Bipartite Graph Embedding via Mutual Information Maximization

Cao, Jiangxia; Lin, Xi-Jie; Guo, Shu; Liu, Luchen; Liu, Tingwen; Wang, Bin

doi:10.1145/3437963.3441783

Cited by 82 publications

(40 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…sedenions S ∈ H 16 , and trigintaduonions T ∈ H 32 . To verify the performance of CDRec, we compare it with the following methods, covering the classical models (MF [17], DMF [39], and NeuCF [14]), the GCN-based models (GCMC [3], NGCF [38], LightGCN [11], and BiGI [4]), and the hypercomplex-valued models (CCF [42], QCF [42], QFM [6], QNFM [6], and QGNN [23]). The characteristics of the comparison methods are included in supplement C.2.…”

Section: Comparison Methodsmentioning

confidence: 99%

See 1 more Smart Citation

CDRec: Cayley-Dickson Recommender

Li¹,

Yang²,

Huo³

et al. 2021

Preprint

View full text Add to dashboard Cite

In this paper, we propose a recommendation framework named Cayley-Dickson Recommender. We introduce Cayley-Dickson construction which uses a recursive process to define hypercomplex algebras and their mathematical operations. We also design a graph convolution operator to learn representations in the hypercomplex space. To the best of our knowledge, it is the first time that Cayley-Dickson construction and graph convolution techniques have been used in hypercomplex recommendation. Compared with the stateof-the-art recommendation methods, our method achieves superior performance on real-world datasets.

show abstract

Section: Comparison Methodsmentioning

confidence: 99%

“…Learning high-quality user and item representations forms the crux of CF. Most CF models primarily learn representations in the real number system, ignoring the great potential of alternative algebra systems [3,4,11,14,38].…”

Section: Introductionmentioning

confidence: 99%

CDRec: Cayley-Dickson Recommender

Li¹,

Yang²,

Huo³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…It is usually applied to measure the correlation coefficient among random variables [25]. More specifically, it quantifies the amount of information obtained about one random variable by observing the other [26]. Thus, the MI for variable X and Y can be calculated using the following formula:…”

Section: B Point-wise Mutual Informationmentioning

confidence: 99%

“…To decompile the Android apps, AndroGuard was used to extract DPs from the manifest file and invocation relations. To facilitate the analysis of all apps, we executed Dypermin to obtain API mappings for multiple Android versions (API level [16][17][18][19][20][21][22][23][24][25][26][27][28]. Dypermin extracts API mappings from the SDK source code via Java annotations.…”

Section: Empirical Evaluationmentioning

confidence: 99%

Enhancing Fidelity of Description in Android Apps With Category-Based Common Permissions

Chen

Khan

et al. 2021

IEEE Access

View full text Add to dashboard Cite

Application description analysis is applied for various purposes in software engineering domains. Besides the inherent challenges from the ambiguity of natural language, sparse permission semantics raise the difficulties of predicting functionalities and permission usages from app descriptions. More specifically, the functionalities common to the app's category are intentionally abbreviated by developers due to the limited number of characters, and the permissions are often over-claimed. These are the main reasons that cause false positives in predicting permissions from app descriptions. Such unmentioned permissions can only be detected as suspicious in previous studies where effective assistance for developers in refining app descriptions and preventing potential security risks is not provided. In this paper, we propose the FideDroid, a framework to identify category-based common permissions to offset those essential functionalities while assessing the fidelity of app descriptions. Our framework augments the labeled dataset of app descriptions to improve the prediction of permissions. FideDroid compares inferred permissions with used ones to reveal the suspicious and unnecessary permissions based on the prediction. It helps developers to refine app descriptions and maintain permission usages. In our experiments on large real-world apps, we analyzed and revealed that the category-based common permissions may cover more unmentioned functionalities without considering all possible permissions during app description analysis. In addition, we discovered three factors causing the inconsistency between descriptions and permission usages to be: 1) human interventions in writing description; 2) bad practices on permission usages; and 3) prolific developers. These findings will facilitate developers to refine app descriptions and optimize permission usages in the apps.

show abstract

“…Recommender systems have been successfully applied into a great number of online services for providing precise personalized recommendations [52]. Traditional matrix factorization (MF) models and popular deep learning models are among the most widely used techniques, predicting which items a user will be interested in via learning the low-dimensional representations of users and items [7,17,23,27,53]. These models typically work well when adequate user interactions are available, but suffer from cold-start problems.…”

Section: Introductionmentioning

confidence: 99%

Task-adaptive Neural Process for User Cold-Start Recommendation

Lin

Zhou

et al. 2021

Proceedings of the Web Conference 2021

Self Cite

View full text Add to dashboard Cite

User cold-start recommendation is a long-standing challenge for recommender systems due to the fact that only a few interactions of cold-start users can be exploited. Recent studies seek to address this challenge from the perspective of meta learning, and most of them follow a manner of parameter initialization, where the model parameters can be learned by a few steps of gradient updates. While these gradient-based meta-learning models achieve promising performances to some extent, a fundamental problem of them is how to adapt the global knowledge learned from previous tasks for the recommendations of cold-start users more effectively.In this paper, we develop a novel meta-learning recommender called task-adaptive neural process (TaNP). TaNP is a new member of the neural process family, where making recommendations for each user is associated with a corresponding stochastic process. TaNP directly maps the observed interactions of each user to a predictive distribution, sidestepping some training issues in gradient-based meta-learning models. More importantly, to balance the trade-off between model capacity and adaptation reliability, we introduce a novel task-adaptive mechanism. It enables our model to learn the relevance of different tasks and customize the global knowledge to the task-related decoder parameters for estimating user preferences. We validate TaNP on multiple benchmark datasets in different experimental settings. Empirical results demonstrate that TaNP yields consistent improvements over several state-ofthe-art meta-learning recommenders. CCS CONCEPTS• Information systems → Recommender systems; • Computing methodologies → Neural networks.

show abstract

Bipartite Graph Embedding via Mutual Information Maximization

Cited by 82 publications

References 13 publications

CDRec: Cayley-Dickson Recommender

CDRec: Cayley-Dickson Recommender

Enhancing Fidelity of Description in Android Apps With Category-Based Common Permissions

Task-adaptive Neural Process for User Cold-Start Recommendation

Contact Info

Product

Resources

About