检索结果-内蒙古大学图书馆

arXiv 2024年

作者： He, Liyang Chen, Enhong Huang, Zhenya Wang, Fei Wang, Shijin Liu, Jiayu Sha, Jing School of Computer Science and Technology University of Science and Technology of China State Key Laboratory of Cognitive Intelligence Hefei China Anhui Province Key Laboratory of Big Data Analysis and Application University of Science and Technology of China State Key Laboratory of Cognitive Intelligence Hefei China State Key Laboratory of Cognitive Intelligence Hefei China School of Data Science University of Science and Technology of China State Key Laboratory of Cognitive Intelligence Hefei China iFLYTEK AI Research State Key Laboratory of Cognitive Intelligence Hefei China

Unsupervised semantic hashing has emerged as an indispensable technique for fast image search, which aims to convert images into binary hash codes without relying on labels. Recent advancements in the field demonstrate that employing large-scale backbones (e.g., ViT) in unsupervised semantic hashing models can yield substantial improvements. However, the inference delay has become increasingly difficult to overlook. Knowledge distillation provides a means for practical model compression to alleviate this delay. Nevertheless, the prevailing knowledge distillation approaches are not explicitly designed for semantic hashing. They ignore the unique search paradigm of semantic hashing, the inherent necessities of the distillation process, and the property of hash codes. In this paper, we propose an innovative Bit-mask Robust Contrastive knowledge Distillation (BRCD) method, specifically devised for the distillation of semantic hashing models. To ensure the effectiveness of two kinds of search paradigms in the context of semantic hashing, BRCD first aligns the semantic spaces between the teacher and student models through a contrastive knowledge distillation objective. Additionally, to eliminate noisy augmentations and ensure robust optimization, a cluster-based method within the knowledge distillation process is introduced. Furthermore, through a bit-level analysis, we uncover the presence of redundancy bits resulting from the bit independence property. To mitigate these effects, we introduce a bit mask mechanism in our knowledge distillation objective. Finally, extensive experiments not only showcase the noteworthy performance of our BRCD method in comparison to other knowledge distillation methods but also substantiate the generality of our methods across diverse semantic hashing models and backbones. The code for BRCD is available at https://***/hly1998/BRCD. Copyright © 2024, The Authors. All rights reserved.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Graph Structure Prefix Injection Transformer for Multi-Modal Entity Alignment

SSRN

引用

SSRN 2024年

作者： Li, Zhifei Luo, Xiangyu Zhang, Miao Zhang, Yan Xiao, Kui School of Computer Science and Information Engineering Hubei University Wuhan430062 China School of Cyber Science and Technology Hubei University Wuhan430062 China Key Laboratory of Intelligent Sensing System and Security Hubei University Ministry of Education Wuhan430062 China Hubei Key Laboratory of Big Data Intelligent Analysis and Application Hubei University Wuhan430062 China

Multi-modal entity alignment aims to integrate equivalent entities in diverse multi-modal knowledge graphs. However, previous studies have primarily focused on multi-modal fusion methods without considering the impact of graph structure heterogeneity on entity alignment tasks due to the distinct structural features of different knowledge graphs. Moreover, the multi-modal information that enriches entity representations, including image, textual content, and numerical attributes, frequently incorporates noise due to modal discrepancies. To tackle these challenges, we introduce the Graph Structural Prefix Injection transformer (GSIEA) for multi-modal entity alignment. This novel approach minimizes structural heterogeneity’s impact and optimizes structural data utilization by integrating a structural prefix inject interaction module within the multi-head attention mechanism. Then, we adopt a convolutional enhancement module to extract local features of multi-modal information and complete feature fusion by cross-modal weights. Extensive experiments on two datasets demonstrate that GSIEA outperforms previous baseline models, showcasing an average performance improvement of 3.26% on MRR, with a maximum enhancement of 12.5%. Additionally, the average performance improvement on Hits@1 is 4.96%, with a maximum enhancement of 16.92%. Our code is available at https://***/HubuKG/GSIEA. © 2024, The Authors. All rights reserved.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Backdoor Defense via Deconfounded Representation Learning

Backdoor Defense via Deconfounded Representation Learning

引用

Conference on computer Vision and Pattern Recognition (CVPR)

作者： Zaixi Zhang Qi Liu Zhicai Wang Zepu Lu Qingyong Hu Anhui Province Key Lab of Big Data Analysis and Application School of Computer Science and Technology University of Science and Technology of China State Key Laboratory of Cognitive Intelligence Hefei Anhui China University of Science and Technology of China Hong Kong University of Science and Technology

