Pub Date : 2026-01-29DOI: 10.1016/j.ins.2026.123164
Alaa A. Najim , Safa A. Najim
This paper presents a novel methodology for analyzing bubble movement in turbulent flows by framing multi-object tracking as a global optimization problem on a spatiotemporal graph. Bubble detections across frames are modeled as nodes in a directed graph, with tracking addressed through a minimum cost flow algorithm. Unlike sequential trackers, our approach considers the entire temporal context to achieve globally optimal paths, reducing identity switches and fragmentation caused by occlusions and interactions. The pipeline integrates anisotropic Gaussian filtering and Otsu’s thresholding with a cost function enforcing movement and appearance consistency. Experimental validation across flow rates (, , ) identifies distinct systems: stable homogeneous, transitional interference, and saturated aggregation-dominated. The method simultaneously measures temporal evolution of bubble count , mean diameter , and velocity , revealing critical phenomena like saturation effects where increased airflow leads to bubble aggregation rather than an increased count.
{"title":"A novel graph-theoretic and data visualization framework for spatiotemporal bubble analysis in turbulent flows","authors":"Alaa A. Najim , Safa A. Najim","doi":"10.1016/j.ins.2026.123164","DOIUrl":"10.1016/j.ins.2026.123164","url":null,"abstract":"<div><div>This paper presents a novel methodology for analyzing bubble movement in turbulent flows by framing multi-object tracking as a global optimization problem on a spatiotemporal graph. Bubble detections across frames are modeled as nodes in a directed graph, with tracking addressed through a minimum cost flow algorithm. Unlike sequential trackers, our approach considers the entire temporal context to achieve globally optimal paths, reducing identity switches and fragmentation caused by occlusions and interactions. The pipeline integrates anisotropic Gaussian filtering and Otsu’s thresholding with a cost function enforcing movement and appearance consistency. Experimental validation across flow rates (<span><math><msub><mi>Q</mi><mn>1</mn></msub><mo>=</mo><mn>1.0</mn></math></span>, <span><math><msub><mi>Q</mi><mn>2</mn></msub><mo>=</mo><mn>2.0</mn></math></span>, <span><math><msub><mi>Q</mi><mn>3</mn></msub><mo>=</mo><mn>3.0</mn></math></span>) identifies distinct systems: stable homogeneous, transitional interference, and saturated aggregation-dominated. The method simultaneously measures temporal evolution of bubble count <span><math><mi>N</mi><mo>(</mo><mi>t</mi><mo>)</mo></math></span>, mean diameter <span><math><mover><mi>d</mi><mo>―</mo></mover><mo>(</mo><mi>t</mi><mo>)</mo></math></span>, and velocity <span><math><mover><mi>v</mi><mo>―</mo></mover><mo>(</mo><mi>t</mi><mo>)</mo></math></span>, revealing critical phenomena like saturation effects where increased airflow leads to bubble aggregation rather than an increased count.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"739 ","pages":"Article 123164"},"PeriodicalIF":6.8,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190945","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-29DOI: 10.1016/j.ins.2026.123174
Zi-Biao Feng , Hai-Yang Jia , Sheng-Sheng Wang
Although widely used for reasoning under uncertainty, Extended Belief Rule-Based (EBRB) systems are often constrained by two fundamental deficiencies: suboptimal rule base quality due to redundancy, and a theoretically irrational conjunctive model that leads to logical fallacies. To address this research gap, this paper proposes a novel framework named SNN-EBRB. It introduces a rule generation strategy based on Shared-Nearest-Neighbor Density Peak Clustering (SNNDPC) to directly construct a compact, high-quality rule base from raw data with complex distributions. Concurrently, we establish, for the first time, a theoretical evaluation framework comprising six essential properties for the conjunctive model and design a new model that is rigorously proven to satisfy all of them, fundamentally rectifying the irrational behavior of the traditional model. Extensive experiments on 13 public UCI datasets and a real-world microseismic signal identification case study demonstrate the comprehensive superiority of SNN-EBRB: its performance is statistically superior to the majority of existing EBRB variants and it achieves the number one overall average rank. Furthermore, our method reduces the number of rules by up to 80% while exhibiting millisecond-level inference efficiency, and its introduced activation factor, , provides an effective mechanism to trade off between model performance and interpretability.
