检索结果-内蒙古大学图书馆

Adversarial attribute-image person re-identification

学校读者我要写书评

暂无评论

arXiv 2017年

作者： Yin, Zhou Zheng, Wei-Shi Wu, Ancong Yu, Hong-Xing Wan, Hai Guo, Xiaowei Huang, Feiyue Lai, Jianhuang School of Data and Computer Science Sun Yat-sen University Guangzhou China YouTu Lab Tencent Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education China

While attributes have been widely used for person re-identification (Re-ID) which aims at matching the same person images across disjoint camera views, they are used either as extra features or for performing multi-task learning to assist the image-image matching task. However, how to find a set of person images according to a given attribute description, which is very practical in many surveillance applications, remains a rarely investigated cross-modality matching problem in person Re-ID. In this work, we present this challenge and formulate this task as a joint space learning problem. By imposing an attribute-guided attention mechanism for images and a semantic consistent adversary strategy for attributes, each modality, i.e., images and attributes, successfully learns semantically correlated concepts under the guidance of the other. We conducted extensive experiments on three attribute datasets and demonstrated that the proposed joint space learning method is so far the most effective method for the attribute-image cross-modality person Re-ID problem. Copyright © 2017, The Authors. All rights reserved.

关键词： Semantics

A generalized matrix splitting algorithm

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Yuan, Ganzhao Zheng, Wei-Shi Shen, Li Ghanem, Bernard Guangzhou Guangdong510275 China Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education Beijing221143 China Tencent AI Lab Shenzhen China Saudi Arabia

Composite function minimization captures a wide spectrum of applications in both computer vision and machine learning. It includes bound constrained optimization, `1 norm regularized optimization, and `0 norm regularized optimization as special cases. This paper proposes and analyzes a new Generalized Matrix Splitting Algorithm (GMSA) for minimizing composite functions. It can be viewed as a generalization of the classical Gauss-Seidel method and the Successive Over-Relaxation method for solving linear systems in the literature. Our algorithm is derived from a novel triangle operator mapping, which can be computed exactly using a new generalized Gaussian elimination procedure. We establish the global convergence, convergence rate, and iteration complexity of GMSA for convex problems. In addition, we also discuss several important extensions of GMSA. Finally, we validate the performance of our proposed method on three particular applications: nonnegative matrix factorization, `0 norm regularized sparse coding, and `1 norm regularized Dantzig selector problem. Extensive experiments show that our method achieves state-of-the-art performance in term of both efficiency and efficacy. Copyright © 2018, The Authors. All rights reserved.

关键词： Convex optimization

EEG-based fatigue classification by using parallel hidden Markov model and pattern classifier combination

学校读者我要写书评

暂无评论

EEG-based fatigue classification by using parallel hidden Ma...

19th International Conference on Neural Information Processing, ICONIP 2012

作者： Sun, Hui Lu, Bao-Liang Center for Brain-Like Computing and Machine Intelligence Department of Computer Science and Engineering Shanghai Jiao Tong University 800 Dongchuan Road Shanghai 200240 China MOE-Microsoft Key Lab. for Intelligent Computing and Intelligent Systems Shanghai Jiao Tong University 800 Dongchuan Road Shanghai 200240 China Shanghai Key Laboratory of Scalable Computing and Systems Shanghai Jiao Tong University 800 Dongchuan Road Shanghai 200240 China MOE Key Laboratory of Systems Biomedicine Shanghai Jiao Tong University 800 Dongchuan Road Shanghai 200240 China

ISBN: (纸本)9783642344770

Fatigue is the most important reason leading to traffic accidents. In order to ensure traffic safety, various methods based on electroencephalogram (EEG) are proposed. But most of them, either regression or classification, are focused on the relationship between feature space and observation values, so the changing patterns of features are ignored or discarded. In this paper, we propose a new fatigue classification method by using parallel hidden-Markov-model and pattern classifier combination techniques, where each model represents a particular fatigue-high-related feature. In the experiment, subjects are asked to accomplish some simple tasks, and both their fatigue states and their EEG signals are recorded simultaneously. Experimental results indicate that the mean error rate obtained by using our new method are 11.15% for classifying 3 states and 16.91% for classifying 4 states, respectively, while the existing approach can only reach 16.45% and 23.55% under the same condition. © 2012 Springer-Verlag.

关键词： Electroencephalography

Online vigilance analysis combining video and electrooculography features

学校读者我要写书评

暂无评论

Online vigilance analysis combining video and electrooculogr...

