检索结果-内蒙古大学图书馆

2024 Findings of the Association for Computational Linguistics: NAACL 2024

作者： Yao, Yao Li, Zuchao Zhao, Hai Department of Computer Science and Engineering Shanghai Jiao Tong University China MoE Key Lab of Artificial Intelligence AI Institute Shanghai Jiao Tong University China National Engineering Research Center for Multimedia Software School of Computer Science Wuhan University Wuhan430072 China

ISBN: (纸本)9798891761193

With the widespread use of language models (LMs) in NLP tasks, researchers have discovered the potential of Chain-of-thought (CoT) to assist LMs in accomplishing complex reasoning tasks by generating intermediate steps. However, human thought processes are often non-linear, rather than simply sequential chains of thoughts. Therefore, we propose Graph-of-Thought (GoT) reasoning, which models human thought processes not only as a chain but also as a graph. By representing thought units as nodes and connections between them as edges, our approach captures the non-sequential nature of human thinking and allows for a more realistic modeling of thought processes. GoT adopts a two-stage framework with an additional GoT encoder for thought graph representation and fuses the graph representation with the original input representation through a gated fusion mechanism. We evaluate GoT's performance on a text-only reasoning task (AQUA-RAT) and a multimodal reasoning task (scienceQA). Our model achieves significant improvement over the strong CoT baseline on the AQUA-RAT test set and boosts accuracy from 85.19% to 87.59% using the T5-base model over the state-of-the-art Multimodal-CoT (Zhang et al., 2023) on the scienceQA test set. Our code is publicly available at https://***/Zoeyyao27/Graphof-Thought. © 2024 Association for Computational Linguistics.

关键词： Rats

来源：评论

学校读者我要写书评

暂无评论

Neighborhood co-occurrence modeling in 3D point cloud segmentation

引用

Computational Visual Media 2022年第2期8卷 303-315页

作者： Jingyu Gong Zhou Ye Lizhuang Ma Department of Computer Science and Engineering Shanghai Jiao Tong UniversityShanghai 200240China Shanghai CLS Fintech Co. LTDShanghai 200030China MoE Key Lab of Artificial Intelligence Shanghai Jiao Tong UniversityShanghai 200240China

A significant performance boost has been achieved in point cloud semantic segmentation by utilization of the encoder-decoder architecture and novel convolution operations for point ***,co-occurrence relationships within a local region which can directly influence segmentation results are usually ignored by current *** this paper,we propose a neighborhood co-occurrence matrix(NCM)to model local co-occurrence relationships in a point *** generate target NCM and prediction NCM from semantic labels and a prediction map ***,Kullback-Leibler(KL)divergence is used to maximize the similarity between the target and prediction NCMs to learn the co-occurrence ***,for large scenes where the NCMs for a sampled point cloud and the whole scene differ greatly,we introduce a reverse form of KL divergence which can better handle the difference to supervise the prediction *** integrate our method into an existing backbone and conduct comprehensive experiments on three datasets:Semantic3D for outdoor space segmentation,and S3DIS and ScanNet v2 for indoor scene *** indicate that our method can significantly improve upon the backbone and outperform many leading competitors.

关键词： 3D vision point cloud co-occurrence relation modeling semantic segmentation

来源：评论

学校读者我要写书评

暂无评论

The Immense Impact of Reverse Edges on Large Hierarchical Networks

引用

engineering 2024年第5期36卷 240-249页

作者： Haosen Cao Bin-Bin Hu Xiaoyu Mo Duxin Chen Jianxi Gao Ye Yuan Guanrong Chen Tamás Vicsek Xiaohong Guan Hai-Tao Zhang MOE Engineering Research Center of Autonomous Intelligent Unmanned Systems State Key Laboratory of Intelligent Manufacturing Equipment and TechnologySchool of Artificial Intelligence and AutomationHuazhong University of Science and TechnologyWuhan 430074China School of Mechanical and Aerospace Engineering Nanyang Technological UniversitySingapore 639798Singapore Jiangsu Key Laboratory of Networked Collective Intelligence School of MathematicsSoutheast UniversityNanjing 210096China Department of Computer Science Rensselaer Polytechnic InstituteTroyNY 12180USA Department of Electronic Engineering City University of Hong KongHong Kong 999077China Department of Biological Physics Eötvös UniversityBudapest 1117Hungary MOE Key Laboratory for Intelligent Networks and Network Security School of Automation Science and EngineeringXi’an Jiaotong UniversityXi’an 710049China

