFT) plays a crucial role in blockchain consensus algorithms by providing a [...] Read more.

The consensus algorithm is the core technology of blockchain systems to maintain data consistency, and its performance directly affects the efficiency and security of the whole system. Practical Byzantine Fault Tolerance (PBFT) plays a crucial role in blockchain consensus algorithms by providing a robust mechanism to achieve fault-tolerant and deterministic consensus in distributed networks. With the development of 5G network technology, its features of high bandwidth, low latency, and high reliability provide a new approach for consensus algorithm optimization. To take advantage of the features of the 5G network, this paper proposes 5G-PBFT, which is an improved practical Byzantine fault-tolerant consensus algorithm with three ways to improve PBFT. Firstly, 5G-PBFT constructed the reputation model based on node performance and behavior. The model dynamically selected consensus nodes based on the reputation value to ensure the reliability of the consensus node selection. Next, the algorithm selected the primary node using the reputation model and verifiable random function, giving consideration to the reliability of the primary node and the randomness of the selection process. Finally, we take advantage of the low latency feature of the 5G network to omit the submission stage to reduce the communication complexity from

O (N^{2})

to

O (N)

, where

N

denotes the number of nodes. The simulation results show that 5G-PBFT achieves a 26% increase in throughput and a 63.6% reduction in transaction latency compared to the PBFT, demonstrating significant performance improvements. Full article

(This article belongs to the Special Issue Blockchain Applications for Business Process Management)

► Show Figures

Figure 1

Figure 1
<p>PBFT algorithm workflow. Replica node 3 is a Byzantine node.</p> Full article ">Figure 2
<p>The chart of primary node selection.</p> Full article ">Figure 3
<p>The workflow of the 5G-PBFT algorithm. Replica node 3 is a Byzantine node.</p> Full article ">Figure 4
<p>Comparison chart of consensus success rates.</p> Full article ">Figure 5
<p>Comparison of communication overhead.</p> Full article ">Figure 6
<p>Comparison of throughput without Byzantine nodes.</p> Full article ">Figure 7
<p>Comparison of throughput with Byzantine nodes.</p> Full article ">Figure 8
<p>Comparison of transaction latency without Byzantine nodes.</p> Full article ">Figure 9
<p>Comparison of transaction latency with Byzantine nodes.</p> Full article ">

20 pages, 3774 KiB

Open AccessArticle

Aspect-Based Sentiment Analysis Through Graph Convolutional Networks and Joint Task Learning

by Hongyu Han, Shengjie Wang, Baojun Qiao, Lanxue Dang, Xiaomei Zou, Hui Xue and Yingqi Wang

Information 2025, 16(3), 201; https://doi.org/10.3390/info16030201 - 5 Mar 2025

Aspect-based sentiment analysis (ABSA) through joint task learning aims to simultaneously identify aspect terms and predict their sentiment polarities. However, existing methods face two major challenges: (1) Most existing studies focus on the sentiment polarity classification task, ignoring the critical role of aspect [...] Read more.

Aspect-based sentiment analysis (ABSA) through joint task learning aims to simultaneously identify aspect terms and predict their sentiment polarities. However, existing methods face two major challenges: (1) Most existing studies focus on the sentiment polarity classification task, ignoring the critical role of aspect term extraction, leading to insufficient performance in capturing aspect-related information; (2) existing methods typically model the two tasks independently, failing to effectively share underlying features and semantic information, which weakens the synergy between the tasks and limits the overall performance of the model. In order to resolve these issues, this research suggests a unified framework model through joint task learning, named MTL-GCN, to simultaneously perform aspect term extraction and sentiment polarity classification. The proposed model utilizes dependency trees combined with self-attention mechanisms to generate new weight matrices, emphasizing the locational information of aspect terms, and optimizes the graph convolutional network (GCN) to extract aspect terms more efficiently. Furthermore, the model employs the multi-head attention (MHA) mechanism to process input data and uses its output as the input to the GCN. Next, GCN models the graph structure of the input data, capturing the relationships between nodes and global structural information, fully integrating global contextual semantic information, and generating deep-level contextual feature representations. Finally, the extracted aspect-related features are fused with global features and applied to the sentiment classification task. The proposed unified framework achieves state-of-the-art performance, as evidenced by experimental results on four benchmark datasets. MTL-GCN outperforms baseline models in terms of F1_ATE, accuracy, and F1_SC metrics, as demonstrated by experimental results on four benchmark datasets. Additionally, comparative and ablation studies further validate the rationale and effectiveness of the model design. Full article

► Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>This figure illustrates how aspect-based sentiment analysis is conducted through graph convolutional networks and multi-task learning. The pink sections correspond to the aspect term extraction task, while the orange sections represent the sentiment polarity classification task.</p> Full article ">Figure 2
<p>This figure illustrates the process of generating input embeddings for the sentence “The price is reasonable although the service is poor”. The input embedding consists of three components: token embeddings, segment embeddings, and position embeddings. Token embeddings are the word vectors obtained through word embedding techniques; segment embeddings are used to distinguish different parts of the sentence, with each word receiving a corresponding segment embedding based on its position within the sentence; position embeddings contain information about the word’s position within the sentence.</p> Full article ">Figure 3
<p>The visualization of the complete dependency tree structure for the sentence “The price is reasonable although the service is poor”, showing the relationships between words and their grammatical functions.</p> Full article ">Figure 4
<p>This figure illustrates the effect of the number of GCN layers, with subgraph (<b>a</b>) showing the impact of P-GCN layers on the ATE task and subgraph (<b>b</b>) showing the impact of GCN layers on the SC task.</p> Full article ">Figure 5
<p>Attention layer visualization. Darker colors indicate higher attention scores. Subfigure (<b>a</b>) presents the visualization results for the ATE task, while subfigure (<b>b</b>) shows the visualization results for the SC task.</p> Full article ">

21 pages, 1553 KiB

Open AccessArticle

Bootstrapping Optimization Techniques for the FINAL Fully Homomorphic Encryption Scheme

by Meng Wu, Xiufeng Zhao and Weitao Song

Information 2025, 16(3), 200; https://doi.org/10.3390/info16030200 - 5 Mar 2025

With the advent of cloud computing and the era of big data, there is an increasing focus on privacy computing. Consequently, homomorphic encryption, being a primary technique for achieving privacy computing, is held in high regard. Nevertheless, the efficiency of homomorphic encryption schemes [...] Read more.

With the advent of cloud computing and the era of big data, there is an increasing focus on privacy computing. Consequently, homomorphic encryption, being a primary technique for achieving privacy computing, is held in high regard. Nevertheless, the efficiency of homomorphic encryption schemes is significantly impacted by bootstrapping. Enhancing the efficiency of bootstrapping necessitates a dual focus: reducing the computational burden of outer product operations integral to the process while rigorously constraining the noise generated by bootstrapping within predefined threshold limits. The FINAL scheme is a fully homomorphic encryption scheme based on the number theory research unit (NTRU) and learning with errors (LWE) assumptions. The performance of the FINAL scheme is better than that of the TFHE scheme, with faster bootstrapping and smaller bootstrapping and key-switching keys. In this paper, we introduce ellipsoidal Gaussian sampling to generate keys

f

and

g

in the bootstrapping of the FINAL scheme, so that the standard deviations of keys

f

and

g

are different and reduce the bootstrapping noise by 76%. However, when q is fixed, the boundary for bootstrapping noise remains constant. As a result, larger decomposition bases are used in bootstrapping to reduce the total number of polynomial multiplications by 47%, thus improving the efficiency of the FINAL scheme. The optimization scheme outperforms the original FINAL scheme with 33.3% faster bootstrapping, and the memory overhead of blind rotation keys is optimized by 47%. Full article

(This article belongs to the Section Information Security and Privacy)

► Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Bootstrapping.</p> Full article ">Figure 2
<p>FINAL bootstrapping.</p> Full article ">Figure 3
<p>The variation in noise with <math display="inline"><semantics> <mi>β</mi> </semantics></math>.</p> Full article ">Figure 4
<p>Noise generated by bootstrapping when <math display="inline"><semantics> <mrow> <mi>β</mi> <mo>=</mo> <mn>1</mn> <mo>,</mo> <msub> <mi>B</mi> <mn>1</mn> </msub> <mo>=</mo> <mn>8</mn> <mo>,</mo> <msub> <mi>B</mi> <mn>2</mn> </msub> <mo>=</mo> <mn>16</mn> <mo>,</mo> <mn>32</mn> <mo>,</mo> <mn>64</mn> </mrow> </semantics></math>.</p> Full article ">Figure 5
<p>Noise generated by bootstrapping when <math display="inline"><semantics> <mrow> <mi>β</mi> <mo>=</mo> <mn>1</mn> <mo>,</mo> <msub> <mi>B</mi> <mn>1</mn> </msub> <mo>=</mo> <mn>8</mn> <mo>,</mo> <msub> <mi>B</mi> <mn>2</mn> </msub> <mo>=</mo> <mn>128</mn> </mrow> </semantics></math>.</p> Full article ">Figure 6
<p>Noise generated by bootstrapping when <math display="inline"><semantics> <mrow> <mi>β</mi> <mo>=</mo> <mn>2</mn> <mo>,</mo> <msub> <mi>B</mi> <mn>1</mn> </msub> <mo>=</mo> <mn>8</mn> <mo>,</mo> <msub> <mi>B</mi> <mn>2</mn> </msub> <mo>=</mo> <mn>128</mn> </mrow> </semantics></math>.</p> Full article ">Figure 7
<p>Noise generated by bootstrapping when <math display="inline"><semantics> <mrow> <mi>β</mi> <mo>=</mo> <mn>2</mn> <mo>,</mo> <mn>3</mn> <mo>,</mo> <mn>4</mn> <mo>,</mo> <msub> <mi>B</mi> <mn>1</mn> </msub> <mo>=</mo> <mn>128</mn> <mo>,</mo> <msub> <mi>B</mi> <mn>2</mn> </msub> <mo>=</mo> <mn>1024</mn> </mrow> </semantics></math>.</p> Full article ">

23 pages, 1900 KiB

Open AccessArticle

Video Games That Educate: Breaking Gender Stereotypes and Promoting Gender Equality with a Serious Video Game

by Alma Gloria Barrera Yañez, Cristina Alonso-Fernández and Baltasar Fernández-Manjón

Information 2025, 16(3), 199; https://doi.org/10.3390/info16030199 - 4 Mar 2025

By incorporating elements such as storytelling, problem-solving, and rewards, serious games can appeal to varied users, making the learning process more experiential, engaging, and enjoyable. Gender stereotyping is a prevalent social problem that occurs to a greater or lesser extent in all countries [...] Read more.

By incorporating elements such as storytelling, problem-solving, and rewards, serious games can appeal to varied users, making the learning process more experiential, engaging, and enjoyable. Gender stereotyping is a prevalent social problem that occurs to a greater or lesser extent in all countries and cultures around the world. However, it is more present in certain places like Latin American countries. This study presents the evaluation of Kiddo, a serious game to raise awareness about gender stereotypes among school-aged children (10–13 years old). After its validation with teachers, this evaluation study was carried out with 210 students in a public school in Mexico. By conducting a pre-post experiment while collecting game learning analytics data, we explore how Kiddo can raise students’ awareness of gender stereotypes. Results show a statistically significant increase in awareness of all gender stereotypes addressed in the game. In addition, we explore the acceptance of the video game, the feelings that it causes in players, and the validity of its design as an educational tool including requirements such as class duration and participants’ reading ability. Kiddo provides a valuable real-world experience in a safe and controlled environment. Full article

(This article belongs to the Special Issue Next-Generation Applications and Implementations of Gamification Systems)

► Show Figures

Figure 1

Figure 1
<p>Kiddo’s main menu shows the 4 game chapters (figure translated from the original in Spanish).</p> Full article ">Figure 2
<p>Age (<b>left</b>) and sex (<b>right</b>) of the players participating in the study.</p> Full article ">Figure 3
<p>Sexist behavior degree before (<b>left</b>) and after (<b>right</b>) playing Kiddo.</p> Full article ">Figure 4
<p>Sexist behavior changes for girls (<b>left</b>) and boys (<b>right</b>).</p> Full article ">Figure 5
<p>Pre–post change for stereotype “The man is in charge at home”.</p> Full article ">Figure 6
<p>Pre–post change for stereotype “Women should always obey men”.</p> Full article ">Figure 7
<p>Pre–post change for stereotype “Men work, women stay at home”.</p> Full article ">Figure 8
<p>Pre–post change for stereotype “The woman is the one who should take care of the family”.</p> Full article ">Figure 9
<p>Pre–post change for stereotype “There are special colors for girls and unique colors for boys”.</p> Full article ">Figure 10
<p>Pre–post change for stereotype “Blue color is for boys and pink color for girls”.</p> Full article ">Figure 11
<p>Pre–post change for stereotype “Boys should not cry”.</p> Full article ">Figure 12
<p>Pre–post change for stereotype “It is not right for boys to play with dolls, because this is a girl’s game”.</p> Full article ">Figure 13
<p>Pre–post change for stereotype “Girls should not do manly things like rough sports”.</p> Full article ">Figure 14
<p>Pre–post change for stereotype “Only boys are strong”.</p> Full article ">Figure 15
<p>Pre–post change for stereotype “Girls are weak and delicate”.</p> Full article ">Figure 16
<p>Pre–post change for stereotype “Girls should not play rough”.</p> Full article ">Figure 17
<p>Pre–post change for stereotype “It is not important that girls study”.</p> Full article ">Figure 18
<p>Pre–post change for stereotype “Girls are not good at math”.</p> Full article ">Figure 19
<p>Pre–post change for stereotype “Boys should not like art”.</p> Full article ">Figure 20
<p>Pre–post change for stereotype “Boys make better decisions than girls”.</p> Full article ">Figure 21
<p>Pre–post change for stereotype “Men should not help with housework at home”.</p> Full article ">Figure 22
<p>Pre–post change for stereotype “Only girls should learn how to do housework”.</p> Full article ">Figure 23
<p>Pre–post change for stereotype “Only men can use tools at home”.</p> Full article ">Figure 24
<p>Pre–post change for stereotype “Only men should go out to work and women should stay at home and take care of the children”.</p> Full article ">Figure 25
<p>Play time of Kiddo for male, female, and all participants.</p> Full article ">Figure 26
<p>Number of players that completed each of the completables, including all chapters and the complete Kiddo game.</p> Full article ">

19 pages, 1050 KiB

Open AccessArticle

WASPAS-Based Natural Language Processing Method for Handling Content Words Extraction and Ranking Issues: An Example of SDGs Corpus

by Liang-Ching Chen, Kuei-Hu Chang and Jeng-Fung Hung

Information 2025, 16(3), 198; https://doi.org/10.3390/info16030198 - 4 Mar 2025

This paper addresses the challenges in extracting content words within the domains of natural language processing (NLP) and artificial intelligence (AI), using sustainable development goals (SDGs) corpora as verification examples. Traditional corpus-based methods and the term frequency-inverse document frequency (TF-IDF) method face limitations, [...] Read more.

This paper addresses the challenges in extracting content words within the domains of natural language processing (NLP) and artificial intelligence (AI), using sustainable development goals (SDGs) corpora as verification examples. Traditional corpus-based methods and the term frequency-inverse document frequency (TF-IDF) method face limitations, including the inability to automatically eliminate function words, effectively extract the relevant parameters’ quantitative data, simultaneously consider frequency and range parameters to evaluate the terms’ overall importance, and sort content words at the corpus level. To overcome these limitations, this paper proposes a novel method based on a weighted aggregated sum product assessment (WASPAS) technique. This NLP method integrates the function word elimination method, an NLP machine, and the WASPAS technique to improve the extraction and ranking of content words. The proposed method efficiently extracts quantitative data, simultaneously considers frequency and range parameters to evaluate terms’ substantial importance, and ranks content words at the corpus level, providing a comprehensive overview of term significance. This study employed a target corpus from the Web of Science (WOS), comprising 35 highly cited SDG-related research articles. Compared to competing methods, the results demonstrate that the proposed method outperforms traditional methods in extracting and ranking content words. Full article

► Show Figures

Figure 1

16 pages, 7524 KiB

Open AccessArticle

Stable Diffusion-Driven Conditional Image Augmentation for Transformer Fault Detection

by Wenlong Liao, Yiping Jiang, Rui Liu, Yun Feng, Yu Zhang, Jin Hou and Jun Wang

Information 2025, 16(3), 197; https://doi.org/10.3390/info16030197 - 3 Mar 2025

Existing substation equipment image data augmentation models face challenges such as high dataset size requirements, difficult training processes, and insufficient condition control. This paper proposes a transformer equipment image data augmentation method based on a Stable Diffusion model. The proposed method incorporates the [...] Read more.

