The article considers the impulsive synchronization for inertial neural networks with unbounded delay and actuator saturation via sampled-data control. Based on an differential inequality, difficulties caused by effect may be effectively avoid. By applying polytopic representation technique, term is first considered into design of controller, less conservative linear matrix inequality (LMI) criteria that guarantee asymptotical model hybrid control are given. As special cases, saturating also...

The goal of cross-view image matching based on geo-localization is to determine the location a given ground-view (front view) by it with group satellite-view images (vertical geographic tags. Due rapid development unmanned aerial vehicle (UAV) technology in recent years, has provided real viewpoint close 45 degrees (oblique bridge visual gap between views. However, existing methods ignore direct geometric space correspondence UAV-satellite views, and only use brute force for feature...

This article investigates the event-triggered synchronization of delayed neural networks (NNs). A novel integral-based scheme (IETS) is proposed where integral system states, and past triggered data over a period time are used. With IETS, problem becomes distributed delay problem. Using Bessel-Legendre inequalities, sufficient conditions for existence controller that ensures asymptotic provided in form linear matrix inequalities (LMIs). Illustrative examples used to demonstrate advantages...

Fully mining visual cues to aid in content understanding is crucial for video captioning. However, most state-of-the-art captioning methods are limited generating captions purely based on straightforward information while ignoring the scenario and context information. To fill gap, we propose a novel, simple but effective scenario-aware recurrent transformer (SART) model execute Our contains “scenario understanding” module obtain global perspective across multiple frames, providing specific...

An observer-based dissipativity control for Takagi–Sugeno (T–S) fuzzy neural networks with distributed time-varying delays is studied in this article. First, the network channel are modeled as a delay its kernel. To make full use of kernels delay, Lyapunov–Krasovskii functional (LKF) established kernel delay. It noted that novel LKF and delay-dependent reciprocally convex inequality plays an important role dealing global asymptotical stability strict <inline-formula...

The impulsive stabilization of nonlinear time-delay system with input saturation via delay-dependent polytopic approach is studied in this article. Different from representation technique, technique able to estimate a larger domain attraction. Based on approach, the actuator term first introduced into design controller, which expressed as convex combination product state vectors and auxiliary matrices. By applying Lyapunov–Krasovskii functional (LKF) new series less conservative linear...

Training generative adversarial networks (GANs) for noise-to-image synthesis is a challenge task, primarily due to the instability of GANs’ training process. One key issues generator’s sensitivity input data, which can cause sudden fluctuations in loss value with certain inputs. This suggests an inadequate ability resist disturbances generator, causing discriminator’s oscillate and negatively impacting discriminator. Then, negative feedback discriminator also not conducive updating...

User preferences behind users' decision-making processes are highly diverse and may range from lower-level concepts with more specific intentions higher-level general intentions. In this case, user tend to be expressed hierarchically. However, learning such different levels behaviors is challenging, remains largely neglected by the existing literature. Meanwhile, behavior data tends sparse because of limited response vast combinations users items, which results in cold-start problems unclear...

Multi-sensory data has exhibited a clear advantage in expressing richer and more complex feelings, on the Emotion Recognition Conversation (ERC) task. Yet, current methods for multimodal dynamics that aggregate modalities or employ additional modality-specific modality-shared networks are still inadequate balancing between sufficiency of processing scalability to incremental multi-sensory type additions. This incurs bottleneck performance improvement ERC. To this end, we present MetaDrop,...

Approximating the uncertainty of an emotional AI agent is crucial for improving reliability such agents and facilitating human-in-the-loop solutions, especially in critical scenarios. However, none existing systems emotion recognition conversation (ERC) has attempted to estimate their predictions. In this article, we present HU-Dialogue, which models hierarchical ERC task. We perturb contextual attention weight values with source-adaptive noises within each modality, as a regularization...

Person re-identification is an important task in video surveillance, focusing on finding the same person across different cameras. However, most existing methods of video-based still have some limitations (e.g., lack effective deep learning framework, robustness model, and treatment for all frames) which make them unable to achieve better recognition performance. In this paper, we propose a novel self-paced algorithm re-identification, could gradually learn from simple complex samples mature...

The goal of cross-view geo-localization is to determine the location a given ground image by matching with aerial images. However, existing methods ignore variability scenes, additional information and spatial correspondence covisibility non-convisibility areas in ground-aerial pairs. In this context, we propose method called SMDT alignment Transformer. First, utilize semantic segmentation technique segment different areas. Then, convert vertical view images front mixing polar mapping...