检索结果-内蒙古大学图书馆

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Liu, Meng Lee, Kong Aik Wang, Longbiao Zhang, Hanyi Zeng, Chang Dang, Jianwu Tianjin University College of Intelligence and Computing Tianjin Key Laboratory of Cognitive Computing and Application Tianjin China Astar Institute for Infocomm Research Singapore Singapore Institute of Technology Singapore National Institute of Informatics Tokyo Japan

ISBN: (纸本)9781728163277

Visual speech (i.e., lip motion) is highly related to auditory speech due to the co-occurrence and synchronization in speech production. This paper investigates this correlation and proposes a cross-modal speech co-learning paradigm. The primary motivation of our cross-modal co-learning method is modeling one modality aided by exploiting knowledge from another modality. Specifically, two cross-modal boosters are introduced based on an audio-visual pseudo-siamese structure to learn the modality-transformed correlation. Inside each booster, a max-feature-map embedded Transformer variant is proposed for modality alignment and enhanced feature generation. The network is co-learned both from scratch and with pretrained models. Experimental results on the test scenarios demonstrate that our proposed method achieves around 60% and 20% average relative performance improvement over baseline unimodal and fusion systems, respectively. © 2023 IEEE.

关键词： co-learning cross-modal lip biometrics speaker verification visual speech

来源：评论

学校读者我要写书评

暂无评论

Dual Selection Network for Video Object Detection

Dual Selection Network for Video Object Detection

引用

2022 IEEE International Conference on Multimedia and Expo, ICME 2022

作者： Hou, Tianxiang Qi, Qiang Lu, Yang Du, Kaiwen Wang, Hanzi School of Informatics Xiamen University Fujian Key Laboratory of Sensing and Computing for Smart City Xiamen China

ISBN: (数字)9781665485630

ISBN: (纸本)9781665485630

Some off-the-shelf video object detection methods usually enhance the degraded proposal features of target frames by aggregating the proposal features from support frames. However, the proposals generated by region proposal network may not be accurate, resulting in inaccurate proposal features and limited performance. To mitigate this, we propose a novel dual selection network (DSNet) for video object detection, which contains two successive stages: selecting proposals that fit objects more closely, and selecting proposal features that are more conducive to feature aggregation. Correspondingly, the proposal selection module (PSM) aims to select better proposals by exploiting their boundary information, and the selective aggregation module (SAM) aims to select better proposal features for aggregation. Consequently, DSNet can generate more robust proposal features through the novel dual selection mechanism implemented by PSM and SAM. Extensive experiments show that our DSNet obtains 83.7% mAP and achieves superior performance over several state-of-the-art methods. © 2022 IEEE.

关键词： Object recognition

来源：评论

学校读者我要写书评

暂无评论

Wildfire Detection and Burned Area Estimation Based on Multi-source Spatial Data 19

Wildfire Detection and Burned Area Estimation Based on Multi...

引用

2022 IEEE SmartWorld, 19th IEEE International Conference on Ubiquitous Intelligence and computing, 2022 IEEE International Conference on Autonomous and Trusted Vehicles Conference, 22nd IEEE International Conference on Scalable computing and Communications, 2022 IEEE International Conference on Digital Twin, 8th IEEE International Conference on Privacy computing and 2022 IEEE International Conference on Metaverse, SmartWorld/UIC/ATC/ScalCom/DigitalTwin/PriComp/Metaverse 2022

作者： Weng, Lijuan Luo, Ruixiang Huang, Menghan Wang, Cheng Chen, Longbiao Xiamen University Fujian Key Laboratory of Sensing and Computing for Smart City School of Informatics Xiamen China

ISBN: (纸本)9798350346558

Forest is an important part of the global ecosystem and is of great significance to the sustainable development of the ecological environment. As the rising temperatures of the earth, forests become increasingly dry and wildfires occur more frequently, which brings great losses to global forest resources and the safety of individuals' lives. Therefore, wildfire detection is of great significance for environment protection. Recently, thermal infrared remote sensing shows the advantages of wide coverage and long observation time, which provides an important technology for the identification of large-scale and all-weather wildfires. In this paper, we use remote sensing data to detect wildfires by leveraging deep learning algorithm and estimate burned areas. Specifically, We first investigate the characteristics of wildfires in thermal infrared remote sensing imagery and make a dataset, and then train a Faster R-CNN model to detect hotspots. Next, we filter out nonforest areas to discover wildfire by fusing the global forest cover map. Finally, we estimate the burned area through the changes of the Normalized Difference Vegetation Index (NDVI) and Normalized Burn Ratio (NBR) of remote sensing images before and after wildfire. Results show that our method can accurately detect wildfires with F1-Score of 0.881 and error of the burned area estimation is 1.18%, and it is certified on social media. © 2022 IEEE.