{"title":"A novel approach to conjunctive relation modeling and rule generation in extended belief rule-based expert systems for classification problems","authors":"Zi-Biao Feng , Hai-Yang Jia , Sheng-Sheng Wang","doi":"10.1016/j.ins.2026.123174","DOIUrl":"10.1016/j.ins.2026.123174","url":null,"abstract":"<div><div>Although widely used for reasoning under uncertainty, Extended Belief Rule-Based (EBRB) systems are often constrained by two fundamental deficiencies: suboptimal rule base quality due to redundancy, and a theoretically irrational conjunctive model that leads to logical fallacies. To address this research gap, this paper proposes a novel framework named SNN-EBRB. It introduces a rule generation strategy based on Shared-Nearest-Neighbor Density Peak Clustering (SNNDPC) to directly construct a compact, high-quality rule base from raw data with complex distributions. Concurrently, we establish, for the first time, a theoretical evaluation framework comprising six essential properties for the conjunctive model and design a new model that is rigorously proven to satisfy all of them, fundamentally rectifying the irrational behavior of the traditional model. Extensive experiments on 13 public UCI datasets and a real-world microseismic signal identification case study demonstrate the comprehensive superiority of SNN-EBRB: its performance is statistically superior to the majority of existing EBRB variants and it achieves the number one overall average rank. Furthermore, our method reduces the number of rules by up to 80% while exhibiting millisecond-level inference efficiency, and its introduced activation factor, <span><math><mi>λ</mi></math></span>, provides an effective mechanism to trade off between model performance and interpretability.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"739 ","pages":"Article 123174"},"PeriodicalIF":6.8,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191059","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-29DOI: 10.1016/j.ins.2026.123168
Wei Wang , Zixin Huang , Ning Chen
This paper investigates the model-free finite-horizon optimal control problem of discrete-time linear time-invariant systems with a prescribed degree of stability. Initially, a novel finite-horizon cost function is formulated, and the corresponding time-varying Riccati equation (TVRE) is derived. It is proven that the solution to the TVRE ensures the exponential stability of the closed-loop system with a prescribed degree. Subsequently, a time-varying Q-function is designed, and a Q-learning-based backward-in-time algorithm is developed to estimate solutions for the TVRE and the optimal time-varying state feedback gains, all without requiring knowledge of the system dynamics. Finally, a simulation study is conducted to validate the efficacy of the proposed algorithm. It shows that as the setting parameter decreases, the degree of stability increases, and the convergence to the equilibrium point becomes faster.