Hierarchical networks are frequently encountered in animal groups,gene networks,and artificial engineering systems such as multiple robots,unmanned vehicle systems,smart grids,wind farm networks,and so *** structure of a large directed hierarchical network is often strongly influenced by reverse edges from lower-to higher-level nodes,such as lagging birds’howl in a flock or the opinions of lowerlevel individuals feeding back to higher-level ones in a social *** study reveals that,for most large-scale real hierarchical networks,the majority of the reverse edges do not affect the synchronization process of the entire network;the synchronization process is influenced only by a small part of these reverse edges along specific *** surprisingly,a single effective reverse edge can slow down the synchronization of a huge hierarchical network by over 60%.The effect of such edges depends not on the network size but only on the average in-degree of the involved *** overwhelming majority of active reverse edges turn out to have some kind of“bunching”effect on the information flows of hierarchical networks,which slows down synchronization *** finding refines the current understanding of the role of reverse edges in many natural,social,and engineering hierarchical networks,which might be beneficial for precisely tuning the synchronization rhythms of these *** study also proposes an effective way to attack a hierarchical network by adding a malicious reverse edge to it and provides some guidance for protecting a network by screening out the specific small proportion of vulnerable nodes.

关键词： Synchronizability Large hierarchical networks Reverse edges Information flows Complex networks

来源：评论

学校读者我要写书评

暂无评论

Rethinking and improving robustness of convolutional neural networks: a shapley value-based approach in frequency domain 22

Rethinking and improving robustness of convolutional neural ...

引用

Proceedings of the 36th International Conference on Neural Information Processing Systems

作者： Yiting Chen Qibing Ren Junchi Yan Department of Computer Science and Engineering MoE Key Lab of Artificial Intelligence Shanghai Jiao Tong University

ISBN: (纸本)9781713871088

The existence of adversarial examples poses concerns for the robustness of convolutional neural networks (CNN), for which a popular hypothesis is about the frequency bias phenomenon: CNNs rely more on high-frequency components (HFC) for classification than humans, which causes the brittleness of CNNs. However, most previous works manually select and roughly divide the image frequency spectrum and conduct qualitative analysis. In this work, we introduce Shapley value, a metric of cooperative game theory, into the frequency domain and propose to quantify the positive (negative) impact of every frequency component of data on CNNs. Based on the Shapley value, we quantify the impact in a fine-grained way and show intriguing instance disparity. Statistically, we investigate adversarial training(AT) and the adversarial attack in the frequency domain. The observations motivate us to perform an in-depth analysis and lead to multiple novel hypotheses about i) the cause of adversarial robustness of the AT model; ii) the fairness problem of AT between different classes in the same dataset; iii) the attack bias on different frequency components. Finally, we propose a Shapley-value guided data augmentation technique for improving the robustness. Experimental results on image classification benchmarks show its effectiveness.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Improving Speech Enhancement Using Audio Tagging Knowledge From Pre-Trained Representations and Multi-Task Learning

Improving Speech Enhancement Using Audio Tagging Knowledge F...

引用

2023 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2023

作者： Lin, Shaoxiong Zhang, Chao Qian, Yanmin Tsinghua University Department of Electronic Engineering Beijing China Shanghai Jiao Tong University MoE Key Lab of Artificial Intelligence Ai Institute Department of Computer Science and Engineering Shanghai China Suzhou Institute of Artificial Intelligence Shanghai Jiao Tong University Suzhou215000 China

ISBN: (纸本)9798350306897

In deep-learning-based speech enhancement (SE), an audio-knowledge-ignorant approach is often used, which estimates a denoising model to transform the noisy input speech into clean output speech without understanding the audio events that constitute the background noises. In this paper, an audio-knowledge-aware approach is proposed to improve SE, which explicitly leverages the knowledge from audio taggings to understand the background noises. Based on the recent progress in audio pattern analysis, the audio tagging knowledge is obtained using either additional input representations extracted by pre-trained audio tagging models, or from multi-task learning with extra audio event classification or regression tasks. Experimental results based on the DNS-2020 dataset and the pre-trained Wavegram-Logmel-CNN audio tagging model show that the proposed approach leads to considerable improvements in the STOI, SDR, and SI-SNR metrics. © 2023 IEEE.

