检索结果-内蒙古大学图书馆

Contrastive Learning and Multi-Choice Negative Sampling recommendation

INTERNATIONAL JOURNAL OF ADVANCED COMPUTER SCIENCE AND APPLICATIONS 2024年第5期15卷 905-912页

作者： Xue, Yun Cai, Xiaodong Fang, Sheng Zhou, Li Guilin Univ Elect Technol Sch Informat & Commun Guilin Peoples R China Nanning West Bank Fenggu Business Data Co Ltd Nanning Peoples R China

Most existing recommendation models that directly model user interests on user-item interaction data usually ignore the natural noise present in the interaction data, leading to bias in the model's learning of user preferences during data propagation and aggregation. In addition, the currently adopted negative sampling strategy does not consider the relationship between the prediction scores of positive samples and the degree of difficulty of negative samples, and is unable to adaptively select a suitable negative sample for each positive sample, leading to a decrease in the model recommendation performance. In order to solve the above problems, this paper proposes a Contrastive Learning and Multi-choice Negative Sampling recommendation. Firstly, an improved topology-aware pruning strategy is used to process the user-item bipartite graph, which uses the topology information of the graph to remove noise and improve the accuracy of model prediction. In addition, a new multivariate selective negative sampling module is designed, which ensures that each positive sample selects a negative sample of appropriate hardness through two sampling principles, improving the model embedding space representation capability, which in turn leads to improved model recommendation accuracy. Experimental results on the Urban-Book and Yelp2018 datasets show that the proposed algorithm significantly improves all the metrics compared to the state-of-the-art model, which proves the effectiveness and sophistication of the algorithm in different scenarios.

关键词： recommendation algorithms comparative learning negative sampling pruning strategies

来源：评论

学校读者我要写书评

暂无评论

Auditing YouTube's recommendation system for ideologically congenial, extreme, and problematic recommendations

引用

PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA 2023年第50期120卷 e2213020120-e2213020120页

作者： Haroon, Muhammad Wojcieszak, Magdalena Chhabra, Anshuman Liu, Xin Mohapatra, Prasant Shafiq, Zubair Univ Calif Davis Dept Comp Sci Davis CA 95616 USA Univ Calif Davis Dept Commun Davis CA 95616 USA

algorithms of social media platforms are often criticized for recommending ide-ologically congenial and radical content to their users. Despite these concerns, evidence on such filter bubbles and rabbit holes of radicalization is inconclusive. We conduct an audit of the platform using 100,000 sock puppets that allow us to systematically and at scale isolate the influence of the algorithm in recommendations. We test 1) whether recommended videos are congenial with regard to users' ideology, especially deeper in the watch trail and whether 2) recommendations deeper in the trail become progressively more extreme and come from problematic channels. We find that YouTube's algorithm recommends congenial content to its partisan users, although some moderate and cross-cutting exposure is possible and that congenial recommendations increase deeper in the trail for right-leaning users. We do not find meaningful increases in ideological extremity of recommendations deeper in the trail, yet we show that a growing proportion of recommendations comes from channels categorized as problematic (e.g., "IDW," "Alt-right," "Conspiracy," and "QAnon"), with this increase being most pronounced among the very-right users. Although the proportion of these problematic recommendations is low (max of 2.5%), they are still encountered by over 36.1% of users and up to 40% in the case of very-right ***'s algorithm is often accused of putting users in filter bubbles and generating rabbit holes of radicalization. However, evidence on these issues is inconclusive. We conduct a systematic audit of the platform using 100,000 sock puppets that allow us to isolate the influence of the algorithm in recommendations to ideologically congenial and increasingly extreme and problematic videos. YouTube's algorithm recommends ideologically congenial content to partisan users, and congenial recommendations increase deeper in the recommendation trail for right-leaning users. Although we do n

关键词： YouTube congeniality recommendation algorithms extremity ideology

来源：评论

学校读者我要写书评

暂无评论

Collaborative Filtering Algorithm Based on the Preference List in the Digital Library 1

引用

3rd International Conference on Information Computing and Applications (ICICA 2012)

作者： Gao Feng Liu Zhiyou Liu Huixin Hebei United Univ Affiliated Hosp Tangshan Hebei Peoples R China Hebei United Univ Lib Tangshan Hebei Peoples R China

ISBN: (数字)9783642340383

ISBN: (纸本)9783642340376

Traditional digital library services are built on the explicit needs of the user. The information needs of the user with specific digital resources associated with key words. This is a passive information retrieval service, only to meet the basic needs of users, not through the user's interest in reading and reading goals to provide targeted services. Through the user log information, we can mine user preferences of different digital resources, to establish a list of user preferences. By the user preference list of association and similarity computation, this paper presents a collaborative filtering algorithm based on user preference list to help readers discover more useful knowledge and information in the mass of digital resources in the digital library system.

