Neural Networks最新文献_第6页

PerCNet: Periodic complete representation for crystal graphs PerCNet：晶体图的周期性完整表示

IF 6 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neural Networks

Pub Date : 2024-10-30 DOI: 10.1016/j.neunet.2024.106841

Jiao Huang , Qianli Xing , Jinglong Ji , Bo Yang

Crystal molecules are considered as graph structures in different representation methods. A reasonable crystal representation method should capture the local and global information. However, existing methods only consider the local information of crystal molecules by modeling the bond distance and bond angle of first-order neighbors of atoms, which leads to the issue that different crystals will have the same representation. To solve this many-to-one issue, we consider the global information by further considering dihedral angles. We propose a periodic complete representation of graph modeling and a calculation algorithm for infinite extended crystal materials. A theoretical proof for the representation that satisfies the periodic completeness is provided. Based on the proposed representation, we then propose a network for predicting crystal material properties, PerCNet, with a specially designed message-passing mechanism. To our best known, we are the first work that ensures the representation corresponds one-to-one with the crystal material based on graph modeling. Extensive experiments are conducted on two large-scale real-world material benchmark datasets. The PerCNet achieves the best performance among baseline methods in terms of MAE. Our code is available at https://github.com/JiaoHuang111/PerCNet.

在不同的表示方法中，晶体分子被视为图结构。合理的晶体表示方法应该捕捉局部和全局信息。然而，现有的方法只考虑了晶体分子的局部信息，即对原子一阶相邻的键距和键角进行建模，这导致了不同晶体具有相同表示方法的问题。为了解决这个多对一的问题，我们通过进一步考虑二面角来考虑全局信息。我们为无限扩展晶体材料提出了周期性的完整图建模表示法和计算算法。我们提供了满足周期完整性的表示法的理论证明。在此基础上，我们提出了一种预测晶体材料特性的网络 PerCNet，该网络具有专门设计的消息传递机制。据我们所知，我们是第一项基于图建模确保表征与晶体材料一一对应的工作。我们在两个大规模真实材料基准数据集上进行了广泛的实验。就 MAE 而言，PerCNet 在基线方法中取得了最佳性能。我们的代码见 https://github.com/JiaoHuang111/PerCNet。

{"title":"PerCNet: Periodic complete representation for crystal graphs","authors":"Jiao Huang , Qianli Xing , Jinglong Ji , Bo Yang","doi":"10.1016/j.neunet.2024.106841","DOIUrl":"10.1016/j.neunet.2024.106841","url":null,"abstract":"<div><div>Crystal molecules are considered as graph structures in different representation methods. A reasonable crystal representation method should capture the local and global information. However, existing methods only consider the local information of crystal molecules by modeling the bond distance and bond angle of first-order neighbors of atoms, which leads to the issue that different crystals will have the same representation. To solve this many-to-one issue, we consider the global information by further considering dihedral angles. We propose a periodic complete representation of graph modeling and a calculation algorithm for infinite extended crystal materials. A theoretical proof for the representation that satisfies the periodic completeness is provided. Based on the proposed representation, we then propose a network for predicting crystal material properties, PerCNet, with a specially designed message-passing mechanism. To our best known, we are the first work that ensures the representation corresponds one-to-one with the crystal material based on graph modeling. Extensive experiments are conducted on two large-scale real-world material benchmark datasets. The PerCNet achieves the best performance among baseline methods in terms of MAE. Our code is available at <span><span>https://github.com/JiaoHuang111/PerCNet</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"181 ","pages":"Article 106841"},"PeriodicalIF":6.0,"publicationDate":"2024-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142607194","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}

引用次数: 0

Enabling deformation slack in tracking with temporally even correlation filters. 利用时间上均匀的相关滤波器实现跟踪中的变形松弛。

IF 6 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neural Networks

Pub Date : 2024-10-29 DOI: 10.1016/j.neunet.2024.106839

Yuanming Zhang, Huihui Pan, Jue Wang

Discriminative correlation filters with temporal regularization have recently attracted much attention in mobile video tracking, due to the challenges of target occlusion and background interference. However, rigidly penalizing the variability of templates between adjacent frames makes trackers lazy for target evolution, leading to inaccurate responses or even tracking failure when deformation occurs. In this paper, we address the problem of instant template learning when the target undergoes drastic variations in appearance and aspect ratio. We first propose a temporally even model featuring deformation slack, which theoretically supports the ability of the template to respond quickly to variations while suppressing disturbances. Then, an optimal derivation of our model is formulated, and the closed form solutions are deduced to facilitate the algorithm implementation. Further, we introduce a cyclic shift methodology for mirror factors to achieve scale estimation of varying aspect ratios, thereby dramatically improving the cross-area accuracy. Comprehensive comparisons on seven datasets demonstrate our excellent performance: DroneTB-70, VisDrone-SOT2019, VOT-2019, LaSOT, TC-128, UAV-20L, and UAVDT. Our approach runs at 16.9 frames per second on a low-cost Central Processing Unit, which makes it suitable for tracking on drones. The code and raw results will be made publicly available at: https://github.com/visualperceptlab/TEDS.

