Authentication email has already been sent, please check your email box: and activate it as soon as possible.
You can login to My Profile and manage your email alerts.
If you haven’t received the email, please:
|
|
There are 15 papers published in subject: > since this site started. |
Select Subject |
Select/Unselect all | For Selected Papers |
Saved Papers
Please enter a name for this paper to be shown in your personalized Saved Papers list
|
1. Learning Fair and Efficient Policies in Sequential Public Goods Dilemmas | |||
CHEN Yi-Tian,LIU Xuan,CHEN Xin-Ning,ZHANG Shi-Geng | |||
Computer Science and Technology 12 May 2023 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Rational individuals can obtain higher rewards in the short term by defecting in social dilemmas, which, however, leads to low collective utility or even task failure. Many recent works have induced cooperative behaviors in social dilemmas though, they work only in stateless matrix games but fail in sequential social dilemmas. In tasks of sequential social dilemmas involving large number of players and complex states, cooperation is no longer simply one-step action and is hard to learn. Some works take payoffs equality into agents’ reward signals in decentralized multi-agent reinforcement learning to prevent some agents from taking up too much resources and starving others. However, this payoffs equality cannot lead to effective cooperative strategy, because it will force well-learned agents to sacrifice their high efficiency for equality if some agents have extremely low performance. In this work, we consider sequential public goods dilemmas in which group members can donate voluntarily for public welfare. We take fairness into account for training, well-learned agents obtain adequate rewards without being constrained by the policies of others, and meanwhile, the laggards have more access to learning owing to sufficient public goods. We empirically show that our method has excellent performance both in terms of collective efficiency and fairness. Compared to baselines, our agents acquire more universal and sustainable policies in sequential public goods dilemmas. | |||
TO cite this article:CHEN Yi-Tian,LIU Xuan,CHEN Xin-Ning, et al. Learning Fair and Efficient Policies in Sequential Public Goods Dilemmas[OL].[12 May 2023] http://en.paper.edu.cn/en_releasepaper/content/4760736 |
2. Multi-grained Location Matching with Universal Structural Coordinate Encoder for Referring Expression Grounding | |||
Yihong Zhao,Xiaojie Wang | |||
Computer Science and Technology 07 March 2023 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:\justifying Referring expression grounding is a multimodal matching task involving language and vision, with the goal of locating the object in an image that is closest to the current referring expression(RE). The key to this task is not only to use the attribute of the subject in the text, but also to fully utilize the complex location information (absolute and relative location) in the image. Existing methods only encode location feature using information such as 5-dimensional coordinate and object area, which ignore some possible fine-grained clues, such as the overlap between two objects, which can be helpful in distinguishing. This paper proposes a general structure modeling approach based on mask information that is applicable to both absolute and relative location. By modeling at a fine-grained level, this paper achieves the use of the same structure for both types of location information, thereby improving modular training efficiency. Specifically, for any two objects in an image, the model extracts small-scale binary feature constructed by mask information, which correspond to the subject and object parts of the relationship, respectively. Then, it performs phrase-guided object attention on this feature and update the initial representation of the objects through multi-layer message passing to obtain cross-feature information. Conducting experiments on three of the most commonly used related datasets, results show that compared to previous methods, the model can improve the performance of modular-based referring expression grounding models in a generalizable manner, further achieving superior performance. | |||
TO cite this article:Yihong Zhao,Xiaojie Wang. Multi-grained Location Matching with Universal Structural Coordinate Encoder for Referring Expression Grounding[OL].[ 7 March 2023] http://en.paper.edu.cn/en_releasepaper/content/4759356 |
3. Dynamic gradient compression federated learning under privacy protection | |||
ZHOU Tao,PENG Haipeng | |||
Computer Science and Technology 15 March 2022 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Methods based on deep learning have been widely used in various practical projects. Due to privacy policy reasons, traditional centralized learning may not be suitable for some engineering application scenarios with sensitive data, such as smart medical care, image recognition, etc. Federated learning has received extensive attention as a new collaborative learning method, which can break down data barriers between different institutions and improve model performance. However, the private information of individual clients can be inferred from their shared parameters, and at the same time, the communication consumption of federated learning systems is very high due to large batches of communication interactions. This paper proposes a dynamic gradient exchange privacy-preserving federated learning framework, which combines two technical theories of differential privacy and gradient compression. During the training process, differential privacy is used to interfere with the gradient parameters of the client, and dynamic gradient exchange is used to reject part of the "lazy" client communication. Theoretical analysis and experimental results demonstrate the superiority of the dynamic gradient exchange privacy-preserving federated learning framework in terms of accuracy, privacy security, and communication savings. | |||
TO cite this article:ZHOU Tao,PENG Haipeng. Dynamic gradient compression federated learning under privacy protection[OL].[15 March 2022] http://en.paper.edu.cn/en_releasepaper/content/4756701 |