关键词： Remote sensing

来源：评论

学校读者我要写书评

暂无评论

Viewing Pattern Assisted Proactive Partial Caching for 360-degree Videos in MEC Networks

引用

IEEE Internet of Things Journal 2025年第12期12卷 19280-19293页

作者： Yin, Guoxiao Zhou, Xiaotian Zhang, Haixia Li, Dongyang Yuan, Dongfeng Shandong University School of Control Science and Engineering Jinan250061 China Shandong University Shandong Key Laboratory of Intelligent Communication and Sensing-Computing Integration Shandong Jinan China College of Oceanography and Space Informatics Qingdao266580 China Shandong University School of Qilu Transportation Jinan250002 China

Caching 360-degree videos at the network edge can reduce user content request latency and mitigate transmission congestion in backbone networks. Given the fact that user only views a part of content of 360-degree scope at any time, caching the entire video is resource inefficient. To address this, we focus a Multi-access Edge computing (MEC)-based 360-degree video service system, where the edge server only caches a portion of each video that is most likely falling in the Field of View (FoV) of users. To minimize the average video request latency of all users in the system, we formulate a large-scale 0-1 knapsack problem, which is NP-hard. To tackle it, we proposed a heuristic algorithm where the user viewing patterns extracted from the historical request information are taken into account. Specifically, we first design a cascading cache space allocation method to assign the total cache space of edge server to each segment of videos. After that, the original problem is decomposed into several small-scale yet individual tile caching subproblems with compressed solution space. Then, they are solved by using the dynamic programming algorithm with moderate complexity. To further enhance the caching performance, the PSO-based algorithm is designed to fine tune the parameters involved in the proposed caching algorithm. In addition, we introduce a content-based method to calculate the request probability of the newly generated videos. The effectiveness of the proposed algorithm is evaluated through simulations based on a real world dataset, where the results demonstrate a substantial improvement in both video request latency and cache hit rate compared to the benchmark methods. © 2014 IEEE.

关键词： Heuristic algorithms

来源：评论

学校读者我要写书评

暂无评论

AGCL: Aspect Graph Construction and Learning for Aspect-level Sentiment Classification 31

AGCL: Aspect Graph Construction and Learning for Aspect-leve...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Jian, Zhongquan Wu, Daihang Wang, Shaopan Wang, Yancheng Yao, Junfeng Wang, Meihong Wu, Qingqiang Institute of Artificial Intelligence Xiamen University China School of Informatics Xiamen University China School of Film Xiamen University China College of Management Mahidol University Thailand School of Computing and Data Science Xiamen University Malaysia Malaysia Key Laboratory of Digital Protection and Intelligent Processing of Intangible Cultural Heritage of Fujian and Taiwan Ministry of Culture and Tourism Xiamen University China Xiamen Key Laboratory of Intelligent Storage and Computing School of Informatics Xiamen University China

ISBN: (纸本)9798891761964

Prior studies on Aspect-level Sentiment Classification (ALSC) emphasize modeling interrelationships among aspects and contexts but overlook the crucial role of aspects themselves as essential domain knowledge. To this end, we propose AGCL, a novel Aspect Graph Construction and Learning method, aimed at furnishing the model with finely tuned aspect information to bolster its task-understanding ability. AGCL's pivotal innovations reside in Aspect Graph Construction (AGC) and Aspect Graph Learning (AGL), where AGC harnesses intrinsic aspect connections to construct the domain aspect graph, and then AGL iteratively updates the introduced aspect graph to enhance its domain expertise, making it more suitable for the ALSC task. Hence, this domain aspect graph can serve as a bridge connecting unseen aspects with seen aspects, thereby enhancing the model's generalization capability. Experiment results on three widely used datasets demonstrate the significance of aspect information for ALSC and highlight AGL's superiority in aspect learning, surpassing state-of-the-art baselines greatly. Code is available at https://***/jian-projects/agcl. © 2025 Association for Computational Linguistics.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Sensing Road Obstacles After Natural Disasters: A Survey