Deep neural networks (DNNs) are recently shown to be vulnerable to backdoor attacks, where attackers embed hidden backdoors in the DNN model by injecting a few poisoned examples into the training dataset. While extensive efforts have been made to detect and remove backdoors from backdoored DNNs, it is still not clear whether a backdoor-free clean model can be directly obtained from poisoned datasets. In this paper, we first construct a causal graph to model the generation process of poisoned data and find that the backdoor attack acts as the confounder, which brings spurious associations between the input images and target labels, making the model predictions less reliable. Inspired by the causal understanding, we propose the Causality-inspired Backdoor Defense (CBD), to learn deconfounded representations for reliable classification. Specifically, a backdoored model is intentionally trained to capture the confounding effects. The other clean model dedicates to capturing the desired causal effects by minimizing the mutual information with the confounding representations from the backdoored model and employing a sample-wise re-weighting scheme. Extensive experiments on multiple benchmark datasets against 6 state-of-the-art attacks verify that our proposed defense method is effective in reducing backdoor threats while maintaining high accuracy in predicting benign samples. Further analysis shows that CBD can also resist potential adaptive attacks. The code is available at https://***/zaixizhang/CBD.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Knowledge-Injected Curriculum Pretraining Framework for Question Answering

arXiv

引用

arXiv 2024年

作者： Lin, Xin Su, Tianhuang Huang, Zhenya Xue, Shangzi Liu, Haifeng Chen, Enhong School of Computer Science and Technology University of Science and Technology of China State Key Laboratory of Cognitive Intelligence Hefei China Guangdong OPPO Mobile Telecommunications Corp. Ltd Shenzhen China University of Science and Technology of China Hefei China Anhui Province Key Laboratory of Big Data Analysis and Application University of Science and Technology of China State Key Laboratory of Cognitive Intelligence Hefei China

Knowledge-based question answering (KBQA) is a key task in natural language processing research, and also an approach to access the web data and knowledge, which requires exploiting knowledge graphs (KGs) for reasoning. In the literature, one promising solution for KBQA is to incorporate the pretrained language model (LM) with KGs by generating KG-centered pretraining corpus, which has shown its superiority. However, these methods often depend on specific techniques and resources to work, which may not always be available and restrict its application. Moreover, existing methods focus more on improving language understanding with KGs, while neglect the more important human-like complex reasoning. To this end, in this paper, we propose a general Knowledge-Injected Curriculum Pretraining framework (KICP) to achieve comprehensive KG learning and exploitation for KBQA tasks, which is composed of knowledge injection (KI), knowledge adaptation (KA) and curriculum reasoning (CR). Specifically, the KI module first injects knowledge into the LM by generating KG-centered pretraining corpus, and generalizes the process into three key steps that could work with different implementations for flexible application. Next, the KA module learns knowledge from the generated corpus with LM equipped with an adapter as well as keeps its original natural language understanding ability to reduce the negative impacts of the difference between the generated and natural corpus. Last, to enable the LM with complex reasoning, the CR module follows human reasoning patterns to construct three corpora with increasing difficulties of reasoning, and further trains the LM from easy to hard in a curriculum manner to promote model learning. We provide an implementation of the general framework, and evaluate the proposed KICP on four real-word datasets. The results demonstrate that our framework can achieve higher performances, and have good generalization ability to other QA tasks. Copyright © 2024, The

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Large-Scale OD Matrix Estimation with A Deep Learning Method

arXiv

引用

arXiv 2023年

作者： Xiong, Zheli Lian, Defu Chen, Enhong Chen, Gang Cheng, Xiaomin The School of Data Science University of Science and Technology of China Anhui Hefei230000 China The Anhui Province Key Laboratory of Big Data Analysis and Application School of Computer Science and Technology University of Science and Technology of China Anhui Hefei230000 China The Yangtze River Delta Information Intelligence Innovation Research Institute China

The estimation of origin-destination (OD) matrices is a crucial aspect of Intelligent Transport Systems (ITS). It involves adjusting an initial OD matrix by regressing the current observations like traffic counts of road sections (e.g., using least squares). However, the OD estimation problem lacks sufficient constraints and is mathematically underdetermined. To alleviate this problem, some researchers incorporate a prior OD matrix as a target in the regression to provide more structural constraints. However, this approach is highly dependent on the existing prior matrix, which may be outdated. Others add structural constraints through sensor data, such as vehicle trajectory and speed, which can reflect more current structural constraints in real-time. Our proposed method integrates deep learning and numerical optimization algorithms to infer matrix structure and guide numerical optimization. This approach combines the advantages of both deep learning and numerical optimization algorithms. The neural network(NN) learns to infer structural constraints from probe traffic flows, eliminating dependence on prior information and providing real-time performance. Additionally, due to the generalization capability of NN, this method is economical in engineering. We conducted tests to demonstrate the good generalization performance of our method on a large-scale synthetic dataset. Subsequently, we verified the stability of our method on real traffic data. Our experiments provided confirmation of the benefits of combining NN and numerical *** Codes I.2.1 Copyright © 2023, The Authors. All rights reserved.