{"title":"Model-free finite-horizon optimal control of linear systems with prescribed degree of stability","authors":"Wei Wang , Zixin Huang , Ning Chen","doi":"10.1016/j.ins.2026.123168","DOIUrl":"10.1016/j.ins.2026.123168","url":null,"abstract":"<div><div>This paper investigates the model-free finite-horizon optimal control problem of discrete-time linear time-invariant systems with a prescribed degree of stability. Initially, a novel finite-horizon cost function is formulated, and the corresponding time-varying Riccati equation (TVRE) is derived. It is proven that the solution to the TVRE ensures the exponential stability of the closed-loop system with a prescribed degree. Subsequently, a time-varying Q-function is designed, and a Q-learning-based backward-in-time algorithm is developed to estimate solutions for the TVRE and the optimal time-varying state feedback gains, all without requiring knowledge of the system dynamics. Finally, a simulation study is conducted to validate the efficacy of the proposed algorithm. It shows that as the setting parameter decreases, the degree of stability increases, and the convergence to the equilibrium point becomes faster.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"739 ","pages":"Article 123168"},"PeriodicalIF":6.8,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191058","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-29DOI: 10.1016/j.ins.2026.123145
Xiaopeng Yi , Chongyang Liu , Huey Tyng Cheong , Kok Lay Teo
In this paper, we investigate dynamic optimization problems governed by nonlinear fractional switched systems with multiple time-delays and subject to terminal state inequality constraints, where both the switching times and system parameters are treated as decision variables. For this problem, we first transform it into an equivalent form on a normalized time horizon with fixed switching points using a novel time-scaling transformation, with time-delays expressed in terms of subsystem durations in the original time horizon. A third-order numerical integration scheme is then applied to discretize the transformed problem, resulting in a discrete-time dynamic optimization problem. Furthermore, gradients of the cost and constraint functions with respect to the decision variables are derived, and a gradient-based optimization algorithm is developed to solve the resulting problem. Lastly, three representative numerical examples are provided to showcase the effectiveness and broad applicability of the proposed method.
{"title":"Dynamic optimization of nonlinear fractional switched systems with multiple time-delays","authors":"Xiaopeng Yi , Chongyang Liu , Huey Tyng Cheong , Kok Lay Teo","doi":"10.1016/j.ins.2026.123145","DOIUrl":"10.1016/j.ins.2026.123145","url":null,"abstract":"<div><div>In this paper, we investigate dynamic optimization problems governed by nonlinear fractional switched systems with multiple time-delays and subject to terminal state inequality constraints, where both the switching times and system parameters are treated as decision variables. For this problem, we first transform it into an equivalent form on a normalized time horizon with fixed switching points using a novel time-scaling transformation, with time-delays expressed in terms of subsystem durations in the original time horizon. A third-order numerical integration scheme is then applied to discretize the transformed problem, resulting in a discrete-time dynamic optimization problem. Furthermore, gradients of the cost and constraint functions with respect to the decision variables are derived, and a gradient-based optimization algorithm is developed to solve the resulting problem. Lastly, three representative numerical examples are provided to showcase the effectiveness and broad applicability of the proposed method.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"739 ","pages":"Article 123145"},"PeriodicalIF":6.8,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191153","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-29DOI: 10.1016/j.ins.2026.123171
Yu Su , Junyi Zhang
While edge streaming is becoming increasingly prevalent in missions that demand real-time stream processing, conventional data pipelines often suffer from inefficiencies caused by redundant memory copies, thread contention, and high latency in edge computing applications. To address these challenges, this paper proposes a new zero-copy lock-free data pipeline framework (LZDP) co-grounded in an adaptable edge streaming architecture. The contributions of this work lie in: (a) the pipelined structural zero-copy mechanism by leveraging (b) layered folding executors with memory atomic frames and synchronized atomic operators, eliminating the redundant data copying; (c) an extensible template-driven in-band control closed-loop by reconstructing the datapath for signal synchronous routing strategy to ensure multi-thread safety; and (d) lock-free synchronization under the unified architecture of the multi-producer–consumer model, implemented by a pipelined zero-coupling conception, featuring hot-pluggable components, tailored in stream-batch integrated processing. Microbenchmarks and experiments varying thread count with message sizes demonstrate that the conceived pipeline achieves 3-4x performance improvements over comparative frameworks in latency and throughput. By replacing internal execution queue backends with alternative approaches within unified LZDP logic and operator chains, experiments and ablation studies validate the synergistic effects and throughput scalability of the targeted conceptions. Further tests reveal that deploying generic streaming middleware for architectural consistency inevitably entails additional performance trade-offs, whereas the conceived architecture for specific edge scenarios offers distinct advantages in deployment cost, lightweight, and adaptability.