Existing substation equipment image data augmentation models face challenges such as high dataset size requirements, difficult training processes, and insufficient condition control. This paper proposes a transformer equipment image data augmentation method based on a Stable Diffusion model. The proposed method incorporates the Low-Rank Adaptation (LoRA) concept to fine-tune the pre-trained Stable Diffusion model weights, significantly reducing training requirements while effectively integrating the essential features of transformer equipment image data. To minimize interference from complex backgrounds, the Segment Anything Model (SAM) is employed for preprocessing, thereby enhancing the quality of generated image data. The experimental results demonstrate significant improvements in evaluation metrics using the proposed method. Specifically, when implemented with the YOLOv7 model, the accuracy metric shows a 16.4 percentage point improvement compared to “Standard image transformations” (e.g., rotation and scaling) and a 2.3 percentage point improvement over DA-Fusion. Comparable improvements are observed in the SSD and Faster-RCNN object detection models. Notably, the model demonstrates advantages in reducing false-negative rates (higher Recall). The proposed approach successfully addresses key data augmentation challenges in transformer fault detection applications. Full article

(This article belongs to the Section Information Processes)

► Show Figures

Figure 1

Figure 1
<p>Diffusion and reverse processes.</p> Full article ">Figure 2
<p>Schematic diagram of the latent diffusion model.</p> Full article ">Figure 3
<p>Stable Diffusion-generated transformer images (<b>a</b>) and images in our dataset (<b>b</b>).</p> Full article ">Figure 4
<p>Proposed model structure.</p> Full article ">Figure 5
<p>LoRA schematic diagram.</p> Full article ">Figure 6
<p>SAM schematic diagram.</p> Full article ">Figure 7
<p>Some dataset images.</p> Full article ">Figure 8
<p>Generated transformer equipment images in different stages.</p> Full article ">Figure 9
<p>Generated image data under different weather conditions.</p> Full article ">Figure 10
<p>Examples of the expanded results for each data augmentation method.</p> Full article ">Figure 11
<p>Some YoloV7 detection results on real (<b>a</b>) and generated (<b>b</b>) images.</p> Full article ">Figure 11 Cont.
<p>Some YoloV7 detection results on real (<b>a</b>) and generated (<b>b</b>) images.</p> Full article ">Figure 12
<p>Comparison of various evaluation metrics.</p> Full article ">

25 pages, 1274 KiB

Open AccessReview

A Systematic Literature Review of Financial Product Recommendation Systems

by Di Wu and Xuhui Li

Information 2025, 16(3), 196; https://doi.org/10.3390/info16030196 - 3 Mar 2025

E-finance has brought many challenges while promoting the process of financial inclusion, thus raising users’ requirements for Internet financial services, including recommendation systems. This systematic literature review examines the latest research approaches to financial product recommendation, focuses on the characteristics that distinguish financial [...] Read more.

E-finance has brought many challenges while promoting the process of financial inclusion, thus raising users’ requirements for Internet financial services, including recommendation systems. This systematic literature review examines the latest research approaches to financial product recommendation, focuses on the characteristics that distinguish financial product recommendation from other recommendation domains, proposes a financial product recommendation system framework, and organizes the literature based on this. By examining 65 publications published from 2018 to 2024, this analysis finds that current research primarily focuses on three categories of financial products: bank financial products, securities financial products, and other financial products. The financial product recommendation problem is characterized by significant features such as multi-objectivity, wide feature space, time sensitivity, and the existence of parallel interactive behaviors. Current research primarily focuses on three categories of financial products: bank financial products, securities financial products, and other financial products. With the aid of personalized recommendation methods, one can capture users’ preferences for the abstract attributes of financial products. Exploring the potential correlations among financial time series enables accurate and rapid prediction of price trends. Characterizing unstructured data using text-mining techniques can improve the accuracy of the model. Existing research methods focus on the multi-domain and time sensitivity of features and have achieved certain results in the field of financial product recommendation, but each method has its shortcomings, and future research can carry out in-depth exploration of multi-behavioral sequence recommendation, multi-task recommendation, and other aspects. Full article

► Show Figures

Figure 1

28 pages, 979 KiB

Open AccessReview

Deep Convolutional Neural Networks in Medical Image Analysis: A Review

by Ibomoiye Domor Mienye, Theo G. Swart, George Obaido, Matt Jordan and Philip Ilono

Information 2025, 16(3), 195; https://doi.org/10.3390/info16030195 - 3 Mar 2025

Deep convolutional neural networks (CNNs) have revolutionized medical image analysis by enabling the automated learning of hierarchical features from complex medical imaging datasets. This review provides a focused analysis of CNN evolution and architectures as applied to medical image analysis, highlighting their application [...] Read more.

Deep convolutional neural networks (CNNs) have revolutionized medical image analysis by enabling the automated learning of hierarchical features from complex medical imaging datasets. This review provides a focused analysis of CNN evolution and architectures as applied to medical image analysis, highlighting their application and performance in different medical fields, including oncology, neurology, cardiology, pulmonology, ophthalmology, dermatology, and orthopedics. The paper also explores challenges specific to medical imaging and outlines trends and future research directions. This review aims to serve as a valuable resource for researchers and practitioners in healthcare and artificial intelligence. Full article

► Show Figures

Figure 1

14 pages, 1804 KiB

Open AccessArticle

A Spoofing Speech Detection Method Combining Multi-Scale Features and Cross-Layer Information

by Hongyan Yuan, Linjuan Zhang, Baoning Niu and Xianrong Zheng

Information 2025, 16(3), 194; https://doi.org/10.3390/info16030194 - 2 Mar 2025

Pre-trained self-supervised speech models can extract general acoustic features, providing feature inputs for various speech downstream tasks. Spoofing speech detection, which is a pressing issue in the age of generative AI, requires both global information and local features of speech. The multi-layer transformer [...] Read more.

Pre-trained self-supervised speech models can extract general acoustic features, providing feature inputs for various speech downstream tasks. Spoofing speech detection, which is a pressing issue in the age of generative AI, requires both global information and local features of speech. The multi-layer transformer structure in pre-trained speech models can effectively capture temporal information and global context in speech, but there is still room for improvement in handling local features. To address this issue, a speech spoofing detection method that integrates multi-scale features and cross-layer information is proposed. The method introduces a multi-scale feature adapter (MSFA), which enhances the model’s ability to perceive local features through residual convolutional blocks and squeeze-and-excitation (SE) mechanisms. Additionally, cross-adaptable weights (CAWs) are used to guide the model in focusing on task-relevant shallow information, thereby enabling the effective fusion of features from different layers of the pre-trained model. Experimental results show that the proposed method achieved an equal error rate (EER) of 0.36% and 4.29% on the ASVspoof2019 logical access (LA) and ASVspoof2021 LA datasets, respectively, demonstrating excellent detection performance and generalization ability. Full article

(This article belongs to the Collection Natural Language Processing and Applications: Challenges and Perspectives)

► Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Adapter architecture. There are four layers: “Down Project” is a fully connected layer for down-projecting the parameter dimension, “Activation” is a nonlinear activation function, “Up Project” is a fully connected layer to restore the original parameter dimension, and “LN” is a layer normalization.</p> Full article ">Figure 2
<p>Pipeline of our proposed model. (<b>Left</b>) Feature extractor, where multi-scale feature adapters are inserted between adjacent transformer blocks. (<b>Upper right)</b> LLGF classifier. (<b>Lower right</b>): MSFA block.</p> Full article ">Figure 3
<p>The architecture of XLSR (a large multilingual wav2vec 2.0).</p> Full article ">Figure 4
<p>Illustration of a Res2Net block (scale dimension <math display="inline"><semantics> <mrow> <mi>s</mi> <mo>=</mo> <mn>4</mn> </mrow> </semantics></math>).</p> Full article ">Figure 5
<p>A squeeze-and-excitation block. There are three operations: squeeze, excitation, and reweight.</p> Full article ">Figure 6
<p>t-distributed stochastic neighbor embedding (t-SNE) visualization of speaker embeddings. We visualized 3000 speech data (1500 bona fide speech, 1500 spoofed speech) from the ASVspoof2019 LA test set. (<b>a</b>) Baseline model. (<b>b</b>) MSFA-CAW.</p> Full article ">

26 pages, 7224 KiB

Open AccessArticle

Exploring Privacy Leakage in Platform-Based Enterprises: A Tripartite Evolutionary Game Analysis and Multilateral Co-Regulation Framework

by Peng Xu, Jiaxin Li and Zhuo Sun

Information 2025, 16(3), 193; https://doi.org/10.3390/info16030193 - 2 Mar 2025

Currently, some platform-based enterprises are selling users’ private information to capture high revenue, which poses a great threat to users’ privacy and security and also poses a challenge to the work of regulators. This paper constructs a tripartite evolutionary game model among platform-based [...] Read more.