由于面临目标遮挡和背景干扰的挑战，带有时间正则化的判别相关滤波器最近在移动视频跟踪领域引起了广泛关注。然而，硬性惩罚相邻帧之间模板的可变性会使跟踪器懒于考虑目标的演变，导致响应不准确，甚至在发生变形时跟踪失败。在本文中，我们将解决目标在外观和长宽比发生剧烈变化时的即时模板学习问题。我们首先提出了一个具有形变松弛特征的时间均匀模型，从理论上支持模板在抑制干扰的同时快速响应变化的能力。然后，我们对模型进行了优化推导，并推导出了闭式解，以方便算法的实施。此外，我们还引入了镜像因子循环移动方法，以实现不同长宽比的比例估算，从而显著提高了横面积精度。七个数据集的综合比较证明了我们的卓越性能：这些数据集包括：DroneTB-70、VisDrone-SOT2019、VOT-2019、LaSOT、TC-128、UAV-20L 和 UAVDT。我们的方法可在低成本的中央处理器上以每秒 16.9 帧的速度运行，因此适合在无人机上进行跟踪。代码和原始结果将在以下网站公开：https://github.com/visualperceptlab/TEDS。

{"title":"Enabling deformation slack in tracking with temporally even correlation filters.","authors":"Yuanming Zhang, Huihui Pan, Jue Wang","doi":"10.1016/j.neunet.2024.106839","DOIUrl":"https://doi.org/10.1016/j.neunet.2024.106839","url":null,"abstract":"<p><p>Discriminative correlation filters with temporal regularization have recently attracted much attention in mobile video tracking, due to the challenges of target occlusion and background interference. However, rigidly penalizing the variability of templates between adjacent frames makes trackers lazy for target evolution, leading to inaccurate responses or even tracking failure when deformation occurs. In this paper, we address the problem of instant template learning when the target undergoes drastic variations in appearance and aspect ratio. We first propose a temporally even model featuring deformation slack, which theoretically supports the ability of the template to respond quickly to variations while suppressing disturbances. Then, an optimal derivation of our model is formulated, and the closed form solutions are deduced to facilitate the algorithm implementation. Further, we introduce a cyclic shift methodology for mirror factors to achieve scale estimation of varying aspect ratios, thereby dramatically improving the cross-area accuracy. Comprehensive comparisons on seven datasets demonstrate our excellent performance: DroneTB-70, VisDrone-SOT2019, VOT-2019, LaSOT, TC-128, UAV-20L, and UAVDT. Our approach runs at 16.9 frames per second on a low-cost Central Processing Unit, which makes it suitable for tracking on drones. The code and raw results will be made publicly available at: https://github.com/visualperceptlab/TEDS.</p>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"181 ","pages":"106839"},"PeriodicalIF":6.0,"publicationDate":"2024-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142607187","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}

引用次数: 0

Synchronization of time-delay dynamical networks via hybrid delayed impulses 通过混合延迟脉冲实现时延动态网络的同步化

IF 6 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neural Networks

Pub Date : 2024-10-28 DOI: 10.1016/j.neunet.2024.106835

Huannan Zheng , Wei Zhu , Xiaodi Li

This paper investigates the synchronization problem of time-delay dynamical networks by means of hybrid delayed impulses, where synchronizing impulses and desynchronizing impulses can occur simultaneously. Some sufficient synchronization conditions are established based on Razumikhin-type inequality and Lyapunov function. These conditions do not place any limitation on the magnitude of time-delay in dynamical networks. To be specific, it can be less than or greater than the length of impulses intervals and has no magnitude relationship with delays in impulses. Moreover, results indicate that delays in impulses have positive contributions to synchronization. The effectiveness of the theoretical results is demonstrated by two numerical examples.

本文通过混合延迟脉冲来研究时延动态网络的同步问题，其中同步脉冲和非同步脉冲可以同时发生。本文基于拉祖米金型不等式和 Lyapunov 函数，建立了一些充分同步条件。这些条件对动态网络中的时延大小没有任何限制。具体来说，它可以小于或大于脉冲间隔的长度，与脉冲延迟的大小没有关系。此外，研究结果表明，脉冲延迟对同步有积极作用。两个数值实例证明了理论结果的有效性。

引用次数: 0

CNN-Informer: A hybrid deep learning model for seizure detection on long-term EEG CNN-Informer：用于长期脑电图癫痫发作检测的混合深度学习模型。

IF 6 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neural Networks

Pub Date : 2024-10-28 DOI: 10.1016/j.neunet.2024.106855

Chuanyu Li , Haotian Li , Xingchen Dong , Xiangwen Zhong , Haozhou Cui , Dezan Ji , Landi He , Guoyang Liu , Weidong Zhou