{"title":"A zero-copy lock-free data pipeline for edge streaming","authors":"Yu Su , Junyi Zhang","doi":"10.1016/j.ins.2026.123171","DOIUrl":"10.1016/j.ins.2026.123171","url":null,"abstract":"<div><div>While edge streaming is becoming increasingly prevalent in missions that demand real-time stream processing, conventional data pipelines often suffer from inefficiencies caused by redundant memory copies, thread contention, and high latency in edge computing applications. To address these challenges, this paper proposes a new zero-copy lock-free data pipeline framework (LZDP) co-grounded in an adaptable edge streaming architecture. The contributions of this work lie in: (a) the pipelined structural zero-copy mechanism by leveraging (b) layered folding executors with memory atomic frames and synchronized atomic operators, eliminating the redundant data copying; (c) an extensible template-driven in-band control closed-loop by reconstructing the datapath for signal synchronous routing strategy to ensure multi-thread safety; and (d) lock-free synchronization under the unified architecture of the multi-producer–consumer model, implemented by a pipelined zero-coupling conception, featuring hot-pluggable components, tailored in stream-batch integrated processing. Microbenchmarks and experiments varying thread count with message sizes demonstrate that the conceived pipeline achieves 3-4x performance improvements over comparative frameworks in latency and throughput. By replacing internal execution queue backends with alternative approaches within unified LZDP logic and operator chains, experiments and ablation studies validate the synergistic effects and throughput scalability of the targeted conceptions. Further tests reveal that deploying generic streaming middleware for architectural consistency inevitably entails additional performance trade-offs, whereas the conceived architecture for specific edge scenarios offers distinct advantages in deployment cost, lightweight, and adaptability.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"739 ","pages":"Article 123171"},"PeriodicalIF":6.8,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191060","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-29DOI: 10.1016/j.ins.2026.123170
You-hong Li , Jian-Qiang Wang , Le Gao , Tian-Yu Wang , Hao-Ming Mo
Community detection in organizational networks is vital for optimizing team structures, yet existing methods face critical challenges: Static models ignore temporal dynamics, dynamic single-layer approaches overlook cross-layer interactions, and multi-objective frameworks often optimize goals in isolation, leading to suboptimal real-world performance. We propose the Multi-Objective Dynamic Multi-Layer Hypergraph Modeling Framework (MO-DMLHM), integrating three innovations: (1) Adaptive Dynamic Hypergraph Modeling with dual-scale decay and adaptive time windowing to capture spatiotemporal dynamics; (2) Four-Dimensional Multi-Objective Optimization balancing modularity, cross-layer consistency, stability, and efficiency via Pareto-optimal NSGA-III; (3) Hybrid Encoding Evolutionary Algorithm jointly optimizing hyperedge activation and node membership through spectral clustering-guided mutation and betweenness centrality-driven crossover. Experiments on diverse organizational networks show MO-DMLHM outperforms state-of-the-art methods in detection accuracy, cross-layer alignment, and stability, reducing coordination costs by nearly 40%. Ablation studies confirm the necessity of dynamic modeling, multi-objective optimization, and hybrid encoding. MO-DMLHM resolves structural-community decoupling in dynamic multi-layer systems, advancing complex network analysis and enabling adaptive governance in organizations, with extensions to smart cities, biological networks, and financial risk management.