关键词： Statistical tests

来源：评论

学校读者我要写书评

暂无评论

引用

IEEE International Conference on Multimedia and Expo (ICME)

作者： Xiaoke Zhu Danyang Li Xiaopan Chen Fumin Qi Fan Zhang Xiao-Yuan Jing School of Computer and Information Engineering Henan University Kaifeng China Henan Engineering Research Center of Intelligent Technology and Application Henan University Kaifeng China Department of System Operation National Supercomuting Center in Shenzhen Shenzhen China Henan Key Laboratory of Big Data Analysis and Processing Henan University Kaifeng China School of Computer Science Wuhan University Wuhan China

ISBN: (数字)9798350390155

ISBN: (纸本)9798350390162

Facial image based kinship verification aims to decide whether there exists kinship between the given facial images. In practice, the cross-generation differences will cause adverse effects on kinship verification, which limits the performance. Therefore, how to mine the implied similarity from facial images with large cross-generation divergence is an important problem in kinship verification, which has not yet been well studied. In view of this, we propose a Similarity Mining via Implicit matching pattern LEarning (SMILE) approach for kinship verification. Specifically, SMILE mainly consists of two modules, including a Semi-coupled Multi-pattern Similarity Learning (SMSL) module and a Cross-Generation Feature Normalization (CGFN) module. The SMSL module is designed to learn multiple semi-coupled matching patterns for mining the implicit facial similarity information from different perspectives. The CGFN module aims to reduce the divergence between facial images of parent and child. Extensive experiments demonstrate that the proposed approach outperforms the existing state-of-the-art methods.

关键词： Pattern matching Similarity learning

来源：评论

学校读者我要写书评

暂无评论

Full-atom protein pocket design via iterative refinement 23

Full-atom protein pocket design via iterative refinement

引用

Proceedings of the 37th International Conference on Neural Information Processing Systems

作者： Zaixi Zhang Zepu Lu Zhongkai Hao Marinka Zitnik Qi Liu Anhui Province Key Lab of Big Data Analysis and Application School of Computer Science and Technology University of Science and Technology of China and State Key Laboratory of Cognitive Intelligence Hefei Anhui China and Harvard University Anhui Province Key Lab of Big Data Analysis and Application School of Computer Science and Technology University of Science and Technology of China and State Key Laboratory of Cognitive Intelligence Hefei Anhui China Dept. of Comp. Sci. and Tech. Institute for AI THBI Lab BNRist Center Tsinghua-Bosch Joint ML Center Tsinghua Harvard University

The design and optimization of functional proteins that bind specific ligand molecules is paramount in therapeutics and bio-engineering. A critical yet formidable task in this endeavor is the design of the protein pocket, which is the cavity region of the protein where the ligand binds. Current methods are plagued by inefficient generation, inadequate context modeling of the ligand molecule, and the inability to generate side-chain atoms. Here, we present the Full-Atom Iterative Refinement (FAIR) method, designed to address these challenges by facilitating the co-design of protein pocket sequences, specifically residue types, and their corresponding 3D structures. FAIR operates in two steps, proceeding in a coarse-to-fine manner (transitioning from protein backbone to atoms, including side chains) for a full-atom generation. In each iteration, all residue types and structures are simultaneously updated, a process termed full-shot refinement. In the initial stage, the residue types and backbone coordinates are refined using a hierarchical context encoder, complemented by two structure refinement modules that capture both inter-residue and pocket-ligand interactions. The subsequent stage delves deeper, modeling the side-chain atoms of the pockets and updating residue types to ensure sequence-structure congruence. Concurrently, the structure of the binding ligand is refined across iterations to accommodate its inherent flexibility. Comprehensive experiments show that FAIR surpasses existing methods in designing superior pocket sequences and structures, producing average improvement exceeding 10% in AAR and RMSD metrics.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Dual Test-time Training for Out-of-distribution Recommender System

arXiv

引用

arXiv 2024年

作者： Yang, Xihong Wang, Yiqi Chen, Jin Fan, Wenqi Zhao, Xiangyu Zhu, En Liu, Xinwang Lian, Defu School of Computer National University of Defense Technology Changsha410073 China School of Business and Management Hong Kong University of Science and Technology Hong Kong The Hong Kong Polytechnic University Hong Kong Department of Data Science City University of Hong Kong Hong Kong Anhui Province Key Laboratory of Big Data Analysis and Application School of Computer Science and Technology University of Science and Technology of China Anhui Hefei230000 China