4. Hierarchical Federated Learning with Gaussian Differential Privacy | |||
ZHOU Tao,PENG Hai-Peng | |||
Computer Science and Technology 28 February 2022 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Federated learning is a privacy preserving machine learning technology. Each participant can build the model without disclosing the underlying data, and only shares the weight update and gradient information of the model with the server. However, a lot of work shows that the attackers can easily obtain the client's contributions and the relevant privacy training data from the public shared gradient, so the gradient exchange is no longer safe. In order to ensure the security of Federated learning, in the differential privacy method, noise is added to the model update to obscure the contribution of the client, thereby resisting member reasoning attacks, preventing malicious clients from knowing other client information, and ensuring private output. This paper proposes a new differential privacy aggregation scheme, which adopts a more fine-grained hierarchy update strategy. For the first time, the $f$-differential privacy ($f$-DP) method is used for the privacy analysis of federated aggregation. Adding Gaussian noise disturbance model update in order to protect the privacy of the client level. We prove that the $f$-DP differential privacy method improves the previous privacy analysis by experiments. It accurately captures the loss of privacy at every communication round in federal training, and overcome the problem of ensuring privacy at the cost of reducing model utility in most previous work. At the same time, it provides a federal model updating scheme with wider applicability and better utility. When enough users participate in federated learning, the client-level privacy guarantee is achieved while minimizing model loss. | |||
TO cite this article:ZHOU Tao,PENG Hai-Peng. Hierarchical Federated Learning with Gaussian Differential Privacy[OL].[28 February 2022] http://en.paper.edu.cn/en_releasepaper/content/4756405 |
5. An attribute reduction algorithm based on particle swarm optimization | |||
Liu Jingyu | |||
Computer Science and Technology 09 April 2020 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Due to the explosive growth of data, the attribute dimensions of the data sets are getting higher, and the volume is larger, which leads to increased training overhead and decreased prediction accuracy of machine learning algorithms. And most of the current attribute reduction algorithms are based on a single attribute reduction, which is not easy to obtain the global optimum and has a large amount of calculation. Based on this, this paper proposes an attribute reduction algorithm based on particle swarm optimization (ARPSO). This algorithm designs the importance function of the attribute set based on the variable precision rough set, and uses particle swarm optimization algorithm to construct the optimization space, optimizes the attribute set in the data set globally, and reduces the redundant attributes of the data set to reduce the training overhead of machine learning algorithms and improve their prediction accuracy. The experimental results show that the attribute reduction performance of the ARPSO algorithm is significantly better than the common attribute reduction algorithms, which verifies the effectiveness of it. | |||
TO cite this article:Liu Jingyu. An attribute reduction algorithm based on particle swarm optimization[OL].[ 9 April 2020] http://en.paper.edu.cn/en_releasepaper/content/4751524 |
6. Personalized Review Recommendation based on Users' Aspect Sentiment | |||
CHUNLI HUANG, WENJUN JIANG,JIE WU,GUOJUN WANG | |||
Computer Science and Technology 02 April 2020 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Product reviews play an important role in guiding users' purchase decision-making in e-commerce platforms.However, it is challenging for users to find helpful reviews that meet their preferences and experiences among an overwhelming amount of reviews.While some existing personalized review recommendation models neglect an user's aspect preferences or the user-product interactions for measuring user similarity.Moreover, those works predict review helpfulness at the review-level (a review is taken as a whole); few of them consider the aspect-level.To address the above issues, this paper propose an users' aspect sentiment similarity-based personalized review recommendation model ($A2SPR$), which quantifies review helpfulness and recommends reviews that are customized for each individual.Firstly, the paper analyze users' aspect preferences from reviews and improve user similarity with users' fine-grained sentiment similarity and product relevance.Furthermore, the review helpfulness score is redefined at the aspect level, which indicates the review's reference value for users' purchase decisions. Finally, recommending the top $k$ helpful reviews for individuals based on the review helpfulness score. To validate the performance of the proposed model, eight baselines are developed and compared.Experimental results show that our model performs better than those baselines in both the coverage and precision. | |||
TO cite this article:CHUNLI HUANG, WENJUN JIANG,JIE WU, et al. Personalized Review Recommendation based on Users' Aspect Sentiment[OL].[ 2 April 2020] http://en.paper.edu.cn/en_releasepaper/content/4751464 |
7. DCDG-EA: Dynamic Convergence-Diversity Guided Evolutionary Algorithm for Many-Objective Optimization | |||
LI Zhi-Yong,LIN Ke | |||