Timely detecting epileptic seizures can significantly reduce accidental injuries of epilepsy patients and offer a novel intervention approach to improve their quality of life. Investigation on seizure detection based on deep learning models has achieved great success. However, there still remain challenging issues, such as the high computational complexity of the models and overfitting caused by the scarce availability of ictal electroencephalogram (EEG) signals for training. Therefore, we propose a novel end-to-end automatic seizure detection model named CNN-Informer, which leverages the capability of Convolutional Neural Network (CNN) to extract EEG local features of multi-channel EEGs, and the low computational complexity and memory usage ability of the Informer to capture the long-range dependencies. In view of the existence of various artifacts in long-term EEGs, we filter those raw EEGs using Discrete Wavelet Transform (DWT) before feeding them into the proposed CNN-Informer model for feature extraction and classification. Post-processing operations are further employed to achieve the final detection results. Our method is extensively evaluated on the CHB-MIT dataset and SH-SDU dataset with both segment-based and event-based criteria. The experimental outcomes demonstrate the superiority of the proposed CNN-Informer model and its strong generalization ability across two EEG datasets. In addition, the lightweight architecture of CNN-Informer makes it suitable for real-time implementation.

及时发现癫痫发作可以大大减少癫痫患者的意外伤害，并为改善他们的生活质量提供一种新的干预方法。基于深度学习模型的癫痫发作检测研究已取得巨大成功。然而，目前仍存在一些具有挑战性的问题，如模型的计算复杂度较高，以及用于训练的发作性脑电图（EEG）信号稀缺导致的过拟合等。因此，我们提出了一种名为 "CNN-Informer "的新型端到端癫痫发作自动检测模型，该模型利用卷积神经网络（CNN）提取多通道脑电图局部特征的能力，以及 Informer 的低计算复杂度和内存占用能力捕捉长程依赖关系。鉴于长期脑电图中存在各种假象，我们在将原始脑电图输入所提出的 CNN Informer 模型进行特征提取和分类之前，使用离散小波变换（DWT）对其进行了过滤。我们还进一步采用了后处理操作，以实现最终的检测结果。我们的方法在 CHB-MIT 数据集和 SH-SDU 数据集上进行了广泛评估，评估标准既有基于片段的标准，也有基于事件的标准。实验结果证明了所提出的 CNN-Informer 模型的优越性及其在两个脑电图数据集上的强大泛化能力。此外，CNN-Informer 的轻量级架构使其适合实时实施。

{"title":"CNN-Informer: A hybrid deep learning model for seizure detection on long-term EEG","authors":"Chuanyu Li , Haotian Li , Xingchen Dong , Xiangwen Zhong , Haozhou Cui , Dezan Ji , Landi He , Guoyang Liu , Weidong Zhou","doi":"10.1016/j.neunet.2024.106855","DOIUrl":"10.1016/j.neunet.2024.106855","url":null,"abstract":"<div><div>Timely detecting epileptic seizures can significantly reduce accidental injuries of epilepsy patients and offer a novel intervention approach to improve their quality of life. Investigation on seizure detection based on deep learning models has achieved great success. However, there still remain challenging issues, such as the high computational complexity of the models and overfitting caused by the scarce availability of ictal electroencephalogram (EEG) signals for training. Therefore, we propose a novel end-to-end automatic seizure detection model named CNN-Informer, which leverages the capability of Convolutional Neural Network (CNN) to extract EEG local features of multi-channel EEGs, and the low computational complexity and memory usage ability of the Informer to capture the long-range dependencies. In view of the existence of various artifacts in long-term EEGs, we filter those raw EEGs using Discrete Wavelet Transform (DWT) before feeding them into the proposed CNN-Informer model for feature extraction and classification. Post-processing operations are further employed to achieve the final detection results. Our method is extensively evaluated on the CHB-MIT dataset and SH-SDU dataset with both segment-based and event-based criteria. The experimental outcomes demonstrate the superiority of the proposed CNN-Informer model and its strong generalization ability across two EEG datasets. In addition, the lightweight architecture of CNN-Informer makes it suitable for real-time implementation.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"181 ","pages":"Article 106855"},"PeriodicalIF":6.0,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142565350","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}

引用次数: 0

Highly valued subgoal generation for efficient goal-conditioned reinforcement learning 为高效目标条件强化学习生成高价值子目标

IF 6 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neural Networks

Pub Date : 2024-10-28 DOI: 10.1016/j.neunet.2024.106825

Yao Li , YuHui Wang , XiaoYang Tan

Goal-conditioned reinforcement learning is widely used in robot control, manipulating the robot to accomplish specific tasks by maximizing accumulated rewards. However, the useful reward signal is only received when the desired goal is reached, leading to the issue of sparse rewards and affecting the efficiency of policy learning. In this paper, we propose a method to generate highly valued subgoals for efficient goal-conditioned policy learning, enabling the development of smart home robots or automatic pilots in our daily life. The highly valued subgoals are conditioned on the context of the specific tasks and characterized by suitable complexity for efficient goal-conditioned action value learning. The context variable captures the latent representation of the particular tasks, allowing for efficient subgoal generation. Additionally, the goal-conditioned action values regularized by the self-adaptive ranges generate subgoals with suitable complexity. Compared to Hindsight Experience Replay that uniformly samples subgoals from visited trajectories, our method generates the subgoals based on the context of tasks with suitable difficulty for efficient policy training. Experimental results show that our method achieves stable performance in robotic environments compared to baseline methods.