关键词： Preference list Collaborative filtering recommendation algorithms Digital libraries

来源：评论

学校读者我要写书评

暂无评论

On the Joint Recovery of Community Structure and Community Features 56

On the Joint Recovery of Community Structure and Community F...

引用

56th Annual Allerton Conference on Communication, Control, and Computing (Allerton)

作者： Yoon, Jisang Lee, Kangwook Suh, Changho Kakaocorp R&D Ctr Sungnam South Korea Korea Adv Inst Sci & Technol Sch EE Daejeon South Korea

ISBN: (纸本)9781538665961

We study the problem of recovering both K communities and their features from a labeled graph observation. We assume that the edges of an observed graph are generated as per the symmetric Stochastic Block Model (SBM), and that the label of each node is a noisy and partially-observed version of the corresponding community feature. We characterize the information-theoretic limit of this problem, and then propose a computationally efficient algorithm that achieves the information-theoretic limit.

关键词： Community recovery Stochastic Block Model Matrix completion recommendation algorithms Signal processing

来源：评论

学校读者我要写书评

暂无评论

Opportunistic Multi-aspect Fairness through Personalized Re-ranking 20

Opportunistic Multi-aspect Fairness through Personalized Re-...

引用

Proceedings of the 28th ACM Conference on User Modeling, Adaptation and Personalization (UMAP)

作者： Sonboli, Nasim Eskandanian, Farzad Burke, Robin Liu, Weiwen Mobasher, Bamshad Univ Colorado Boulder CO 80309 USA DePaul Univ Chicago IL USA Chinese Univ Hong Kong Shatin Hong Kong Peoples R China

ISBN: (纸本)9781450368612

As recommender systems have become more widespread and moved into areas with greater social impact, such as employment and housing, researchers have begun to seek ways to ensure fairness in the results that such systems produce. This work has primarily focused on developing recommendation approaches in which fairness metrics are jointly optimized along with recommendation accuracy. However, the previous work had largely ignored how individual preferences may limit the ability of an algorithm to produce fair recommendations. Furthermore, with few exceptions, researchers have only considered scenarios in which fairness is measured relative to a single sensitive feature or attribute (such as race or gender). In this paper, we present a re-ranking approach to fairness-aware recommendation that learns individual preferences across multiple fairness dimensions and uses them to enhance provider fairness in recommendation results. Specifically, we show that our opportunistic and metric-agnostic approach achieves a better trade-off between accuracy and fairness than prior re-ranking approaches and does so across multiple fairness dimensions.

关键词： recommendation algorithms personalized reranking personalized diversity post processing fairness-aware reranking fairness opportunistic fairness reranking provider-side fairness fairness-aware recommendation algorithms

来源：评论

学校读者我要写书评

暂无评论

The development of media quality in the Digital Age

引用

List Forum fur Wirtschafts- und Finanzpolitik 2024年第3期50卷 183-203页

作者： Barhoumi, Sahra Düsseldorf Institute for Competition Economics Universität Düsseldorf Düsseldorf Germany

The Digital Age has revolutionized the media landscape, reshaping how consumers access and interact with content. Consumers now have access to a wide range of media content, from physical copies to digital downloads and streaming. This has led to content creators detaching from traditional publishing and self-publishing their content on platforms like Wattpad, SoundCloud, and YouTube. The Digital Age has also added more remedies to signal quality to consumers, such as online reviews and personalized recommendations. This paper aims to analyze the impact of the Digital Age on the quality of media content in three different media fields (books, music, and films and series) by conducting an online survey among 245 German-speaking internet users in May 2022. Findings reveal a significant increase in access to content creators and content availability for consumers in the Digital Age. Moreover, the Digital Age introduces mechanisms that assist consumers in evaluating content quality, mitigating the adverse selection problem. The study concludes that media content quality in the Digital Age surpasses that of the Pre-Digital Age. For an enhanced consumer experience in the field of books, media providers should reconsider their algorithms, prioritizing content over purchasing patterns. Caution is advised in reinforcing intellectual property rights, emphasizing the need for judicious application to preserve content creation incentives. © The Author(s) 2024.

关键词： Content consumption Content creation Online platforms recommendation algorithms Self-publishing Survey

来源：评论

学校读者我要写书评

暂无评论

Improving Social Collaborations in Virtual Learning Environments

Improving Social Collaborations in Virtual Learning Environm...