Computer Science and Technology 09 April 2018 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Maintaining a good balance between convergence and diversity is particularly crucial to the performance of the evolutionary algorithms (EAs). However, traditional multi-objective evolutionary algorithms, which have demonstrated their competitive performance in a variety of practical problems involving two or three objectives, face significant challenges in many-objective optimization problems (MaOPs). This paper proposes a dynamic convergence-diversity guided evolutionary algorithm (DCDG-EA) for MaOPs by employing decomposition technique. The objective space of MaOPs is divided into $K$ subspaces by a set of uniformly distributed reference vectors. Each subspace has its own subpopulation and evolves in parallel with the other subspaces. In DCDG-EA, the balance between convergence and diversity is achieved through convergence-diversity based operator selection (CDOS) strategy and convergence-diversity based individual selection (CDIS) strategy. In CDOS, each operator in a set of operators is assigned a selection probability which is related to its convergence and diversity reward. On the basis of selection probability, an appropriate operator is selected to generate offspring. Furthermore, CDIS greatly overcomes the inefficiency of Pareto dominance. It updates each subpopulation by using two independent distance measures that respectively represent convergence and control diversity. The experimental results on DTLZ benchmark problems with up to 15 objectives show that our algorithm is highly competitive in comparison with the selected four state-of-the-art evolutionary algorithms in terms of convergence and diversity. | |||
TO cite this article:LI Zhi-Yong,LIN Ke. DCDG-EA: Dynamic Convergence-Diversity Guided Evolutionary Algorithm for Many-Objective Optimization[OL].[ 9 April 2018] http://en.paper.edu.cn/en_releasepaper/content/4744497 |
8. Re-ranking Answers by Discarding Biases in cQA Sites | |||
ZHAO Chenyang,XU Liutong | |||
Computer Science and Technology 11 January 2018 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:The vote mechanism employed to rank answers in community-based question answering websites is not very accurate because users will not vote to answers entirely base on their quality. Both the position and the appearance of an answer have an effect on the probability of users voting to it. Except the position bias and appearance bias, the following relationship between users impacts the voting results, too. As a result, the top answers obtained by vote mechanism is not reliable, especially when the votes is not sufficient. To rank answers based on their quality, this paper discussed the influences of the relationship between users to the vote mechanism and proposed a vote process model. Firstly, some assumptions about user's vote activities are made, then the vote process model is processed based on these assumptions to model user's voting process. Through the model inference process, the final equation to calculate answer's quality is get. Finally, an expectation-maximization algorithm is used to calculate the parameters in the final equation. By modeling user voting process,the vote process model can eliminate the influences of biases mentioned above and get the real quality evaluation of answers. Experiments on real dataset demonstrates the effectieness of the model proposed in this paper. In particular, when 30 percent of training data is used, the vote process model achieves a 10.1 percent improvement in precision and a 7.5 percent improvement in MRR compared with the joint click model, which is the state of the art click model. | |||
TO cite this article:ZHAO Chenyang,XU Liutong. Re-ranking Answers by Discarding Biases in cQA Sites[OL].[11 January 2018] http://en.paper.edu.cn/en_releasepaper/content/4743183 |
9. A Modeling Project of Chinese's Handwriting Character Online Recognition | |||
Huang Yishi,Liang Yan | |||
Computer Science and Technology 10 July 2013 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Provide a total solution for Chinese's handwriting character online recognition in the mobile phone. Set the basic definitions to form a special but very simple method to describe the graph. Classify and sum up a group of strokes for the regular script. Describe the independent strokes of it. Avoid using the traditional methods in the two-dimensional graphics. Give a part solution for the semi-cursive scripts and cursive in an open way. Produce a practical modeling system of handwriting recognition for mobile daily users. | |||
TO cite this article:Huang Yishi,Liang Yan. A Modeling Project of Chinese's Handwriting Character Online Recognition[OL].[10 July 2013] http://en.paper.edu.cn/en_releasepaper/content/4551483 |
10. An accelerator for attribute reduction based on perspective of objects and attributes | |||
Liang Jiye,Mi Junrong,Wei Wei,Wang Feng | |||
Computer Science and Technology 17 September 2012 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Feature selection is an active area of research in pattern recognition, machine learning and artificial intelligence, which greatly improve the performance of forecasting or classification. In rough set theory, attribute reduction, as a special form of feature selection, aims to retain the discernability of the original attribute set. To solve this problem, many heuristic attribute reduction algorithms have been proposed in the literature. However, these methods are computationally time-consuming for large scale datasets. Recently, an accelerator was introduced by computing reducts on gradually reducing the size of the universe. Although the accelerator can considerably shorten the computational time, it remains a challenging issue. To further enhance the efficiency of these algorithms, we develop a new accelerator for attribute reduction, which simultaneously reduces the size of the universe and the number of attributes at each iteration of the process of reduction. Based on the new accelerator, several representative heuristic attribute reduction algorithms are accelerated. Experiments show that these accelerating algorithms can significantly reduce computational time while maintaining their results the same as before. | |||
TO cite this article:Liang Jiye,Mi Junrong,Wei Wei, et al. An accelerator for attribute reduction based on perspective of objects and attributes[OL].[17 September 2012] http://en.paper.edu.cn/en_releasepaper/content/4489295 |
Select/Unselect all | For Selected Papers |
Saved Papers
Please enter a name for this paper to be shown in your personalized Saved Papers list
|
About Sciencepaper Online | Privacy Policy | Terms & Conditions | Contact Us
© 2003-2012 Sciencepaper Online. unless otherwise stated