目标条件强化学习广泛应用于机器人控制，通过最大化累积奖励来操纵机器人完成特定任务。然而，只有在达到预期目标时才会收到有用的奖励信号，这就导致了奖励稀疏的问题，影响了策略学习的效率。在本文中，我们提出了一种生成高价值子目标的方法，以实现高效的目标条件策略学习，从而开发出智能家居机器人或日常生活中的自动驾驶员。高价值子目标以特定任务的上下文为条件，具有适合目标条件行动值高效学习的复杂性。上下文变量捕捉了特定任务的潜在表征，从而可以高效地生成子目标。此外，由自适应范围规范化的目标条件行动值可生成具有适当复杂度的子目标。与从访问过的轨迹中均匀采样子目标的 "后见之明经验重放 "相比，我们的方法是根据任务的上下文生成子目标，具有适当的难度，从而实现高效的策略训练。实验结果表明，与基线方法相比，我们的方法在机器人环境中实现了稳定的性能。

{"title":"Highly valued subgoal generation for efficient goal-conditioned reinforcement learning","authors":"Yao Li , YuHui Wang , XiaoYang Tan","doi":"10.1016/j.neunet.2024.106825","DOIUrl":"10.1016/j.neunet.2024.106825","url":null,"abstract":"<div><div>Goal-conditioned reinforcement learning is widely used in robot control, manipulating the robot to accomplish specific tasks by maximizing accumulated rewards. However, the useful reward signal is only received when the desired goal is reached, leading to the issue of sparse rewards and affecting the efficiency of policy learning. In this paper, we propose a method to generate highly valued subgoals for efficient goal-conditioned policy learning, enabling the development of smart home robots or automatic pilots in our daily life. The highly valued subgoals are conditioned on the context of the specific tasks and characterized by suitable complexity for efficient goal-conditioned action value learning. The context variable captures the latent representation of the particular tasks, allowing for efficient subgoal generation. Additionally, the goal-conditioned action values regularized by the self-adaptive ranges generate subgoals with suitable complexity. Compared to Hindsight Experience Replay that uniformly samples subgoals from visited trajectories, our method generates the subgoals based on the context of tasks with suitable difficulty for efficient policy training. Experimental results show that our method achieves stable performance in robotic environments compared to baseline methods.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"181 ","pages":"Article 106825"},"PeriodicalIF":6.0,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142565351","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}

引用次数: 0

Adaptive window adjustment with boundary DoU loss for cascade segmentation of anatomy and lesions in prostate cancer using bpMRI 利用 bpMRI 对前列腺癌的解剖结构和病灶进行级联分割时，采用边界 DoU 损失进行自适应窗口调整

IF 6 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neural Networks

Pub Date : 2024-10-28 DOI: 10.1016/j.neunet.2024.106831

Wenhao Li , Bowen Zheng , Quanyou Shen , Xiaoran Shi , Kun Luo , Yuqian Yao , Xinyan Li , Shidong Lv , Jie Tao , Qiang Wei

Accurate segmentation of prostate anatomy and lesions using biparametric magnetic resonance imaging (bpMRI) is crucial for the diagnosis and treatment of prostate cancer with the aid of artificial intelligence. In prostate bpMRI, different tissues and pathologies are best visualized within specific and narrow ranges for each sequence, which have varying requirements for image window settings. Currently, adjustments to window settings rely on experience, lacking an efficient method for universal automated adjustment. Hence, we propose an Adaptive Window Adjustment (AWA) module capable of adjusting window settings to accommodate different image modalities, sample data, and downstream tasks. Moreover, given the pivotal role that loss functions play in optimizing model performance, we investigate the performance of different loss functions in segmenting prostate anatomy and lesions. Our study validates the superiority of the Boundary Difference over Union (DoU) Loss in these tasks and extends its applicability to 3D medical imaging. Finally, we propose a cascaded segmentation approach tailored for prostate anatomy and lesions. This approach leverages anatomical structure information to enhance lesion segmentation accuracy. Experimental results on the Prostate158, ProstateX, and PI-CAI datasets confirm the effectiveness of the proposed methods. Our code of methods is available at https://github.com/WenHao-L/AWA_BoundaryDoULoss.

利用双参数磁共振成像（bpMRI）对前列腺解剖结构和病变进行准确分割，对于借助人工智能诊断和治疗前列腺癌至关重要。在前列腺双参数磁共振成像（bpMRI）中，不同的组织和病变在每个序列的特定和狭窄范围内可视化效果最佳，这对图像窗口设置有不同的要求。目前，窗口设置的调整主要依靠经验，缺乏通用自动调整的有效方法。因此，我们提出了一种自适应窗口调整（AWA）模块，能够调整窗口设置以适应不同的图像模式、样本数据和下游任务。此外，鉴于损失函数在优化模型性能方面的关键作用，我们研究了不同损失函数在分割前列腺解剖结构和病变时的性能。我们的研究验证了边界差分损失函数（Boundary Difference over Union，DoU）在这些任务中的优越性，并将其应用扩展到三维医学成像中。最后，我们提出了一种针对前列腺解剖结构和病变的级联分割方法。这种方法利用解剖结构信息来提高病变分割的准确性。在 Prostate158、ProstateX 和 PI-CAI 数据集上的实验结果证实了所提方法的有效性。我们的方法代码见 https://github.com/WenHao-L/AWA_BoundaryDoULoss。