{"title":"MO-DMLHM: Multi-objective dynamic hypergraph modeling for cross-layer community detection in organizational networks","authors":"You-hong Li , Jian-Qiang Wang , Le Gao , Tian-Yu Wang , Hao-Ming Mo","doi":"10.1016/j.ins.2026.123170","DOIUrl":"10.1016/j.ins.2026.123170","url":null,"abstract":"<div><div>Community detection in organizational networks is vital for optimizing team structures, yet existing methods face critical challenges: Static models ignore temporal dynamics, dynamic single-layer approaches overlook cross-layer interactions, and multi-objective frameworks often optimize goals in isolation, leading to suboptimal real-world performance. We propose the Multi-Objective Dynamic Multi-Layer Hypergraph Modeling Framework (MO-DMLHM), integrating three innovations: (1) Adaptive Dynamic Hypergraph Modeling with dual-scale decay and adaptive time windowing to capture spatiotemporal dynamics; (2) Four-Dimensional Multi-Objective Optimization balancing modularity, cross-layer consistency, stability, and efficiency via Pareto-optimal NSGA-III; (3) Hybrid Encoding Evolutionary Algorithm jointly optimizing hyperedge activation and node membership through spectral clustering-guided mutation and betweenness centrality-driven crossover. Experiments on diverse organizational networks show MO-DMLHM outperforms state-of-the-art methods in detection accuracy, cross-layer alignment, and stability, reducing coordination costs by nearly 40%. Ablation studies confirm the necessity of dynamic modeling, multi-objective optimization, and hybrid encoding. MO-DMLHM resolves structural-community decoupling in dynamic multi-layer systems, advancing complex network analysis and enabling adaptive governance in organizations, with extensions to smart cities, biological networks, and financial risk management.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"739 ","pages":"Article 123170"},"PeriodicalIF":6.8,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190944","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-28DOI: 10.1016/j.ins.2026.123129
Tianhao Yuan, Xia Wang, Qiyang Sun, Yuyang Li
Hyperspectral imaging (HSI) contains abundant spatial and spectral cues, making it advantageous for a wide range of applications, including Earth observation, medicine, as well as agricultural analysis. However, conventional HSI systems are often constrained by high costs, limited acquisition efficiency, and weak adaptability to dynamic scenes. To mitigate these limitations, we propose a progressive multi-feature fusion model, termed PMFF-SRNet, for RGB-to-hyperspectral reconstruction. The proposed model progressively recovers spectral information through multiple stages, which helps reduce spectral redundancy while improving reconstruction efficiency. Furthermore, a Local–Global Spectral Attention (LGSA) module is employed to model spectral features at different granularities, where grouped self-attention focuses on local band interactions while spectral order information contributes to long-range dependency modeling. In addition, a Discrete Wavelet Attention (DWA) module is incorporated into the skip connections to enhance texture and edge restoration by exploiting the multi-scale characteristics of wavelet transforms. Results obtained on the NTIRE benchmarks indicate that PMFF-SRNet achieves competitive reconstruction performance across multiple evaluation metrics, while maintaining a lightweight and computationally efficient architecture. These findings demonstrate the strong potential of PMFF-SRNet for practical hyperspectral reconstruction tasks.
{"title":"PMFF-SRNet: A progressive multi-feature fusion network for hyperspectral image reconstruction","authors":"Tianhao Yuan, Xia Wang, Qiyang Sun, Yuyang Li","doi":"10.1016/j.ins.2026.123129","DOIUrl":"10.1016/j.ins.2026.123129","url":null,"abstract":"<div><div>Hyperspectral imaging (HSI) contains abundant spatial and spectral cues, making it advantageous for a wide range of applications, including Earth observation, medicine, as well as agricultural analysis. However, conventional HSI systems are often constrained by high costs, limited acquisition efficiency, and weak adaptability to dynamic scenes. To mitigate these limitations, we propose a progressive multi-feature fusion model, termed PMFF-SRNet, for RGB-to-hyperspectral reconstruction. The proposed model progressively recovers spectral information through multiple stages, which helps reduce spectral redundancy while improving reconstruction efficiency. Furthermore, a Local–Global Spectral Attention (LGSA) module is employed to model spectral features at different granularities, where grouped self-attention focuses on local band interactions while spectral order information contributes to long-range dependency modeling. In addition, a Discrete Wavelet Attention (DWA) module is incorporated into the skip connections to enhance texture and edge restoration by exploiting the multi-scale characteristics of wavelet transforms. Results obtained on the NTIRE benchmarks indicate that PMFF-SRNet achieves competitive reconstruction performance across multiple evaluation metrics, while maintaining a lightweight and computationally efficient architecture. These findings demonstrate the strong potential of PMFF-SRNet for practical hyperspectral reconstruction tasks.