关键词： Speech enhancement

来源：评论

学校读者我要写书评

暂无评论

CAM-GUI: A Conversational Assistant on Mobile GUI 18th

CAM-GUI: A Conversational Assistant on Mobile GUI

引用

18th National Conference on Man-Machine Speech Communication, NCMMSC 2023

作者： Zhu, Zichen Sun, Liangtai Yang, Jingkai Peng, Yifan Zou, Weilin Li, Ziyuan Li, Wutao Chen, Lu Ma, Yingzi Zhang, Danyang Fan, Shuai Yu, Kai X-LANCE Lab Department of Computer Science and Engineering MoE Key Lab of Artificial Intelligence SJTU AI Institute Shanghai Jiao Tong University Shanghai China Sichuan University Sichuan Chengdu China AISpeech Ltd. Suzhou China

ISBN: (纸本)9789819706006

Smartphone assistants are becoming more and more popular in our daily lives. These assistants mostly rely on the API-based Task-Oreiented Dialogue (TOD) systems, which limits the generality of these assistants, and the development of APIs would cost much labor and time. In this paper, we develop a Conversational Assistant on Mobile GUI (CAM-GUI), which can directly perform GUI operations on real devices, without the need of TOD-related backend APIs. To evaluate the performance of our assistant, we collect a dataset containing dialogues and GUI operation traces. From the experiment demonstrations and user studies, we show that the CAM-GUI reaches promising results. Our system demonstration can be downloaded from this link. You can watch our demo through this link. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Human computer interaction

来源：评论

学校读者我要写书评

暂无评论

ExpoEv: Enhancing Social Recommendation Service with Social Exposure and Feature Evolution

ExpoEv: Enhancing Social Recommendation Service with Social ...

引用

2023 IEEE International Conference on Web Services, ICWS 2023

作者： Ma, Li Zheng, Zuowu Huang, Xiuqi Zhang, Zhaoxiang Gao, Xiaofeng Guo, Jianxiong Chen, Guihai Shanghai Jiao Tong University MoE Key Lab of Artificial Intelligence Department of Computer Science and Engineering Shanghai China Advanced Institute of Natural Sciences Beijing Normal University Zhuhai China

ISBN: (纸本)9798350304855

Social networks are widely recognized as highly effective information sources for social recommendation services. However, previous social recommendation methods assumed that a user's preference factor and social trust factor shared a common latent feature space. Additionally, few studies have explored the incorporation of social information into the item domain for recommendations. To address these gaps, we propose ExpoEv, a deep collaborative filtering recommendation model that integrates social exposure based on feature evolution for social recommendation services. Specifically, we propose a social exposure module for both user and item domains that considers the number of items that a user's social friends interact with. Furthermore, we introduce a feature evolution component that enables the incorporation of social exposure information with social trust and attribute factors in the context of social recommendation services. Experiments demonstrate the effectiveness of our model in the quality of recommendation service. © 2023 IEEE.

关键词： Quality of service

来源：评论

学校读者我要写书评

暂无评论

Painterly Image Harmonization via Adversarial Residual Learning

Painterly Image Harmonization via Adversarial Residual Learn...

引用

IEEE Workshop on Applications of computer Vision (WACV)

作者： Xudong Wang Li Niu Junyan Cao Yan Hong Liqing Zhang Department of Computer Science and Engineering MoE Key Lab of Artificial Intelligence Shanghai Jiao Tong University

Image compositing plays a vital role in photo editing. After inserting a foreground object into another background image, the composite image may look unnatural and inharmonious. When the foreground is photorealistic and the background is an artistic painting, painterly image harmonization aims to transfer the style of background painting to the foreground object, which is a challenging task due to the large domain gap between foreground and background. In this work, we employ adversarial learning to bridge the domain gap between foreground feature map and background feature map. Specifically, we design a dual-encoder generator, in which the residual encoder produces the residual features added to the foreground feature map from main encoder. Then, a pixel-wise discriminator plays against the generator, encouraging the refined foreground feature map to be indistinguishable from background feature map. Extensive experiments demonstrate that our method could achieve more harmonious and visually appealing results than previous methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A dry-electrode enabled ECG-on-Chip with arrhythmia-aware data transmission