{"title":"Adaptive window adjustment with boundary DoU loss for cascade segmentation of anatomy and lesions in prostate cancer using bpMRI","authors":"Wenhao Li , Bowen Zheng , Quanyou Shen , Xiaoran Shi , Kun Luo , Yuqian Yao , Xinyan Li , Shidong Lv , Jie Tao , Qiang Wei","doi":"10.1016/j.neunet.2024.106831","DOIUrl":"10.1016/j.neunet.2024.106831","url":null,"abstract":"<div><div>Accurate segmentation of prostate anatomy and lesions using biparametric magnetic resonance imaging (bpMRI) is crucial for the diagnosis and treatment of prostate cancer with the aid of artificial intelligence. In prostate bpMRI, different tissues and pathologies are best visualized within specific and narrow ranges for each sequence, which have varying requirements for image window settings. Currently, adjustments to window settings rely on experience, lacking an efficient method for universal automated adjustment. Hence, we propose an Adaptive Window Adjustment (AWA) module capable of adjusting window settings to accommodate different image modalities, sample data, and downstream tasks. Moreover, given the pivotal role that loss functions play in optimizing model performance, we investigate the performance of different loss functions in segmenting prostate anatomy and lesions. Our study validates the superiority of the Boundary Difference over Union (DoU) Loss in these tasks and extends its applicability to 3D medical imaging. Finally, we propose a cascaded segmentation approach tailored for prostate anatomy and lesions. This approach leverages anatomical structure information to enhance lesion segmentation accuracy. Experimental results on the Prostate158, ProstateX, and PI-CAI datasets confirm the effectiveness of the proposed methods. Our code of methods is available at <span><span>https://github.com/WenHao-L/AWA_BoundaryDoULoss</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"181 ","pages":"Article 106831"},"PeriodicalIF":6.0,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142551929","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}

引用次数: 0

SeBIR: Semantic-guided burst image restoration SeBIR：语义引导的突发图像修复

IF 6 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neural Networks

Pub Date : 2024-10-26 DOI: 10.1016/j.neunet.2024.106834

Huan Liu, Mingwen Shao, Yecong Wan, Yuexian Liu, Kai Shang

Burst image restoration methods offer the possibility of recovering faithful scene details from multiple low-quality snapshots captured by hand-held devices in adverse scenarios, thereby attracting increasing attention in recent years. However, individual frames in a burst typically suffer from inter-frame misalignments, leading to ghosting artifacts. Besides, existing methods indiscriminately handle all burst frames, struggling to seamlessly remove the corrupted information due to the neglect of multi-frame spatio-temporal varying degradation. To alleviate these limitations, we propose a general semantic-guided model named SeBIR for burst image restoration incorporating the semantic prior knowledge of Segment Anything Model (SAM) to enable adaptive recovery. Specifically, instead of relying solely on a single aligning scheme, we develop a joint implicit and explicit strategy that sufficiently leverages semantic knowledge as guidance to achieve inter-frame alignment. To further adaptively modulate and aggregate aligned features with spatio-temporal disparity, we elaborate a semantic-guided fusion module using the intermediate semantic features of SAM as an explicit guide to weaken the inherent degradation and strengthen the valuable complementary information across frames. Additionally, a semantic-guided local loss is designed to boost local consistency and image quality. Extensive experiments on synthetic and real-world datasets demonstrate the superiority of our method in both quantitative and qualitative evaluations for burst super-resolution, burst denoising, and burst low-light image enhancement tasks.

连拍图像复原方法可以从手持设备在不利场景下捕获的多张低质量快照中恢复忠实的场景细节，因此近年来受到越来越多的关注。然而，连拍中的单个帧通常会出现帧间错位，从而导致重影伪影。此外，现有方法不加区分地处理所有突发帧，由于忽略了多帧时空变化退化，难以无缝去除损坏的信息。为了缓解这些局限性，我们提出了一种名为 SeBIR 的通用语义引导模型，该模型用于突发图像修复，并结合了 "分段任意模型"（Segment Anything Model，SAM）的语义先验知识，从而实现自适应恢复。具体来说，我们不再仅仅依赖单一的对齐方案，而是开发了一种隐式和显式联合策略，充分利用语义知识作为实现帧间对齐的指导。为了进一步自适应地调节和聚合具有时空差异的对齐特征，我们精心设计了一个语义指导的融合模块，使用 SAM 的中间语义特征作为显式指导，以弱化固有的劣化，并加强各帧之间有价值的互补信息。此外，还设计了语义引导的局部损失，以提高局部一致性和图像质量。在合成数据集和真实数据集上进行的大量实验证明了我们的方法在突发超分辨率、突发去噪和突发低照度图像增强任务的定量和定性评估中的优越性。