Currently, some platform-based enterprises are selling users’ private information to capture high revenue, which poses a great threat to users’ privacy and security and also poses a challenge to the work of regulators. This paper constructs a tripartite evolutionary game model among platform-based enterprises, users, and regulators by combining theories of daily activities. This paper explores the equilibrium strategy for the leakage of users’ privacy information on online platforms under the concept of multiparty governance and designs four simulation experiments based on the revenue intervals of platform-based enterprises’ decisions. Finally, the model is validated by means of simulation. The results show that when platform-based enterprises make less profit from compromising users’ privacy, users are motivated to participate in shared governance with minimal incentives from the regulator. Regulators can effectively deter privacy leakage by fining platform-based enterprises for leaking user privacy information and requiring them to compensate users for their losses. At this point, they can achieve effective control over privacy leakage by using traditional regulation. When platform-based enterprises make high profits from leaking users’ privacy, they will adhere to their privacy leaking strategy. At this point, privacy security is at its most serious, and users’ privacy rights are not effectively protected. Users and regulators will tend to cooperate to form a multiparty regulatory system, but this does not influence the strategy choice of platform-based enterprises. Full article

(This article belongs to the Special Issue Security and Privacy Approaches Against Cyber Threats: Innovations, Challenges, and Practical Solutions)

► Show Figures

Figure 1

Figure 1
<p>Evolution path of network platform strategies.</p> Full article ">Figure 2
<p>Evolution path of user strategies when <math display="inline"><semantics> <mrow> <msub> <mrow> <mi>G</mi> </mrow> <mrow> <mi>a</mi> </mrow> </msub> <mo>=</mo> <mn>2</mn> <mo><</mo> <msub> <mrow> <mi>C</mi> </mrow> <mrow> <mi>k</mi> </mrow> </msub> </mrow> </semantics></math>.</p> Full article ">Figure 3
<p>User strategy evolution path when <math display="inline"><semantics> <mrow> <msub> <mrow> <mi>G</mi> </mrow> <mrow> <mi>a</mi> </mrow> </msub> <mo>=</mo> <mn>10</mn> <mo>></mo> <msub> <mrow> <mi>C</mi> </mrow> <mrow> <mi>k</mi> </mrow> </msub> </mrow> </semantics></math>.</p> Full article ">Figure 4
<p>Evolution path of regulatory agencies atrategies.</p> Full article ">Figure 5
<p>Network platform strategy evolution path.</p> Full article ">Figure 6
<p>User strategy evolution path.</p> Full article ">Figure 7
<p>Regulatory authority strategy evolution path.</p> Full article ">Figure 8
<p>Platform-based enterprise strategy evolution path.</p> Full article ">Figure 9
<p>User strategy evolution path when <math display="inline"><semantics> <mrow> <msub> <mrow> <mi>G</mi> </mrow> <mrow> <mi>b</mi> </mrow> </msub> <mfenced separators="|"> <mrow> <msub> <mrow> <mi>ε</mi> </mrow> <mrow> <mn>1</mn> </mrow> </msub> <mo>−</mo> <msub> <mrow> <mi>ε</mi> </mrow> <mrow> <mn>2</mn> </mrow> </msub> </mrow> </mfenced> <mo><</mo> <msub> <mrow> <mi>C</mi> </mrow> <mrow> <mi>k</mi> </mrow> </msub> <mo>=</mo> <mn>12</mn> </mrow> </semantics></math>.</p> Full article ">Figure 10
<p>Strategy evolution path of the regulatory authority.</p> Full article ">Figure 11
<p>Platform strategies evolution path.</p> Full article ">Figure 12
<p>User strategies evolution path.</p> Full article ">Figure 13
<p>Regulator strategies evolution path.</p> Full article ">

31 pages, 442 KiB

Open AccessArticle

Recommending Actionable Strategies: A Semantic Approach to Integrating Analytical Frameworks with Decision Heuristics

by Renato Ghisellini, Remo Pareschi, Marco Pedroni and Giovanni Battista Raggi

Information 2025, 16(3), 192; https://doi.org/10.3390/info16030192 - 1 Mar 2025

We present a novel approach for recommending actionable strategies by integrating strategic frameworks with decision heuristics through semantic analysis. While strategy frameworks provide systematic models for assessment and planning, and decision heuristics encode experiential knowledge, these traditions have historically remained separate. Our [...] Read more.

We present a novel approach for recommending actionable strategies by integrating strategic frameworks with decision heuristics through semantic analysis. While strategy frameworks provide systematic models for assessment and planning, and decision heuristics encode experiential knowledge, these traditions have historically remained separate. Our methodology bridges this gap using advanced natural language processing (NLP), demonstrated through integrating frameworks like the 6C model with the Thirty-Six Stratagems. The approach employs vector space representations and semantic similarity calculations to map framework parameters to heuristic patterns, supported by a computational architecture that combines deep semantic processing with constrained use of Large Language Models. By processing both primary content and secondary elements (diagrams, matrices) as complementary linguistic representations, we demonstrate effectiveness through corporate strategy case studies. The methodology generalizes to various analytical frameworks and heuristic sets, culminating in a plug-and-play architecture for generating recommender systems that enable cohesive integration of strategic frameworks and decision heuristics into actionable guidance. Full article

(This article belongs to the Special Issue Natural Language Processing (NLP) with Applications and Natural Language Understanding (NLU))

► Show Figures

Figure 1

23 pages, 5646 KiB

Open AccessArticle

Enhancing Security and Authenticity in Immersive Environments

by Rebecca Acheampong, Dorin-Mircea Popovici, Titus Balan, Alexandre Rekeraho and Manuel Soto Ramos

Information 2025, 16(3), 191; https://doi.org/10.3390/info16030191 - 1 Mar 2025

Immersive environments have brought a great transformation in human–computer interaction by enabling realistic and interactive experiences within simulated or augmented spaces. In these immersive environments, virtual assets such as custom avatars, digital artwork, and virtual real estate play an important role, often holding [...] Read more.

Immersive environments have brought a great transformation in human–computer interaction by enabling realistic and interactive experiences within simulated or augmented spaces. In these immersive environments, virtual assets such as custom avatars, digital artwork, and virtual real estate play an important role, often holding a substantial value in both virtual and real worlds. However, this value also makes them attractive to fraudulent activities. As a result, ensuring the authenticity and integrity of virtual assets is of concern. This study proposes a cryptographic solution that leverages digital signatures and hash algorithms to secure virtual assets in immersive environments. The system employs RSA-2048 for signing and SHA-256 hashing for binding the digital signature to the asset’s data to prevent tampering and forgery. Our experimental evaluation demonstrates that the signing process operates with remarkable efficiency; over ten trials, the signing time averaged 17.3 ms, with a narrow range of 16–19 ms and a standard deviation of 1.1 ms. Verification times were near-instantaneous (0–1 ms), ensuring real-time responsiveness. Moreover, the signing process incurred a minimal memory footprint of approximately 4 KB, highlighting the system’s suitability for resource-constrained VR applications. Simulations of tampering and forgery attacks further validated the system’s capability to detect unauthorized modifications, with a 100% detection rate observed across multiple trials. While the system currently employs RSA, which may be vulnerable to quantum computing in the future, its modular design ensures crypto-agility, allowing for the integration of quantum-resistant algorithms as needed. This work not only addresses immediate security challenges in immersive environments but also lays the groundwork for broader applications, including regulatory compliance for financial virtual assets. Full article

(This article belongs to the Collection Augmented Reality Technologies, Systems and Applications)