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"739 ","pages":"Article 123129"},"PeriodicalIF":6.8,"publicationDate":"2026-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191062","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-28DOI: 10.1016/j.ins.2026.123163
Chao Wang , Weiwei Fu , Haoyang Li , Linqi Ye , Yang Zhou
Visual Question Answering (VQA) models frequently rely on language priors while overlooking visual content. Current mainstream debiasing methods face limitations: data augmentation techniques demand high manual annotation costs and struggle to achieve balanced mitigation of biases, while ensemble-based approaches only capture language priors through a QA branch without fully identifying comprehensive bias. We propose FAIR, a bias reshaping method that utilizes pseudo-label functions to balance distribution bias and emphasizes learning weights for challenging samples. Moreover, we find that using model logit distributions as a substitute can achieve comparable effects to traditional data distribution annotations required by previous ensemble methods. Experimental results demonstrate that FAIR achieves the best balance among comparable methods, reaching 64.03% accuracy on VQA v2 and 60.96% on VQA-CP v2.
{"title":"Focal equilibrium: Bias reshaping for generalizable and robust visual understanding","authors":"Chao Wang , Weiwei Fu , Haoyang Li , Linqi Ye , Yang Zhou","doi":"10.1016/j.ins.2026.123163","DOIUrl":"10.1016/j.ins.2026.123163","url":null,"abstract":"<div><div>Visual Question Answering (VQA) models frequently rely on language priors while overlooking visual content. Current mainstream debiasing methods face limitations: data augmentation techniques demand high manual annotation costs and struggle to achieve balanced mitigation of biases, while ensemble-based approaches only capture language priors through a QA branch without fully identifying comprehensive bias. We propose <span>FAIR</span>, a bias reshaping method that utilizes pseudo-label functions to balance distribution bias and emphasizes learning weights for challenging samples. Moreover, we find that using model logit distributions as a substitute can achieve comparable effects to traditional data distribution annotations required by previous ensemble methods. Experimental results demonstrate that <span>FAIR</span> achieves the best balance among comparable methods, reaching 64.03% accuracy on VQA v2 and 60.96% on VQA-CP v2.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"739 ","pages":"Article 123163"},"PeriodicalIF":6.8,"publicationDate":"2026-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190948","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-28DOI: 10.1016/j.ins.2026.123161
Meilin Lei , Zhechen Zhu , Yingnan Pan , Yan Lei
Prescribed performance control (PPC) plays a vital role in vehicle platoon systems by ensuring their safe and stable operation, and its effectiveness is commonly limited by spacing policies, initial conditions, and input saturation. This paper investigates an improved PPC strategy under these influences within the framework of the fully actuated system approach. Firstly, an improved exponential spacing policy (IESP) incorporating the leader’s velocity information is proposed to mitigate the effects of velocity fluctuations on inter-vehicle spacing. Subsequently, a novel shifting function is designed such that the spacing error converges inside the prescribed region within the settling time, thus eliminating the dependence on the initial spacing error. The time-varying convergence boundary of the proposed performance function improves the adaptability of the system to sudden changes in the road environment. In addition, the input saturation problem is addressed using the hyperbolic tangent function. Finally, all the signals of the system are proven to be semi-globally ultimately uniformly bounded, ensuring the internal stability, string stability, and traffic flow stability. The effectiveness of the proposed strategy is verified via simulation results.