{"title":"SeBIR: Semantic-guided burst image restoration","authors":"Huan Liu, Mingwen Shao, Yecong Wan, Yuexian Liu, Kai Shang","doi":"10.1016/j.neunet.2024.106834","DOIUrl":"10.1016/j.neunet.2024.106834","url":null,"abstract":"<div><div>Burst image restoration methods offer the possibility of recovering faithful scene details from multiple low-quality snapshots captured by hand-held devices in adverse scenarios, thereby attracting increasing attention in recent years. However, individual frames in a burst typically suffer from inter-frame misalignments, leading to ghosting artifacts. Besides, existing methods indiscriminately handle all burst frames, struggling to seamlessly remove the corrupted information due to the neglect of multi-frame spatio-temporal varying degradation. To alleviate these limitations, we propose a general semantic-guided model named <strong>SeBIR</strong> for burst image restoration incorporating the semantic prior knowledge of Segment Anything Model (SAM) to enable adaptive recovery. Specifically, instead of relying solely on a single aligning scheme, we develop a joint implicit and explicit strategy that sufficiently leverages semantic knowledge as guidance to achieve inter-frame alignment. To further adaptively modulate and aggregate aligned features with spatio-temporal disparity, we elaborate a semantic-guided fusion module using the intermediate semantic features of SAM as an explicit guide to weaken the inherent degradation and strengthen the valuable complementary information across frames. Additionally, a semantic-guided local loss is designed to boost local consistency and image quality. Extensive experiments on synthetic and real-world datasets demonstrate the superiority of our method in both quantitative and qualitative evaluations for burst super-resolution, burst denoising, and burst low-light image enhancement tasks.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"181 ","pages":"Article 106834"},"PeriodicalIF":6.0,"publicationDate":"2024-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142551931","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}

引用次数: 0

Node classification in the heterophilic regime via diffusion-jump GNNs 通过扩散跳跃 GNN 实现异嗜系统中的节点分类

IF 6 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neural Networks

Pub Date : 2024-10-26 DOI: 10.1016/j.neunet.2024.106830

Ahmed Begga, Francisco Escolano, Miguel Ángel Lozano

In the ideal (homophilic) regime of vanilla GNNs, nodes belonging to the same community have the same label: most of the nodes are harmonic (their unknown labels result from averaging those of their neighbors given some labeled nodes). In other words, heterophily (when neighboring nodes have different labels) can be seen as a “loss of harmonicity”.

In this paper, we define “structural heterophily” in terms of the ratio between the harmonicity of the network (Laplacian Dirichlet energy) and the harmonicity of its homophilic version (the so-called “ground” energy). This new measure inspires a novel GNN model (Diffusion-Jump GNN) that bypasses structural heterophily by “jumping” through the network in order to relate distant homologs. However, instead of using hops as standard High-Order (HO) GNNs (MixHop) do, our jumps are rooted in a structural well-known metric: the diffusion distance.

Computing the “diffusion matrix” (DM) is the core of this method. Our main contribution is that we learn both the diffusion distances and the “structural filters” derived from them. Since diffusion distances have a spectral interpretation, we learn orthogonal approximations of the Laplacian eigenvectors while the prediction loss is minimized. This leads to an interplay between a Dirichlet loss, which captures low-frequency content, and a prediction loss which refines that content leading to empirical eigenfunctions. Finally, our experimental results show that we are very competitive with the State-Of-the-Art (SOTA) both in homophilic and heterophilic datasets, even in large graphs.

在虚构 GNN 的理想（嗜同）状态下，属于同一社区的节点具有相同的标签：大多数节点都是谐调的（它们的未知标签是在给定了一些标签节点的情况下，对其邻居的未知标签进行平均的结果）。换句话说，异质性（当相邻节点具有不同标签时）可被视为 "和谐性的损失"。在本文中，我们用网络的谐调性（拉普拉斯 Dirichlet 能量）与其同亲版本的谐调性（所谓的 "地面 "能量）之间的比率来定义 "结构异亲性"。这种新的测量方法启发了一种新颖的 GNN 模型（扩散-跳跃 GNN），它通过在网络中 "跳跃 "来绕过结构异质性，从而将遥远的同源物联系起来。不过，我们的跳跃不是像标准的高阶（HO）GNN（MixHop）那样使用跳数，而是以众所周知的结构度量为基础：扩散距离。计算 "扩散矩阵"（DM）是这一方法的核心。我们的主要贡献在于，我们既学习了扩散距离，也学习了由扩散距离衍生出的 "结构过滤器"。由于扩散距离具有频谱解释，因此我们可以学习拉普拉卡特征向量的正交近似值，同时使预测损失最小化。这就导致了捕捉低频内容的 Dirichlet 损失和细化该内容的预测损失之间的相互作用，从而产生了经验特征函数。最后，我们的实验结果表明，无论是在同亲数据集还是异亲数据集中，即使是在大型图中，我们都能与最新技术（SOTA）相媲美。