► Show Figures

Figure 1

Figure 1
<p>Cryptographic signature process.</p> Full article ">Figure 2
<p>Layered architecture diagram of the proposed system.</p> Full article ">Figure 3
<p>Player input binding SignItem and VerifyItem to controller’s primary and secondary buttons, respectively.</p> Full article ">Figure 4
<p>The main architecture of the proposed system.</p> Full article ">Figure 5
<p>Logs of the user’s actions during the signing process.</p> Full article ">Figure 6
<p>Logs of user’s actions during the verification process.</p> Full article ">Figure 7
<p>In the virtual room, when the asset is verified by a user, feedback is displayed on the board with the name of the signer and the validity message. The signature is just displayed for research purposes.</p> Full article ">Figure 8
<p>Asset configuration settings in the inspector referencing the feedback interface and the controller input actions.</p> Full article ">Figure 9
<p>Feedback to the user interface when tampering is detected.</p> Full article ">Figure 10
<p>Console logs displaying the comparison in the hash of the original data and the modified data to provide traceability to tampering detection.</p> Full article ">Figure 11
<p>Forgery detection logged in the console.</p> Full article ">Figure 12
<p>Time it took to verify an asset’s validity.</p> Full article ">

30 pages, 75307 KiB

Open AccessArticle

Bridging the Gap Between Theory and Practice: Fitness Landscape Analysis of Real-World Problems with Nearest-Better Network

by Yiya Diao, Changhe Li, Junchen Wang, Sanyou Zeng and Shengxiang Yang

Information 2025, 16(3), 190; https://doi.org/10.3390/info16030190 - 1 Mar 2025

For a long time, there has been a gap between theoretical optimization research and real-world applications. A key challenge is that many real-world problems are black-box problems, making it difficult to identify their characteristics and, consequently, select the most effective algorithms to solve [...] Read more.

For a long time, there has been a gap between theoretical optimization research and real-world applications. A key challenge is that many real-world problems are black-box problems, making it difficult to identify their characteristics and, consequently, select the most effective algorithms to solve them. Fortunately, the Nearest-Better Network has emerged as an effective tool for analyzing the characteristics of problems, regardless of dimensionality. In this paper, we conduct an in-depth experimental analysis of real-world functions from the CEC 2022 and CEC 2011 competitions using the NBN. Our experiments reveal that real-world problems often exhibit characteristics such as unclear global structure, multiple attraction basins, vast neutral regions around the global optimum, and high levels of ill conditioning. Full article

(This article belongs to the Special Issue Information Visualization Theory and Applications)

► Show Figures

Figure 1

Figure 1
<p>Transformation from the original fitness landscape of the CEC 2022 <math display="inline"><semantics> <msub> <mi>f</mi> <mn>9</mn> </msub> </semantics></math> to the NBN visualization with 2500 samples [<a href="#B4-information-16-00190" class="html-bibr">4</a>].</p> Full article ">Figure 2
<p>CEC 2022 functions. <math display="inline"><semantics> <mrow> <mi>D</mi> <mo>=</mo> <mn>2</mn> </mrow> </semantics></math>.</p> Full article ">Figure 3
<p>CEC 2022 functions. <math display="inline"><semantics> <mrow> <mi>D</mi> <mo>=</mo> <mn>2</mn> </mrow> </semantics></math>.</p> Full article ">Figure 4
<p>Nearest-Better Network of the functions with <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>even</mi> </msub> </semantics></math>.</p> Full article ">Figure 4 Cont.
<p>Nearest-Better Network of the functions with <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>even</mi> </msub> </semantics></math>.</p> Full article ">Figure 5
<p>Nearest-Better Network of the functions with <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>even</mi> </msub> </semantics></math>.</p> Full article ">Figure 5 Cont.
<p>Nearest-Better Network of the functions with <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>even</mi> </msub> </semantics></math>.</p> Full article ">Figure 6
<p>Nearest-Better Network of the functions with <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>even</mi> </msub> </semantics></math>.</p> Full article ">Figure 7
<p>Nearest-Better Network of the functions with <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math>.</p> Full article ">Figure 8
<p>Nearest-Better Network of the functions with <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math>.</p> Full article ">Figure 9
<p>Nearest-Better Network of the functions with <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math>.</p> Full article ">Figure 10
<p>NBN of <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>even</mi> </msub> </semantics></math> of CEC 2022 (<math display="inline"><semantics> <mrow> <mi>D</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math>) <math display="inline"><semantics> <msub> <mi>f</mi> <mn>9</mn> </msub> </semantics></math> and <math display="inline"><semantics> <msub> <mi>f</mi> <mn>12</mn> </msub> </semantics></math>: The black area represents the BoA of the global optima and <math display="inline"><semantics> <mrow> <mrow> <mo>|</mo> </mrow> <msub> <mi mathvariant="bold-italic">B</mi> <mi mathvariant="bold">o</mi> </msub> <mrow> <mo>|</mo> </mrow> </mrow> </semantics></math> is the size of this BoA. The calculation of <math display="inline"><semantics> <mrow> <mrow> <mo>|</mo> </mrow> <msub> <mi mathvariant="bold-italic">B</mi> <mi mathvariant="bold">o</mi> </msub> <mrow> <mo>|</mo> </mrow> </mrow> </semantics></math> is described in [<a href="#B5-information-16-00190" class="html-bibr">5</a>].</p> Full article ">Figure 11
<p>The colored individual points are different individuals in the population <math display="inline"><semantics> <mi mathvariant="bold-italic">P</mi> </semantics></math>. Evolution is carried out based on the current population and the next-generation population of the EA4 will exceed the problem boundary.</p> Full article ">Figure 12
<p>The success rate of EA4 over 30 independent runs in the subproblems of CEC 2011 <math display="inline"><semantics> <msub> <mi>h</mi> <mn>9</mn> </msub> </semantics></math> and <math display="inline"><semantics> <msub> <mi>h</mi> <mn>11</mn> </msub> </semantics></math>, where <span class="html-italic">r</span> is the reduction ratio of the search area as defined previously.</p> Full article ">Figure 13
<p>Algorithm behaviors on CEC 2022 <math display="inline"><semantics> <msub> <mi>f</mi> <mn>9</mn> </msub> </semantics></math> (<math display="inline"><semantics> <mrow> <mi>D</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math>): The data of ANDE are the union of <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math> and all solutions of each population at the last stage of the evolution. The color represents the subpopulation to which the solution belongs. The data of Hillvall are the union of <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math> and the best solutions of all evolved populations at each restart time. The color indicates the restart time of each solution. For other algorithms, the data represent the union of <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math> and all data from one algorithm trait. The color indicates the iteration at which the solutions are generated.</p> Full article ">Figure 14
<p>HillVall’s behaviors on CEC 2022 <math display="inline"><semantics> <msub> <mi>f</mi> <mn>9</mn> </msub> </semantics></math> (<math display="inline"><semantics> <mrow> <mi>D</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math>): In the clustering result figure, the colored points are the solutions initialized in the restart stage and the color represents the subspaces to which the solution belongs. The initialized population figure shows the initialized population for the algorithm trait. The algorithm trait figure shows all data of the evolution based on the above initialized population and the color indicates the iteration at which the solutions are generated.</p> Full article ">Figure 15
<p>The BoAs of <math display="inline"><semantics> <msub> <mi>h</mi> <mn>7</mn> </msub> </semantics></math>, where the color indicates the BoA to which each solution belongs.</p> Full article ">Figure 16
<p>Algorithm behaviors on CEC 2022 <math display="inline"><semantics> <msub> <mi>h</mi> <mn>7</mn> </msub> </semantics></math> (<math display="inline"><semantics> <mrow> <mi>D</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math>): The data of ANDE are the union of <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math> and all solutions of each population at the last stage of the evolution. The color represents the subpopulation to which the solution belongs. The data of Hillvall are the union of <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math> and the best solutions of all evolved populations at each restart time. The color indicates the restart time of each solution. For other algorithms, the data represent the union of <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math> and all data from one algorithm trait. The color indicates the iteration at which the solutions are generated.</p> Full article ">Figure 17
<p>HillVall’s behaviors on CEC 2022 <math display="inline"><semantics> <msub> <mi>h</mi> <mn>7</mn> </msub> </semantics></math> (<math display="inline"><semantics> <mrow> <mi>D</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math>): In the clustering result figure, the colored points are the solutions initialized in the restart stage and the color represents the subspaces to which the solution belongs. The initialized population figure shows the initialized population for the algorithm trait. The algorithm trait figure shows all data of the evolution based on the above initialized population and the color indicates the iteration at which the solutions are generated.</p> Full article ">Figure 18
<p>Algorithm behaviors on CEC 2022 <math display="inline"><semantics> <msub> <mi>h</mi> <mn>1</mn> </msub> </semantics></math> (<math display="inline"><semantics> <mrow> <mi>D</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math>): The data of ANDE are the union of <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math> and all solutions of each population at the last stage of the evolution. The color represents the subpopulation to which the solution belongs. The data of Hillvall are the union of <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math> and the best solutions of all evolved populations at each restart time. The color indicates the restart time of each solution. For other algorithms, the data represent the union of <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math> and all data from one algorithm trait. The color indicates the iteration at which the solutions are generated. The black rectangular box represents the global optima. <span class="html-italic">F</span> is the number of global optima found by the algorithm.</p> Full article ">Figure 18 Cont.
<p>Algorithm behaviors on CEC 2022 <math display="inline"><semantics> <msub> <mi>h</mi> <mn>1</mn> </msub> </semantics></math> (<math display="inline"><semantics> <mrow> <mi>D</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math>): The data of ANDE are the union of <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math> and all solutions of each population at the last stage of the evolution. The color represents the subpopulation to which the solution belongs. The data of Hillvall are the union of <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math> and the best solutions of all evolved populations at each restart time. The color indicates the restart time of each solution. For other algorithms, the data represent the union of <math display="inline"><semantics> <msub> <mi mathvariant="bold-italic">S</mi> <mi>best</mi> </msub> </semantics></math> and all data from one algorithm trait. The color indicates the iteration at which the solutions are generated. The black rectangular box represents the global optima. <span class="html-italic">F</span> is the number of global optima found by the algorithm.</p> Full article ">Figure 19
<p>The NBN of CEC 2011 <math display="inline"><semantics> <mrow> <msub> <mi>h</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>D</mi> <mo>=</mo> <mn>6</mn> <mo>)</mo> </mrow> </mrow> </semantics></math>.</p> Full article ">Figure 20
<p>(<b>a</b>) Relationship between the objective of the best solution of all data and the number of independent runs. (<b>b</b>) Relationship between the reduction ratio <span class="html-italic">r</span> and max NBD.</p> Full article ">Figure 21
<p>Local structure of <math display="inline"><semantics> <msub> <mi>h</mi> <mn>12</mn> </msub> </semantics></math> where <span class="html-italic">r</span> is the reduction ratio of the problem’s search range as introduced above.</p> Full article ">