{"title":"ESP-based prescribed performance formation control for vehicle platoon systems with input saturation: A fully actuated system approach","authors":"Meilin Lei , Zhechen Zhu , Yingnan Pan , Yan Lei","doi":"10.1016/j.ins.2026.123161","DOIUrl":"10.1016/j.ins.2026.123161","url":null,"abstract":"<div><div>Prescribed performance control (PPC) plays a vital role in vehicle platoon systems by ensuring their safe and stable operation, and its effectiveness is commonly limited by spacing policies, initial conditions, and input saturation. This paper investigates an improved PPC strategy under these influences within the framework of the fully actuated system approach. Firstly, an improved exponential spacing policy (IESP) incorporating the leader’s velocity information is proposed to mitigate the effects of velocity fluctuations on inter-vehicle spacing. Subsequently, a novel shifting function is designed such that the spacing error converges inside the prescribed region within the settling time, thus eliminating the dependence on the initial spacing error. The time-varying convergence boundary of the proposed performance function improves the adaptability of the system to sudden changes in the road environment. In addition, the input saturation problem is addressed using the hyperbolic tangent function. Finally, all the signals of the system are proven to be semi-globally ultimately uniformly bounded, ensuring the internal stability, string stability, and traffic flow stability. The effectiveness of the proposed strategy is verified via simulation results.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"739 ","pages":"Article 123161"},"PeriodicalIF":6.8,"publicationDate":"2026-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191056","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-27DOI: 10.1016/j.ins.2026.123132
Yanan Song, Xiangyuan Chen, Ronghua Xu
The rapid advancement of deep learning-based generative technologies has led to remarkable achievements in deepfake applications using video and image media, particularly in areas such as face swapping and expression transfer. However, these developments have also triggered significant concerns regarding media authenticity and information security. Deepfake content often exhibits various artifacts: in the spatial domain, it may suffer from over-smoothed textures, loss of edge details, or jagged distortions; in the frequency domain, abnormal peaks in high-frequency spectra or noise-induced distortions may appear; and at the semantic level, misaligned keypoints and poor temporal coherence are frequently observed. To address these limitations, this study proposes a network architecture that first performs hybrid-domain feature extraction on deepfake samples. The Xception backbone, optimized through a knowledge distillation strategy to remove redundant layers, is combined with the lightweight MesoNet4 architecture to form a dual-branch backbone that can capture semantic features at different levels. While preserving semantic representation, the overall model size is compressed to just 8.0M parameters, achieving both high-precision detection of deepfake samples (accuracy 99%) and real-time inference performance (single-frame latency 10 ms).
{"title":"Dual-branch Meso-Xception network for hybrid-domain feature of deepfake detection","authors":"Yanan Song, Xiangyuan Chen, Ronghua Xu","doi":"10.1016/j.ins.2026.123132","DOIUrl":"10.1016/j.ins.2026.123132","url":null,"abstract":"<div><div>The rapid advancement of deep learning-based generative technologies has led to remarkable achievements in deepfake applications using video and image media, particularly in areas such as face swapping and expression transfer. However, these developments have also triggered significant concerns regarding media authenticity and information security. Deepfake content often exhibits various artifacts: in the spatial domain, it may suffer from over-smoothed textures, loss of edge details, or jagged distortions; in the frequency domain, abnormal peaks in high-frequency spectra or noise-induced distortions may appear; and at the semantic level, misaligned keypoints and poor temporal coherence are frequently observed. To address these limitations, this study proposes a network architecture that first performs hybrid-domain feature extraction on deepfake samples. The Xception backbone, optimized through a knowledge distillation strategy to remove redundant layers, is combined with the lightweight MesoNet4 architecture to form a dual-branch backbone that can capture semantic features at different levels. While preserving semantic representation, the overall model size is compressed to just 8.0M parameters, achieving both high-precision detection of deepfake samples (accuracy <span><math><mo>≥</mo></math></span> 99%) and real-time inference performance (single-frame latency <span><math><mo>≤</mo></math></span> 10 ms).</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"739 ","pages":"Article 123132"},"PeriodicalIF":6.8,"publicationDate":"2026-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146081919","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号