Sensing Road Obstacles After Natural Disasters: A Survey

引用

2023 International Conference on Artificial Intelligence of Things and Systems, AIoTSys 2023

作者： Muhammad, Auwal Sagir You, Jianyi He, Xin Chen, Longbiao Wang, Cheng School of Informatics Xiamen361005 China Xiamen University Key Laboratory of Multimedia Trusted Perception and Efficient Computing Ministry of Education of China Xiamen361005 China

ISBN: (纸本)9798350312270

Natural disasters, such as earthquakes, hurricanes, and flooding, can cause significant damage to roads and highways, leading to the formation of road obstacles that pose risks to drivers and hinder rescue efforts. With the rapid development of sensing technologies, a number of technologies have been proposed for disaster management, especially in disaster response and recovery. In this paper, we present a comprehensive review of sensing technologies for road obstacles focusing on the events of the aftermath of a natural disaster. Firstly, we discussed the major sensing technologies that are applied in disaster management including Vision, Lidar, Radar, and Ultrasonic. Secondly, we discussed the different techniques used in processing sensing data. Thirdly, we discussed the application of sensing technologies in the events of post-disaster and finally, we discussed the challenges faced in applying sensing technologies and highlighted areas for future studies. © 2023 IEEE.

关键词： Optical radar

来源：评论

学校读者我要写书评

暂无评论

AFFINEQUANT: AFFINE TRANSFORMATION QUANTIZATION FOR LARGE LANGUAGE MODELS 12

AFFINEQUANT: AFFINE TRANSFORMATION QUANTIZATION FOR LARGE LA...

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Ma, Yuexiao Li, Huixia Zheng, Xiawu Ling, Feng Xiao, Xuefeng Wang, Rui Wen, Shilei Chao, Fei Ji, Rongrong Key Laboratory of Multimedia Trusted Perception and Efficient Computing Ministry of Education of China School of Informatics Xiamen University 361005 China ByteDance Inc. China Peng Cheng Laboratory Shenzhen China Institute of Artificial Intelligence Xiamen University China

The significant resource requirements associated with Large-scale Language Models (LLMs) have generated considerable interest in the development of techniques aimed at compressing and accelerating neural networks. Among these techniques, Post-Training Quantization (PTQ) has emerged as a subject of considerable interest due to its noteworthy compression efficiency and cost-effectiveness in the context of training. Existing PTQ methods for LLMs limit the optimization scope to scaling transformations between pre- and post-quantization weights. This constraint results in significant errors after quantization, particularly in low-bit configurations. In this paper, we advocate for the direct optimization using equivalent Affine transformations in PTQ (AffineQuant). This approach extends the optimization scope and thus significantly minimizing quantization errors. Additionally, by employing the corresponding inverse matrix, we can ensure equivalence between the pre- and post-quantization outputs of PTQ, thereby maintaining its efficiency and generalization capabilities. To ensure the invertibility of the transformation during optimization, we further introduce a gradual mask optimization method. This method initially focuses on optimizing the diagonal elements and gradually extends to the other elements. Such an approach aligns with the Levy-Desplanques theorem, theoretically ensuring invertibility of the transformation. As a result, significant performance improvements are evident across different LLMs on diverse datasets. Notably, these improvements are most pronounced when using very low-bit quantization, enabling the deployment of large models on edge devices. To illustrate, we attain a C4 perplexity of 15.76 (2.26↓ vs 18.02 in OmniQuant) on the LLaMA2-7B model of W4A4 quantization without overhead. On zero-shot tasks, AffineQuant achieves an average of 58.61% accuracy (1.98% ↑ vs 56.63 in OmniQuant) when using 4/4-bit quantization for LLaMA-30B, which setting a new st

关键词： Cost effectiveness

来源：评论

学校读者我要写书评

暂无评论

Outlier-Aware Slicing for Post-Training Quantization in Vision Transformer 41

Outlier-Aware Slicing for Post-Training Quantization in Visi...