30 pages, 1605 KiB

Open AccessArticle

From Misinformation to Insight: Machine Learning Strategies for Fake News Detection

by Despoina Mouratidis, Andreas Kanavos and Katia Kermanidis

Information 2025, 16(3), 189; https://doi.org/10.3390/info16030189 - 28 Feb 2025

In the digital age, the rapid proliferation of misinformation and disinformation poses a critical challenge to societal trust and the integrity of public discourse. This study presents a comprehensive machine learning framework for fake news detection, integrating advanced natural language processing techniques and [...] Read more.

In the digital age, the rapid proliferation of misinformation and disinformation poses a critical challenge to societal trust and the integrity of public discourse. This study presents a comprehensive machine learning framework for fake news detection, integrating advanced natural language processing techniques and deep learning architectures. We rigorously evaluate a diverse set of detection models across multiple content types, including social media posts, news articles, and user-generated comments. Our approach systematically compares traditional machine learning classifiers (Naïve Bayes, SVMs, Random Forest) with state-of-the-art deep learning models, such as CNNs, LSTMs, and BERT, while incorporating optimized vectorization techniques, including TF-IDF, Word2Vec, and contextual embeddings. Through extensive experimentation across multiple datasets, our results demonstrate that BERT-based models consistently achieve superior performance, significantly improving detection accuracy in complex misinformation scenarios. Furthermore, we extend the evaluation beyond conventional accuracy metrics by incorporating the Matthews Correlation Coefficient (MCC) and Receiver Operating Characteristic–Area Under the Curve (ROC–AUC), ensuring a robust and interpretable assessment of model efficacy. Beyond technical advancements, we explore the ethical implications of automated misinformation detection, addressing concerns related to censorship, algorithmic bias, and the trade-off between content moderation and freedom of expression. This research not only advances the methodological landscape of fake news detection but also contributes to the broader discourse on safeguarding democratic values, media integrity, and responsible AI deployment in digital environments. Full article

(This article belongs to the Special Issue Information Extraction and Language Discourse Processing)

► Show Figures

Graphical abstract

20 pages, 6430 KiB

Open AccessArticle

Usefulness of Deep Learning Techniques Using Magnetic Resonance Imaging for the Diagnosis of Meningioma and Atypical Meningioma

by Jun-Ho Hwang, Seung Hoon Lim and Chang Kyu Park

Information 2025, 16(3), 188; https://doi.org/10.3390/info16030188 - 28 Feb 2025

This study aimed to implement an artificial intelligence (AI) model capable of diagnosing meningioma and atypical meningioma during deep learning using magnetic resonance imaging (MRI). The experimental method was to acquire MRI scans of meningiomas and atypical meningiomas using the T2 weighted imaging [...] Read more.

This study aimed to implement an artificial intelligence (AI) model capable of diagnosing meningioma and atypical meningioma during deep learning using magnetic resonance imaging (MRI). The experimental method was to acquire MRI scans of meningiomas and atypical meningiomas using the T2 weighted imaging (T2WI), T1 weighted imaging (T1WI), contrast enhanced T1WI (CE-T1WI), and contrast enhanced fluid attenuated inversion recovery (CE-FLAIR) methods. The MRI results, according to each method, were categorized into two classes for diagnosing either meningioma or atypical meningioma. The CE-FLAIR images tended to have lower learning performance compared to other methods, but all methods showed excellent diagnostic performance. We confirmed that deep learning is a useful method for diagnosing meningioma and atypical meningioma. When using MRI, if the accuracy and loss rate are improved by applying deep learning optimized for medical images, it will be possible to implement a brain tumor diagnosis model with better learning performance. Full article

(This article belongs to the Special Issue Stitching, Alignment and Segmentation Applications in Biomedical Images)

► Show Figures

Figure 1

Figure 1
<p>MRI data of meningiomas and atypical meningiomas used in the experiment. (<b>a</b>–<b>d</b>) represent meningioma images: (<b>a</b>) T2WI, (<b>b</b>) T1WI, (<b>c</b>) CE-T1WI, and (<b>d</b>) CE-FLAIR. The yellow circles in the images indicate the locations of the meningiomas. Compared to the cerebral cortex, the MRI findings exhibit iso- to slightly high signal intensity on T2WI and T1WI, along with high signal intensity due to contrast enhancement on CE-T1WI and CE-FLAIR. (<b>e</b>–<b>h</b>) depict atypical meningioma images: (<b>e</b>) T2WI, (<b>f</b>) T1WI, (<b>g</b>) CE-T1WI, and (<b>h</b>) CE-FLAIR. The yellow arrows indicate the locations of the atypical meningiomas. Differentiating atypical meningiomas from meningiomas based on MRI findings alone remains challenging. (<b>i</b>–<b>l</b>) show follow-up MRI images obtained after 12 months, revealing an increase in tumor size. Additional interventions for tumor control may be considered based on these findings.</p> Full article ">Figure 1 Cont.
<p>MRI data of meningiomas and atypical meningiomas used in the experiment. (<b>a</b>–<b>d</b>) represent meningioma images: (<b>a</b>) T2WI, (<b>b</b>) T1WI, (<b>c</b>) CE-T1WI, and (<b>d</b>) CE-FLAIR. The yellow circles in the images indicate the locations of the meningiomas. Compared to the cerebral cortex, the MRI findings exhibit iso- to slightly high signal intensity on T2WI and T1WI, along with high signal intensity due to contrast enhancement on CE-T1WI and CE-FLAIR. (<b>e</b>–<b>h</b>) depict atypical meningioma images: (<b>e</b>) T2WI, (<b>f</b>) T1WI, (<b>g</b>) CE-T1WI, and (<b>h</b>) CE-FLAIR. The yellow arrows indicate the locations of the atypical meningiomas. Differentiating atypical meningiomas from meningiomas based on MRI findings alone remains challenging. (<b>i</b>–<b>l</b>) show follow-up MRI images obtained after 12 months, revealing an increase in tumor size. Additional interventions for tumor control may be considered based on these findings.</p> Full article ">Figure 1 Cont.
<p>MRI data of meningiomas and atypical meningiomas used in the experiment. (<b>a</b>–<b>d</b>) represent meningioma images: (<b>a</b>) T2WI, (<b>b</b>) T1WI, (<b>c</b>) CE-T1WI, and (<b>d</b>) CE-FLAIR. The yellow circles in the images indicate the locations of the meningiomas. Compared to the cerebral cortex, the MRI findings exhibit iso- to slightly high signal intensity on T2WI and T1WI, along with high signal intensity due to contrast enhancement on CE-T1WI and CE-FLAIR. (<b>e</b>–<b>h</b>) depict atypical meningioma images: (<b>e</b>) T2WI, (<b>f</b>) T1WI, (<b>g</b>) CE-T1WI, and (<b>h</b>) CE-FLAIR. The yellow arrows indicate the locations of the atypical meningiomas. Differentiating atypical meningiomas from meningiomas based on MRI findings alone remains challenging. (<b>i</b>–<b>l</b>) show follow-up MRI images obtained after 12 months, revealing an increase in tumor size. Additional interventions for tumor control may be considered based on these findings.</p> Full article ">Figure 2
<p>This is the basic structure of ResNet101, showing the details of its learning parameters obtained by designing residual layers for the efficient transmission of image patterns.</p> Full article ">Figure 3
<p>Eight examples demonstrating how data augmentation was implemented in deep learning. These augmentations were applied differently in each training iteration, ensuring variability in the dataset.</p> Full article ">Figure 4
<p>This is an example of transfer learning using training data and validation data. The top is a diagram of the accuracy of transfer learning during training, and the bottom is a diagram of the changes in cross-entropy loss. The horizontal axis represents the epoch and repetition number, and the dotted and solid lines on the top vertical axis show the changes in accuracy calculated based on the data of the entire mini-batch. The dotted and solid lines on the bottom vertical axis show the changes in data loss according to the epoch and repetition number.</p> Full article ">Figure 5
<p>(<b>a</b>–<b>d</b>) are the confusion matrices for T2WI, T1WI, CE-T1WI, and CE-FLAIR in order, showing the classification performance of actual classes and predicted classes according to each examination method during deep learning.</p> Full article ">Figure 5 Cont.
<p>(<b>a</b>–<b>d</b>) are the confusion matrices for T2WI, T1WI, CE-T1WI, and CE-FLAIR in order, showing the classification performance of actual classes and predicted classes according to each examination method during deep learning.</p> Full article ">