{"title":"Node classification in the heterophilic regime via diffusion-jump GNNs","authors":"Ahmed Begga, Francisco Escolano, Miguel Ángel Lozano","doi":"10.1016/j.neunet.2024.106830","DOIUrl":"10.1016/j.neunet.2024.106830","url":null,"abstract":"<div><div>In the ideal (homophilic) regime of vanilla GNNs, nodes belonging to the same community have the same label: most of the nodes are harmonic (their unknown labels result from averaging those of their neighbors given some labeled nodes). In other words, heterophily (when neighboring nodes have different labels) can be seen as a “loss of harmonicity”.</div><div>In this paper, we define “structural heterophily” in terms of the ratio between the harmonicity of the network (Laplacian Dirichlet energy) and the harmonicity of its homophilic version (the so-called “ground” energy). This new measure inspires a novel GNN model (Diffusion-Jump GNN) that bypasses structural heterophily by “jumping” through the network in order to relate distant homologs. However, instead of using hops as standard High-Order (HO) GNNs (MixHop) do, our jumps are rooted in a structural well-known metric: the diffusion distance.</div><div>Computing the “diffusion matrix” (DM) is the core of this method. Our main contribution is that we learn both the diffusion distances and the “structural filters” derived from them. Since diffusion distances have a spectral interpretation, we learn orthogonal approximations of the Laplacian eigenvectors while the prediction loss is minimized. This leads to an interplay between a Dirichlet loss, which captures low-frequency content, and a prediction loss which refines that content leading to empirical eigenfunctions. Finally, our experimental results show that we are very competitive with the State-Of-the-Art (SOTA) both in homophilic and heterophilic datasets, even in large graphs.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"181 ","pages":"Article 106830"},"PeriodicalIF":6.0,"publicationDate":"2024-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142631414","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Retrieval In Decoder benefits generative models for explainable complex question answering Retrieval In Decoder 有利于可解释复杂问题解答的生成模型。

IF 6 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neural Networks

Pub Date : 2024-10-25 DOI: 10.1016/j.neunet.2024.106833

Jianzhou Feng , Qin Wang , Huaxiao Qiu , Lirong Liu

Large-scale Language Models (LLMs) utilizing the Chain-of-Thought prompting demonstrate exceptional performance in a variety of tasks. However, the persistence of factual hallucinations remains a significant challenge in practical applications. Prevailing retrieval-augmented methods treat the retriever and generator as separate components, which inadvertently restricts the generator’s capabilities to those of the retriever through intensive supervised training. In this work, we propose an unsupervised Retrieval In Decoder framework for multi-granularity decoding called RID, which integrates retrieval directly into the decoding process of generative models. It dynamically adjusts decoding granularity based on retrieval outcomes, and duly corrects the decoding direction through its direct impact on the next token. Moreover, we introduce a reinforcement learning-driven knowledge distillation method for adaptive explanation generation to better apply to Small-scale Language Models (SLMs). The experimental results across six public benchmarks surpass popular LLMs and existing retrieval-augmented methods, which demonstrates the effectiveness of RID in models of different scales and verifies its applicability and scalability.

利用思维链提示的大规模语言模型（LLM）在各种任务中表现出了卓越的性能。然而，在实际应用中，事实幻觉的持续性仍然是一个重大挑战。现行的检索增强方法将检索器和生成器视为独立的组件，这无意中通过强化监督训练将生成器的能力限制在检索器的能力范围内。在这项工作中，我们提出了一种用于多粒度解码的无监督检索解码器框架，称为 RID，它将检索直接集成到生成模型的解码过程中。它能根据检索结果动态调整解码粒度，并通过对下一个标记的直接影响来适当修正解码方向。此外，我们还为自适应解释生成引入了强化学习驱动的知识提炼方法，以便更好地应用于小型语言模型（SLM）。在六个公开基准测试中的实验结果超过了流行的 LLM 和现有的检索增强方法，这证明了 RID 在不同规模模型中的有效性，并验证了其适用性和可扩展性。

{"title":"Retrieval In Decoder benefits generative models for explainable complex question answering","authors":"Jianzhou Feng , Qin Wang , Huaxiao Qiu , Lirong Liu","doi":"10.1016/j.neunet.2024.106833","DOIUrl":"10.1016/j.neunet.2024.106833","url":null,"abstract":"<div><div>Large-scale Language Models (LLMs) utilizing the Chain-of-Thought prompting demonstrate exceptional performance in a variety of tasks. However, the persistence of factual hallucinations remains a significant challenge in practical applications. Prevailing retrieval-augmented methods treat the retriever and generator as separate components, which inadvertently restricts the generator’s capabilities to those of the retriever through intensive supervised training. In this work, we propose an unsupervised Retrieval In Decoder framework for multi-granularity decoding called <em>RID</em>, which integrates retrieval directly into the decoding process of generative models. It dynamically adjusts decoding granularity based on retrieval outcomes, and duly corrects the decoding direction through its direct impact on the next token. Moreover, we introduce a reinforcement learning-driven knowledge distillation method for adaptive explanation generation to better apply to Small-scale Language Models (SLMs). The experimental results across six public benchmarks surpass popular LLMs and existing retrieval-augmented methods, which demonstrates the effectiveness of RID in models of different scales and verifies its applicability and scalability.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"181 ","pages":"Article 106833"},"PeriodicalIF":6.0,"publicationDate":"2024-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142607200","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}