19th International Conference on Neural Information Processing, ICONIP 2012

作者： Du, Ruo-Fei Liu, Ren-Jie Wu, Tian-Xiang Lu, Bao-Liang Center for Brain-like Computing and Machine Intelligence Department of Computer Science and Engineering Shanghai Jiao Tong University 800 Dongchuan Road Shanghai 200240 China MOE-Microsoft Key Lab. for Intelligent Computing and Intelligent Systems Shanghai Jiao Tong University 800 Dongchuan Road Shanghai 200240 China Shanghai Key Laboratory of Scalable Computing and Systems Shanghai Jiao Tong University 800 Dongchuan Road Shanghai 200240 China MOE Key Laboratory of Systems Biomedicine Shanghai Jiao Tong University 800 Dongchuan Road Shanghai 200240 China

ISBN: (纸本)9783642344992

In this paper, we propose a novel system to analyze vigilance level combining both video and Electrooculography (EOG) features. For one thing, the video features extracted from an infrared camera include percentage of closure (PERCLOS) and eye blinks, slow eye movement (SEM), rapid eye movement (REM) are also extracted from EOG signals. For another, other features like yawn frequency, body posture and face orientation are extracted from the video by using Active Shape Model (ASM). The results of our experiments indicate that our approach outperforms the existing approaches based on either video or EOG merely. In addition, the prediction offered by our model is in close proximity to the actual error rate of the subject. We firmly believe that this method can be widely applied to prevent accidents like fatigued driving in the future. © 2012 Springer-Verlag.

关键词： Eye movements

Refining BERT embeddings for document hashing via mutual information maximization

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Ou, Zijing Su, Qinliang Yu, Jianxing Zhao, Ruihui Zheng, Yefeng Liu, Bang School of Computer Science and Engineering Sun Yat-sen University Guangzhou China School of Artificial Intelligence Sun Yat-sen University Guangdong China Tencent Jarvis Lab Rali & Mila Université de Montréal Guangdong Key Lab. of Big Data Analysis and Processing Guangzhou China Key Lab. of Machine Intelligence and Advanced Computing Ministry of Education China

Existing unsupervised document hashing methods are mostly established on generative models. Due to the difficulties of capturing long dependency structures, these methods rarely model the raw documents directly, but instead to model the features extracted from them (e.g. bag-of-words (BOW), TFIDF). In this paper, we propose to learn hash codes from BERT embeddings after observing their tremendous successes on downstream tasks. As a first try, we modify existing generative hashing models to accommodate the BERT embeddings. However, little improvement is observed over the codes learned from the old BOW or TFIDF features. We attribute this to the reconstruction requirement in the generative hashing, which will enforce irrelevant information that is abundant in the BERT embeddings also compressed into the codes. To remedy this issue, a new unsupervised hashing paradigm is further proposed based on the mutual information (MI) maximization principle. Specifically, the method first constructs appropriate global and local codes from the documents and then seeks to maximize their mutual information. Experimental results on three benchmark datasets demonstrate that the proposed method is able to generate hash codes that outperform existing ones learned from BOW features by a substantial margin. Copyright © 2021, The Authors. All rights reserved.

关键词： Embeddings

Converting continuous-space language models into N-gram language models for statistical machine translation

学校读者我要写书评

暂无评论

Converting continuous-space language models into N-gram lang...

2013 Conference on Empirical Methods in Natural Language Processing, EMNLP 2013

作者： Wang, Rui Utiyama, Masao Goto, Isao Sumita, Eiichro Zhao, Hai Lu, Bao-Liang Center for Brain-Like Computing and Machine Intelligence Department of Computer Science and Engineering Shanghai Jiao Tong Unviersity Shanghai200240 China Multilingual Translation Laboratory MASTAR Project National Institute of Information and Communications Technology 3-5 Hikaridai Keihanna Science City Kyoto619-0289 Japan MOE-Microsoft Key Lab. for Intelligent Computing and Intelligent Systems Shanghai Jiao Tong Unviersity Shanghai200240 China

ISBN: (纸本)9781937284978

Neural network language models, or continuous-space language models (CSLMs), have been shown to improve the performance of statistical machine translation (SMT) when they are used for reranking n-best translations. However, CSLMs have not been used in the first pass decoding of SMT, because using CSLMs in decoding takes a lot of time. In contrast, we propose a method for converting CSLMs into back-off n-gram language models (BNLMs) so that we can use converted CSLMs in decoding. We show that they outperform the original BNLMs and are comparable with the traditional use of CSLMs in reranking. © 2013 Association for Computational Linguistics.

关键词： Decoding

Generating software test data by particle swarm optimization

学校读者我要写书评

暂无评论

Lecture Notes in Computer Science (including subseries Lectu...