引用

OnTheMove (OTM) Federated Conference

作者： Gonzalez, Daniel Motz, Regina Tansini, Libertad Univ Republica Fac Ingn Montevideo Uruguay

ISBN: (纸本)9783319261386;9783319261379

This paper shows how to use social collaborations in virtual learning environments to motivate students in their learning process through recommendations of learning objects between peers. A hybrid recommender system is proposed and implemented in order to provide personalized recommendations for university students. An important contribution of this work is to show how to incorporate different recommendation approaches and techniques in order to produce useful recommendations in a real life scenario. Preliminary results indicate that information about groups of students as well as demographic information and previous evaluations of learning objects, processed with a combination of recommendation algorithms, show to be useful for the generation of personalized recommendations of learning objects.

关键词： Learning objects Hybrid recommender systems recommendation algorithms Virtual learning environments

来源：评论

学校读者我要写书评

暂无评论

To Become Fashionable: A Brief Review of Outfit Compatibility

To Become Fashionable: A Brief Review of Outfit Compatibilit...

引用

IEEE Conference on Telecommunications, Optics and Computer Science (TOCS)

作者： Feng, Ruining Sun Yat Sen Univ Sch Data & Comp Sci Guangzhou Peoples R China

ISBN: (纸本)9781728181172

How to dress fashionably has become a major concern for people nowadays. In order to match a fashionable outfit, we need to make our outfit as compatible as possible. Thus, we need to exploit an effective outfit matching scheme and apply it to help people match their own clothes. Outfit compatibility is an emergent computer vision field which tackles this problem. It refers to whether multiple fashion items look good if worn together. Recently, researchers have proposed different compatibility algorithms and recommendation algorithms to make an outfit as compatible as possible and fulfill the users' personal need. In this paper, we present a detailed overview of major compatibility algorithms, such as neural network, and major recommendation algorithms, such as collaborative filtering in outfit compatibility. This paper evaluates the models based on different tasks and datasets and concludes current applications, challenges and possible solutions in outfit compatibility.

关键词： Outfit compatibility Compatibility algorithms recommendation algorithms Neural Networks Graph Deep Learning Collaborative filtering Compatibility prediction

来源：评论

学校读者我要写书评

暂无评论

A Health Management Platform based on CGM Device 3

A Health Management Platform based on CGM Device

引用

IEEE 3rd Information Technology, Networking, Electronic and Automation Control Conference (ITNEC)

作者： Zhang, Liangjie Zhong, Shigen Fujian Agr & Forestry Univ Dongfang Coll Dept Informat Engn Fuzhou Fujian Peoples R China Xiamen Nanyang Coll Sch Informat Engn Xiamen Fujian Peoples R China

ISBN: (纸本)9781538662434

People are more and more concerned about the health of modern life. Diabetes is the enemy of health. Forecast to 2030, diabetes will become the seventh largest cause of death. Diabetes is an important public health problem;it is one of the top four non communicable diseases in the world. So, how to detect blood sugar effectively is a question, how to prevent and treat diabetes is particularly important. In this paper, we used Internet of Things, big data, CGM, recommendation algorithms and other technologies to build a diet-based, drugs supplemented diabetes health management platform. Patients wear CGM device to detect blood sugar and other data, upload to the mobile app. Then mobile app uploads combined data to the diabetes health management platform. The platform through big data analysis to provide recommended diet and early warning, and relatives of the patients can understand the state of their parents anytime and anywhere.

关键词： diabetes CGM health management platform diet recommendation recommendation algorithms

来源：评论

学校读者我要写书评

暂无评论

Vector Space Model Embedding for Recomender System Neural Networks 4

Vector Space Model Embedding for Recomender System Neural Ne...

引用

4th International Conference on Systems and Informatics (ICSAI)

作者： Wang, Yahui Craig, Paul Xian Jiaotong Liverpool Univ Dept Comp Sci & Software Engn Suzhou Peoples R China

ISBN: (纸本)9781538611074

Research shows that recommendation algorithms such as Collaborative Filtering (CF) can be enhanced using neural network (NN) to make more accurate recommendations. This project proposes an adaptive recommendation model based on NN to quickly access different items. A vector space embedding method is used to vectorize users and items before a Deep Neural Network (DNN) rating prediction network model is used to predict users' rating behavior. Multi-domain datasets are utilized in experiments to evaluate results and compared our method with traditional recommendation algorithms. Results show that the improved NN based recommendation model is more effective and achieves a higher score in both similarity calculation and predication.

关键词： recommendation algorithms Versatility Neural Network Embedding Method Rating Prediction Network

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：