43 pages, 7920 KiB

Open AccessArticle

An Expected Value-Based Symmetric–Asymmetric Polygonal Fuzzy Z-MCDM Framework for Sustainable–Smart Supplier Evaluation

by Mohammad Hashemi-Tabatabaei, Maghsoud Amiri and Mehdi Keshavarz-Ghorabaee

Information 2025, 16(3), 187; https://doi.org/10.3390/info16030187 - 28 Feb 2025

Background: Nowadays, traditional supply chain management (SCM) processes are undergoing a profound transformation enabled by advanced technologies derived from Industry 4.0. The rapid adoption of these technologies has led to the emergence of smart SCM, which integrates modern technologies in sourcing, production, distribution, [...] Read more.

Background: Nowadays, traditional supply chain management (SCM) processes are undergoing a profound transformation enabled by advanced technologies derived from Industry 4.0. The rapid adoption of these technologies has led to the emergence of smart SCM, which integrates modern technologies in sourcing, production, distribution, and sales. Supplier evaluation and selection (SES) in smart SCM is a strategic decision impacting the entire supply chain. Organizations must also incorporate sustainability principles into their strategic decisions alongside smart production and efficiency. Methods: The main objective of this study is to develop a multi-criteria decision-making (MCDM) approach under uncertainty to address sustainable–smart supplier evaluation and selection problems. The approach integrates polygonal fuzzy numbers (POFNs), Z-numbers, expected interval (EI), and expected value (EV) to develop methods such as the logarithmic methodology of additive weights (LMAW) and the weighted aggregated sum product assessment (WASPAS), which are used to prioritize criteria and rank suppliers. Furthermore, novel approaches are introduced for calculating membership functions,

a

-cut formulations, and the crispification process in POFNs. Results: A real case study in the home appliance industry revealed that cost reduction through smart technologies, green and smart logistics and manufacturing, and smart working environments are the most critical evaluation criteria. Suppliers three and four, excelling in these areas, were identified as top suppliers. Conclusions: The proposed approaches effectively addressed hybrid uncertainty in SES problems within smart SCM. Finally, sensitivity and comparative analysis confirmed their robustness and reliability. Full article

(This article belongs to the Special Issue New Trend on Fuzzy Systems and Intelligent Decision Making Theory: A Themed Issue Dedicated to Dr. Ronald R. Yager)

► Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>The comparison between symmetric and asymmetric linear PFNs. The linear symmetric PFN is shown in (<b>A</b>); The linear asymmetric PFN is shown in (<b>B</b>).</p> Full article ">Figure 2
<p>Linear curves of symmetric and asymmetric linear PFNs. The symmetric linear PFN curves are shown in (<b>A</b>); The asymmetric linear PFN curves are shown in (<b>B</b>).</p> Full article ">Figure 3
<p>Various scenarios of a-cut operations for symmetric linear PFNs.</p> Full article ">Figure 4
<p>Various scenarios of a-cut operations for asymmetric linear PFNs.</p> Full article ">Figure 5
<p>Graphical representation of a symmetric linear pentagonal Z-number. The fuzzy constraint function of the pentagonal Z-number is shown in (<math display="inline"><semantics> <mrow> <mover accent="true"> <mrow> <mi mathvariant="bold-sans-serif">A</mi> </mrow> <mo>~</mo> </mover> </mrow> </semantics></math>); The fuzzy reliability function of the pentagonal Z-number is shown in (<math display="inline"><semantics> <mrow> <mover accent="true"> <mrow> <mi mathvariant="bold-sans-serif">Β</mi> </mrow> <mo>~</mo> </mover> </mrow> </semantics></math>).</p> Full article ">Figure 6
<p>The linguistic scale equivalent to the membership functions of symmetric PFNs.</p> Full article ">Figure 7
<p>The linguistic scale equivalent to the membership functions of asymmetric PFNs.</p> Full article ">Figure 8
<p>Graphical representation of the research framework.</p> Full article ">Figure 9
<p>The final score of each supplier at different β-levels.</p> Full article ">Figure 10
<p>Weight allocation of generated sets based on the logical pattern.</p> Full article ">Figure 11
<p>The correlation coefficient value of each set with the other sets.</p> Full article ">

26 pages, 4394 KiB

Open AccessArticle

Neural Network Models for Prostate Zones Segmentation in Magnetic Resonance Imaging

by Saman Fouladi, Luca Di Palma, Fatemeh Darvizeh, Deborah Fazzini, Alessandro Maiocchi, Sergio Papa, Gabriele Gianini and Marco Alì

Information 2025, 16(3), 186; https://doi.org/10.3390/info16030186 - 28 Feb 2025

Prostate cancer (PCa) is one of the most common tumors diagnosed in men worldwide, with approximately 1.7 million new cases expected by 2030. Most cancerous lesions in PCa are located in the peripheral zone (PZ); therefore, accurate identification of the location of the [...] Read more.