Deep learning has been widely applied in recommender systems, which has recently achieved revolutionary progress. However, most existing learning-based methods assume that the user and item distributions remain unchanged between the training phase and the test phase. However, the distribution of user and item features can naturally shift in real-world scenarios, potentially resulting in a substantial decrease in recommendation performance. This phenomenon can be formulated as an Out-Of-Distribution (OOD) recommendation problem. To address this challenge, we propose a novel Dual Test-Time-Training framework for OOD Recommendation, termed DT3OR. In DT3OR, we incorporate a model adaptation mechanism during the test-time phase to carefully update the recommendation model, allowing the model to adapt specially to the shifting user and item features. To be specific, we propose a self-distillation task and a contrastive task to assist the model learning both the user’s invariant interest preferences and the variant user/item characteristics during the test-time phase, thus facilitating a smooth adaptation to the shifting features. Furthermore, we provide theoretical analysis to support the rationale behind our dual test-time training framework. To the best of our knowledge, this paper is the first work to address OOD recommendation via a test-time-training strategy. We conduct experiments on five datasets with various backbones. Comprehensive experimental results have demonstrated the effectiveness of DT3OR compared to other state-of-the-art baselines. Copyright © 2024, The Authors. All rights reserved.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

ZuantuSet: A Collection of Historical Chinese Visualizations and Illustrations

arXiv

引用

arXiv 2025年

作者： Mei, Xiyao Zhang, Yu Yang, Chaofan Shi, Rui Yuan, Xiaoru National Key Laboratory of General Artificial Intelligence School of Intelligence Science and Technology Peking University Beijing China Department of Computer Science University of Oxford Oxford United Kingdom Center for Research on Ancient Chinese History Peking University Beijing China National Engineering Laboratory for Big Data Analysis and Application Peking University Beijing China PKU-WUHAN Institute for Artificial Intelligence Wuhan China

Historical visualizations are a valuable resource for studying the history of visualization and inspecting the cultural context where they were created. When investigating historical visualizations, it is essential to consider contributions from different cultural frameworks to gain a comprehensive understanding. While there is extensive research on historical visualizations within the European cultural framework, this work shifts the focus to ancient China, a cultural context that remains underexplored by visualization researchers. To this aim, we propose a semi-automatic pipeline to collect, extract, and label historical Chinese visualizations. Through the pipeline, we curate ZuantuSet, a dataset with over 71K visualizations and 108K illustrations. We analyze distinctive design patterns of historical Chinese visualizations and their potential causes within the context of Chinese history and culture. We illustrate potential usage scenarios for this dataset, summarize the unique challenges and solutions associated with collecting historical Chinese visualizations, and outline future research directions. Copyright © 2025, The Authors. All rights reserved.

关键词： data collection

来源：评论

学校读者我要写书评

暂无评论

Dependency and Association Enhanced Finding Similar Exercises in Online Education System

Dependency and Association Enhanced Finding Similar Exercise...

引用

IEEE International Conference on Cloud Computing and Intelligence Systems (CCIS)

作者： Zihang Tang Hao Wang Weibo Gao Zhenya Huang Qi Liu Jun Yu School of Data Science University of Science and Technology of China Anhui Province Key Laboratory of Big Data Analysis and Application School of Computer Science and Technology University of Science and Technology of China iFLYTEK Research iFLYTEK CO. LTD.

ISBN: (纸本)9781665456579

Finding Similar Exercises (FSE) is a tough nut to crack in online education systems, especially among large scale exercise corpus. Given exercises with associated attributes, contents (texts and images) and knowledge concepts, most solutions learn exercise representations in isolation, where only contents or concepts are exploited. Though a few works try to combine concept-exercise associations for capturing long-range similarity among exercises, they just utilize a single relation among concepts but neglect the multiple dependencies. In this paper, we propose a novel Dependency and Association enhanced Finding Similar Exercises framework, namely DAFSE, which can consider both the concept-exercise associations and the multiple dependency relations among concepts, simultaneously. Specifically, we first build a concept-exercise-content multi-layer graph and embed each concept, exercise and content node through an Embedding Layer. Then we introduce Attentive Combination Layer to effectively preserve the long-range similarity as well as rich domain knowledge to exercise representations by propagating on the multi-layer graph. Finally, we use a Similarity Score Layer to assess the similarity of exercise pairs. Extensive experiments on real-world datasets demonstrate the effectiveness of DAFSE.

关键词： Cloud computing Computational modeling Education Collaboration Labeling Intelligent systems

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：