引用

41st International Conference on Machine Learning, ICML 2024

Post-Training Quantization (PTQ) is a vital technique for network compression and acceleration, gaining prominence as model sizes increase. This paper addresses a critical challenge in PTQ: the severe impact of outliers on the accuracy of quantized transformer architectures. Specifically, we introduce the concept of 'reconstruction granularity' as a novel solution to this issue, which has been overlooked in previous works. Our work provides theoretical insights into the role of reconstruction granularity in mitigating the outlier problem in transformer models. This theoretical framework is supported by empirical analysis, demonstrating that varying reconstruction granularities significantly influence quantization performance. Our findings indicate that different architectural designs necessitate distinct optimal reconstruction granularities. For instance, the multi-stage Swin Transformer architecture benefits from finer granularity, a deviation from the trends observed in ViT and DeiT models. We further develop an algorithm for determining the optimal reconstruction granularity for various ViT models, achieving state-of-the-art (SOTA) performance in PTQ. For example, applying our method to 4-bit quantization, the Swin-Base model achieves a Top-1 accuracy of 82.24% on the ImageNet classification task. This result surpasses the RepQ-ViT by 3.92% (82.24% VS 78.32%). Similarly, our approach elevates the ViT-Small to a Top-1 accuracy of 80.50%, outperforming NoisyQuant by 3.64% (80.50% VS 76.86%). Copyright 2024 by the author(s)

关键词：

来源：评论

学校读者我要写书评

暂无评论

ERQ: Error Reduction for Post-Training Quantization of Vision Transformers 41

ERQ: Error Reduction for Post-Training Quantization of Visio...

引用

41st International Conference on Machine Learning, ICML 2024

作者： Zhong, Yunshan Hu, Jiawei Huang, You Zhang, Yuxin Ji, Rongrong Institute of Artificial Intelligence Xiamen University China Key Laboratory of Multimedia Trusted Perception and Efficient Computing Ministry of Education of China Xiamen University China Department of Artificial Intelligence School of Informatics Xiamen University China Peng Cheng Laboratory China

Post-training quantization (PTQ) for vision transformers (ViTs) has garnered significant attention due to its efficiency in compressing ***, existing methods typically overlook the intricate interdependence between quantized weight and activation, leading to considerable quantization *** this paper, we propose ERQ, a two-step PTQ approach meticulously crafted to sequentially reduce the quantization error arising from activation and weight *** first introduces Activation quantization error reduction (Aqer) that strategically formulates the minimization of activation quantization error as a Ridge Regression problem, tackling it by updating weights with ***, ERQ introduces Weight quantization error reduction (Wqer) that adopts an iterative approach to mitigate the quantization error induced by weight *** each iteration, an empirically derived, efficient proxy is employed to refine the rounding directions of quantized weights, coupled with a Ridge Regression solver to curtail weight quantization *** results attest to the effectiveness of our ***, ERQ surpasses the state-of-the-art GPTQ by 22.36% in accuracy for W3A4 ViT-S. Copyright 2024 by the author(s)

关键词： Error correction

来源：评论

学校读者我要写书评

暂无评论

Adjusting Logit in Gaussian Form for Long-Tailed Visual Recognition

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第10期5卷 5026-5039页

作者： Li, Mengke Cheung, Yiu-Ming Lu, Yang Hu, Zhikai Lan, Weichao Huang, Hui Shenzhen518132 China Hong Kong Baptist University Department of Computer Science 999077 Hong Kong Xiamen University Fujian Key Laboratory of Sensing and Computing for Smart City School of Informatics Xiamen361005 China Shenzhen University College of Computer Science and Software Engineering Shenzhen518060 China

It is not uncommon that real-world data are distributed with a long tail. For such data, the learning of deep neural networks becomes challenging because it is hard to classify tail classes correctly. In the literature, several existing methods have addressed this problem by reducing classifier bias, provided that the features obtained with long-tailed data are representative enough. However, we find that training directly on long-tailed data leads to uneven embedding space. That is, the embedding space of head classes severely compresses that of tail classes, which is not conducive to subsequent classifier learning. This article therefore studies the problem of long-tailed visual recognition from the perspective of feature level. We introduce feature augmentation to balance the embedding distribution. The features of different classes are perturbed with varying amplitudes in Gaussian form. Based on these perturbed features, two novel logit adjustment methods are proposed to improve model performance at a modest computational overhead. Subsequently, the distorted embedding spaces of all classes can be calibrated. In such balanced-distributed embedding spaces, the biased classifier can be eliminated by simply retraining the classifier with class-balanced sampling data. Extensive experiments conducted on benchmark datasets demonstrate the superior performance of the proposed method over the state-of-the-art ones. © 2020 IEEE.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：