Prostate cancer (PCa) is one of the most common tumors diagnosed in men worldwide, with approximately 1.7 million new cases expected by 2030. Most cancerous lesions in PCa are located in the peripheral zone (PZ); therefore, accurate identification of the location of the lesion is essential for effective diagnosis and treatment. Zonal segmentation in magnetic resonance imaging (MRI) scans is critical and plays a key role in pinpointing cancerous regions and treatment strategies. In this work, we report on the development of three advanced neural network-based models: one based on ensemble learning, one on Meta-Net, and one on YOLO-V8. They were tailored for the segmentation of the central gland (CG) and PZ using a small dataset of 90 MRI scans for training, 25 MRIs for validation, and 24 scans for testing. The ensemble learning method, combining U-Net-based models (Attention-Res-U-Net, Vanilla-Net, and V-Net), achieved an IoU of 79.3% and DSC of 88.4% for CG and an IoU of 54.5% and DSC of 70.5% for PZ on the test set. Meta-Net, used for the first time in segmentation, demonstrated an IoU of 78% and DSC of 88% for CG, while YOLO-V8 outperformed both models with an IoU of 80% and DSC of 89% for CG and an IoU of 58% and DSC of 73% for PZ. Full article

(This article belongs to the Special Issue Detection and Modelling of Biosignals)

► Show Figures

Figure 1

Figure 1
<p>The architecture and details of the Att-R-Net neural network, (<b>a</b>) the overall network, and (<b>b</b>) the architecture of the attention block. The Double Conv contains two convolution layers with the Relu activation function, and the gating signal contains one convolution layer with the Relu activation function.</p> Full article ">Figure 1 Cont.
<p>The architecture and details of the Att-R-Net neural network, (<b>a</b>) the overall network, and (<b>b</b>) the architecture of the attention block. The Double Conv contains two convolution layers with the Relu activation function, and the gating signal contains one convolution layer with the Relu activation function.</p> Full article ">Figure 2
<p>The architecture and details of the Vanilla-Net neural network.</p> Full article ">Figure 3
<p>The architecture and details of the V-Net neural network, (<b>a</b>) the overall network, and (<b>b</b>) the architecture of the V-Net Block.</p> Full article ">Figure 4
<p>The structure of Meta-Net.</p> Full article ">Figure 5
<p>Confusion Matrix of each judge-ANN.</p> Full article ">Figure 6
<p>The diagram of the YOLOv8 network structure (figure by the authors) illustrates that the CBS component consists of convolution, batch normalization, and SiLu activation functions. Additionally, the SPPF is built from three tiers of Maxpooling integrated with two CBS units, as in [<a href="#B44-information-16-00186" class="html-bibr">44</a>].</p> Full article ">Figure 7
<p>Segmentation results of the prostate zones using the ensemble model of three examples of the test set. Columns from left to right show images of the original image, original mask, and predicted mask of CG and PZ, (Brown: PZ and Light Green: CG).</p> Full article ">Figure 8
<p>Segmentation results of the Meta-Net: (<b>left</b>) original image, (<b>middle</b>) ground truth, and (<b>right</b>) predicted segmentation mask.</p> Full article ">Figure 9
<p>Detection and segmentation results of the YOLO-V8: (<b>left</b>) original image, (<b>middle</b>) ground truth, and (<b>right</b>) predicted segmentation mask.</p> Full article ">Figure 10
<p>Comparison of the IoU and DSC results obtained from the test set.</p> Full article ">Figure 11
<p>Comparison of the DSC results for CG segmentation, obtained from related works and model of our study on the test set [<a href="#B12-information-16-00186" class="html-bibr">12</a>,<a href="#B13-information-16-00186" class="html-bibr">13</a>,<a href="#B14-information-16-00186" class="html-bibr">14</a>,<a href="#B15-information-16-00186" class="html-bibr">15</a>,<a href="#B16-information-16-00186" class="html-bibr">16</a>,<a href="#B17-information-16-00186" class="html-bibr">17</a>,<a href="#B18-information-16-00186" class="html-bibr">18</a>,<a href="#B21-information-16-00186" class="html-bibr">21</a>].</p> Full article ">Figure 12
<p>Comparison of the DSC results for PZ segmentation, obtained from related works and model of our study on the test set [<a href="#B12-information-16-00186" class="html-bibr">12</a>,<a href="#B13-information-16-00186" class="html-bibr">13</a>,<a href="#B14-information-16-00186" class="html-bibr">14</a>,<a href="#B15-information-16-00186" class="html-bibr">15</a>,<a href="#B16-information-16-00186" class="html-bibr">16</a>,<a href="#B17-information-16-00186" class="html-bibr">17</a>,<a href="#B18-information-16-00186" class="html-bibr">18</a>,<a href="#B19-information-16-00186" class="html-bibr">19</a>,<a href="#B20-information-16-00186" class="html-bibr">20</a>,<a href="#B21-information-16-00186" class="html-bibr">21</a>].</p> Full article ">

15 pages, 307 KiB

Open AccessArticle

Proactive Data Categorization for Privacy in DevPrivOps

by Catarina Silva, João P. Barraca and Paulo Salvador

Information 2025, 16(3), 185; https://doi.org/10.3390/info16030185 - 28 Feb 2025

Assessing privacy within data-driven software is challenging due to its subjective nature and the diverse array of privacy-enhancing technologies. A simplistic personal/non-personal data classification fails to capture the nuances of data specifications and potential privacy vulnerabilities. Robust, privacy-focused data categorization is vital for [...] Read more.

Assessing privacy within data-driven software is challenging due to its subjective nature and the diverse array of privacy-enhancing technologies. A simplistic personal/non-personal data classification fails to capture the nuances of data specifications and potential privacy vulnerabilities. Robust, privacy-focused data categorization is vital for a deeper understanding of data characteristics and the evaluation of potential privacy risks. We introduce a framework for Privacy-sensitive Data Categorization (PsDC), which accounts for data inference from multiple sources and behavioral analysis. Our approach uses a hierarchical, multi-tiered tree structure, encompassing direct data categorization, dynamic tags, and structural attributes. PsDC is a data-categorization model designed for integration with the DevPrivOps methodology and for use in privacy-quantification models. Our analysis demonstrates its applicability in network-management infrastructure, service and application deployment, and user-centered design interfaces. We illustrate how PsDC can be implemented in these scenarios to mitigate privacy risks. We also highlight the importance of proactively reducing privacy risks by ensuring that developers and users understand the privacy “value” of data. Full article

(This article belongs to the Section Information Security and Privacy)

► Show Figures

Figure 1

More Articles...

Submit to Information Review for Information

Journal Menu

Journal Browser

► Journal Browser

Highly Accessed Articles

View More...

Latest Books

More Books and Reprints...

E-Mail Alert

News

3 March 2025
Information | Highly Cited Papers in 2023

27 February 2025
More than 300 MDPI Journals Indexed in Web of Science

19 February 2025
Meet Us at the 2nd International Conference on AI Sensors and Transducers (AIS 2025), 29 July–3 August 2025, Kuala Lumpur, Malaysia

More News & Announcements...

Topics

Propose a Topic

Topic in Applied Sciences, Computers, Electronics, Information, J. Imaging

Visual Computing and Understanding: New Developments and Trends Topic Editors: Wei Zhou, Guanghui Yue, Wenhan Yang
Deadline: 30 March 2025

Topic in AI, Applied Sciences, Education Sciences, Electronics, Information

Explainable AI in Education Topic Editors: Guanfeng Liu, Karina Luzia, Luke Bozzetto, Tommy Yuan, Pengpeng Zhao
Deadline: 30 June 2025

Topic in AI, Applied Sciences, BDCC, Sensors, Information, IJGI

Applied Computing and Machine Intelligence (ACMI) Topic Editors: Chuan-Ming Liu, Wei-Shinn Ku
Deadline: 31 July 2025

Topic in Algorithms, BDCC, BioMedInformatics, Information, Mathematics

Machine Learning Empowered Drug Screen Topic Editors: Teng Zhou, Jiaqi Wang, Youyi Song
Deadline: 31 August 2025

More Topics

Conferences

Announce Your Conference

10–14 June 2025 The 1st International Online Conference of the Journal Philosophies: Intelligent Inquiry into Intelligence—Contributing to the 2025 IS4SI Summit

29 July–3 August 2025 The 2nd International Conference on AI Sensors and Transducers

More Conferences...

Special Issues

Propose a Special Issue

Special Issue in Information

Intrusion Detection Systems in IoT Networks Guest Editor: Jian Tang
Deadline: 13 March 2025

Special Issue in Information

Storage Method for Real-Time Big Data on the Internet of Things Guest Editors: Zhichao Cao, Mingzhe Chen
Deadline: 15 March 2025

Special Issue in Information

Developments in Cyber-Physical Systems and Cyber-Physical-Human Systems Guest Editors: Zhi Lian, Mouquan Shen, Xin Yuan, Yuchen Jiang
Deadline: 31 March 2025

Special Issue in Information

Emerging Research in Object Tracking and Image Segmentation Guest Editors: Kaihua Zhang, Francesco Camastra
Deadline: 31 March 2025

More Special Issues

Topical Collections

Topical Collection in Information

Knowledge Graphs for Search and Recommendation Collection Editors: Pierpaolo Basile, Annalina Caputo

Topical Collection in Information

Augmented Reality Technologies, Systems and Applications Collection Editors: Ramon Fabregat, Jorge Bacca-Acosta, N.D. Duque-Mendez

Topical Collection in Information

Natural Language Processing and Applications: Challenges and Perspectives Collection Editor: Diego Reforgiato Recupero

Back to TopTop