引用

science China(Information sciences) 2025年第2期68卷 338-353页

作者： Xinzi XU Yanxing SUO Yang ZHAO Peiyi ZHOU Xiao HAN Qiao CAI Min WANG Jiajun YUAN Liebin ZHAO Yongfu LI Guoxing WANG Yong LIAN Department of Micro-Nano Electronics and MoE Key Lab of Artificial Intelligence Shanghai Jiao Tong University Pediatric AI Clinical Application and Research Center Shanghai Engineering Research Center of Intelligence Pediatrics (SERCIP)Child Health Advocacy InstituteShanghai Children's Medical CenterSchool of MedicineShanghai Jiao Tong University Shanghai Engineering Research Center of Intelligence Pediatrics (SERCIP) Xin Hua Hospital Affiliate to Shanghai Jiao Tong University School of Medicine

Early detection and treatment of cardiovascular diseases(CVDs) can be significantly enhanced through the use of flexible wearable electrocardiogram(ECG) sensors, potentially reducing CVD-related mortality. This paper introduces an ECG-on-Chip(EoC) solution tailored for flexible ECG sensors, incorporating novel features to address common challenges in wearable ECG technology. The EoC integrates a chopper-stabilized capacitively-coupled instrumentation amplifier(CS-CCIA),ensuring a high common-mode rejection ratio(CMRR) and low noise performance. Performance is further boosted via a positive feedback loop(PFL) and a programmable gain amplifier(PGA) with shared on-chip calibration logic, which enhances input impedance and minimizes gain variability to ensure consistent algorithm performance. Additionally, a secondary chopping technique is employed to further reduce noise, achieving an input-referred noise level of 454 nVrms. The embedded algorithm within the EoC is designed to extract clinically meaningful features, facilitating robust real-time arrhythmia analysis. Fabricated using a 0.18 μm CMOS process, the EoC consumes 14.9 μW with a supply voltage of 1.2 V. The algorithm's efficacy has been validated over 0.4 million heartbeats, demonstrating a sensitivity of over 99.7% for R-peak detection and 98.3% for arrhythmia analysis. To demonstrate practical application, the EoC has been integrated with a commercial Bluetooth low energy(BLE)transceiver, forming a wireless arrhythmia monitoring sensor equipped with dry electrodes. This system consumes 292.04 μW in arrhythmia-aware transmission mode, supporting 13 days of operation with a 30 mAh thin-film battery.

关键词： ECG flexible dry electrodes analog front-end abnormality detection ASIC wireless transmission

来源：评论

学校读者我要写书评

暂无评论

A 5.3 pJ/Spike CMOS Neural Array Employing Time-Modulated Axon-Sharing and Background Mismatch Calibration Techniques

引用

IEEE Transactions on Biomedical Circuits and Systems 2023年第2期17卷 286-298页

作者： Qi, Xiang'ao Zhao, Jian Lou, Yuqing Wang, Guoxing Tang, Kea-Tiong Li, Yongfu Shanghai Jiao Tong University Department of Micro-Nano Electronics The MoE Key Lab of Artificial Intelligence Shanghai200240 China National Tsing Hua University Electrical Engineering Hsinchu30013 Taiwan

Inspired by the human brain, spiking neuron networks are promising to realize energy-efficient and low-latency neuromorphic computing. However, even state-of-the-art silicon neurons are orders of magnitude worse than biological neurons in terms of area and power consumption due to the limitations. Moreover, limited routing in typical CMOS processes is another challenge for realizing the fully-parallel high-throughput synapse connections compared to biological synapses. This paper presents an SNN circuit that utilizes resource-sharing techniques to address the two challenges. Firstly, a comparator sharing neuron circuit with a background calibration technique is proposed to shrink the size of a single neuron without performance degradation. Secondly, a time-modulated axon-sharing synapse system is proposed to realize a fully-parallel connection with limited hardware overhead. To validate the proposed approaches, a CMOS neuron array is designed and fabricated under a 55-nm process. It consists of 48 LIF neurons with 3125 neurons/mm 2 area density, power consumption of 5.3 pJ/spike, and equivalent 2304 fully parallel synapses providing a unit throughput of 5500 events/s/neuron. It proves the proposed approaches are promising to realize a high-throughput high-efficiency SNN with CMOS technology. © 2007-2012 IEEE.

关键词： Neural networks

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：