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial intelligence and Lecture Notes in Bioinformatics) 2014年 8886卷 37-47页

作者： Jia, Ya-Hui Chen, Wei-Neng Zhang, Jun Li, Jing-Jing School of Information Science and Technology Sun Yat-sen University Guangzhou China School of Advanced Computing Sun Yat-sen University Guangzhou China Key Lab. Machine Intelligence and Advanced Computing Ministry of Education China Engineering Research Center of Supercomputing Engineering Software MOE China School of Computer Science South China Normal University Guangzhou China

Search-based method using meta-heuristic algorithms is a hot topic in automatic test data generation. In this paper, we develop an automatic test data generating tool named particle swarm optimization data generation tool (PSODGT). The PSODGT is characterized by the following two features. First, the PSODGT adopts the condition-decision coverage (C/DC) as the criterion of software testing, aiming to build an efficient test data set that covers all conditions. Second, the PSODGT uses a particle swarm optimization (PSO) approach to generate test data set. In addition, a new position initialization technique is developed for PSO. Instead of initializing the test data randomly, the proposed technique uses the previously-found test data that can reach the target condition as the initial positions so that the search speed of PSODGT can be further accelerated. The PSODGT is tested on four practical programs. Experimental results show that the proposed PSO approach is promising. © Springer International Publishing Switzerland 2014.

关键词： Software testing

Unsupervised person re-identification by soft multilab.l learning

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Yu, Hong-Xing Zheng, Wei-Shi Wu, Ancong Guo, Xiaowei Gong, Shaogang Lai, Jian-Huang Sun Yat-sen University China YouTu Lab Tencent Queen Mary University of London United Kingdom Key Laboratory of Machine Intelligence and Advanced Computing Ministryof Education China

Although unsupervised person re-identification (RE-ID) has drawn increasing research attentions due to its potential to address the scalab.lity problem of supervised RE-ID models, it is very challenging to learn discriminative information in the absence of pairwise lab.ls across disjoint camera views. To overcome this problem, we propose a deep model for the soft multilab.l learning for unsupervised RE-ID. The idea is to learn a soft multilab.l (real-valued lab.l likelihood vector) for each unlab.led person by comparing (and representing) the unlab.led person with a set of known reference persons from an auxiliary domain. We propose the soft multilab.l-guided hard negative mining to learn a discriminative embedding for the unlab.led target domain by exploring the similarity consistency of the visual features and the soft multilab.ls of unlab.led target pairs. Since most target pairs are cross-view pairs, we develop the cross-view consistent soft multilab.l learning to achieve the learning goal that the soft multilab.ls are consistently good across different camera views. To enable effecient soft multilab.l learning, we introduce the reference agent learning to represent each reference person by a reference agent in a joint embedding. We evaluate our unified deep model on Market-1501 and DukeMTMC-reID. Our model outperforms the state-of-the-art unsupervised RE-ID methods by clear margins. Copyright © 2019, The Authors. All rights reserved.

关键词： Embeddings

Inserting anybody in diffusion models via celeb basis 23

学校读者我要写书评

暂无评论

Inserting anybody in diffusion models via celeb basis

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

作者： Ge Yuan Xiaodong Cun Yong Zhang Maomao Li Chenyang Qi Xintao Wang Ying Shan Huicheng Zheng School of Computer Science and Engineering Sun Yat-sen University and Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education China and Guangdong Key Laboratory of Information Security Technology and Tencent AI Lab Tencent AI Lab The Hong Kong University of Science and Technology School of Computer Science and Engineering Sun Yat-sen University and Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education China and Guangdong Key Laboratory of Information Security Technology

Exquisite demand exists for customizing the pretrained large text-to-image model, e.g. Stable Diffusion, to generate innovative concepts, such as the users themselves. However, the newly-added concept from previous customization methods often shows weaker combination abilities than the original ones even given several images during training. We thus propose a new personalization method that allows for the seamless integration of a unique individual into the pre-trained diffusion model using just one facial photograph and only 1024 learnable parameters under 3 minutes. So we can effortlessly generate stunning images of this person in any pose or position, interacting with anyone and doing anything imaginable from text prompts. To achieve this, we first analyze and build a well-defined celeb basis from the embedding space of the pre-trained large text encoder. Then, given one facial photo as the target identity, we generate its own embedding by optimizing the weight of this basis and locking all other parameters. Empowered by the proposed celeb basis, the new identity in our customized model showcases a better concept combination ability than previous personalization methods. Besides, our model can also learn several new identities at once and interact with each other where the previous customization model fails to. Project page is at: http://***. Code is at: https://***/ygtxr1997/CelebBasis.

关键词：