引用次数: 0

Adaptive multi-graph contrastive learning for bundle recommendation 用于捆绑推荐的自适应多图对比学习。

IF 6 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neural Networks

Pub Date : 2024-10-24 DOI: 10.1016/j.neunet.2024.106832

Qian Tao , Chenghao Liu , Yuhan Xia , Yong Xu , Lusi Li

Recently, recommending bundles - sets of items that complement each other - instead of individual items to users has drawn much attention in both academia and industry. Models based on Graph Neural Networks (GNNs) for bundle recommendation have achieved great success in capturing users’ preferences by modeling pairwise correlations among users, bundles, and items via information propagation on graphs. However, a notable limitation lies in their insufficient focus on explicitly modeling intricate ternary relationships. Additionally, the loose combination of node embeddings from different graphs tends to introduce noise, as it fails to consider disparities among the graphs. To this end, we propose a novel approach called Adaptive Multi-Graph Contrastive Learning for Bundle Recommendation (AMCBR). Specifically, AMCBR models ternary interactions by constructing multiple graphs, including a bundle preference graph based on direct user-bundle interactions, a collaborative neighborhoods graph featuring user-level and bundle-level subgraphs, and an item-level preference hypergraph capturing indirect user-bundle relationships through items. Then, (hyper)graph convolution is applied to each (hyper)graph to encode diverse potential preferences into node embeddings. To enhance the model’s robustness, an adaptive aggregation module is employed to assign varying weights to node embeddings from different graphs during the fusion process, which enriches the semantic and comprehensive information in the embeddings while mitigating potential noise. Finally, a contrastive learning strategy is proposed to jointly optimize the model, strengthening collaborative links between individual graphs. Extensive experiments on three real datasets demonstrate that AMCBR can outperform the state-of-the-art baselines on the Top-K recommendations.

最近，向用户推荐捆绑商品（相互补充的商品集）而不是单个商品的做法引起了学术界和产业界的广泛关注。基于图神经网络（GNN）的捆绑推荐模型通过图上的信息传播对用户、捆绑商品和商品之间的成对相关性进行建模，在捕捉用户偏好方面取得了巨大成功。然而，其明显的局限性在于没有充分重视对复杂的三元关系进行明确建模。此外，不同图中节点嵌入的松散组合往往会引入噪音，因为它没有考虑到图之间的差异。为此，我们提出了一种名为 "用于捆绑推荐的自适应多图对比学习"（AMCBR）的新方法。具体来说，AMCBR 通过构建多个图（包括基于用户与捆绑包直接交互的捆绑包偏好图、以用户级和捆绑包级子图为特征的协作邻域图，以及通过项目捕捉用户与捆绑包间接关系的项目级偏好超图）来对三元交互进行建模。然后，对每个（超）图进行（超）图卷积，将各种潜在偏好编码为节点嵌入。为了增强模型的鲁棒性，在融合过程中采用了自适应聚合模块，为来自不同图的节点嵌入分配不同的权重，从而丰富了嵌入中的语义和综合信息，同时减少了潜在的噪音。最后，还提出了一种对比学习策略来共同优化模型，加强单个图之间的协作联系。在三个真实数据集上进行的广泛实验表明，AMCBR 在 Top-K 推荐方面的表现优于最先进的基线。

{"title":"Adaptive multi-graph contrastive learning for bundle recommendation","authors":"Qian Tao , Chenghao Liu , Yuhan Xia , Yong Xu , Lusi Li","doi":"10.1016/j.neunet.2024.106832","DOIUrl":"10.1016/j.neunet.2024.106832","url":null,"abstract":"<div><div>Recently, recommending bundles - sets of items that complement each other - instead of individual items to users has drawn much attention in both academia and industry. Models based on Graph Neural Networks (GNNs) for bundle recommendation have achieved great success in capturing users’ preferences by modeling pairwise correlations among users, bundles, and items via information propagation on graphs. However, a notable limitation lies in their insufficient focus on explicitly modeling intricate ternary relationships. Additionally, the loose combination of node embeddings from different graphs tends to introduce noise, as it fails to consider disparities among the graphs. To this end, we propose a novel approach called Adaptive Multi-Graph Contrastive Learning for Bundle Recommendation (AMCBR). Specifically, AMCBR models ternary interactions by constructing multiple graphs, including a bundle preference graph based on direct user-bundle interactions, a collaborative neighborhoods graph featuring user-level and bundle-level subgraphs, and an item-level preference hypergraph capturing indirect user-bundle relationships through items. Then, (hyper)graph convolution is applied to each (hyper)graph to encode diverse potential preferences into node embeddings. To enhance the model’s robustness, an adaptive aggregation module is employed to assign varying weights to node embeddings from different graphs during the fusion process, which enriches the semantic and comprehensive information in the embeddings while mitigating potential noise. Finally, a contrastive learning strategy is proposed to jointly optimize the model, strengthening collaborative links between individual graphs. Extensive experiments on three real datasets demonstrate that AMCBR can outperform the state-of-the-art baselines on the Top-K recommendations.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"181 ","pages":"Article 106832"},"PeriodicalIF":6.0,"publicationDate":"2024-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142607181","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}

引用次数: 0