Attentive Recurrent Social Recommendation

Sun, Peijie; Wu, Le; Wang, Meng

doi:10.1145/3209978.3210023

Cited by 115 publications

(57 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MF models user preferences and item properties by factorizing the user-item interaction matrix into two low-dimensional latent matrices. Recently, numerous deep learning techniques (e.g., MLP [21], CNNs [22], RNNs [23], GNNs [24], Autoencoder [25], and the Attention Mechanism [26]) have been introduced into RS. Compared to traditional RS, deep learning based RS is able to model the nonlinearity of data correlations and learn the underlying complex feature representations [18].…”

Section: Related Work 21 Recommender Systems (Rs)mentioning

confidence: 99%

Attacking Recommender Systems with Augmented User Profiles

Lin

Chen

et al. 2020

Proceedings of the 29th ACM International Conference on Information &Amp; Knowledge Management

View full text Add to dashboard Cite

Recommendation Systems (RS) have become an essential part of many online services. Due to its pivotal role in guiding customers towards purchasing, there is a natural motivation for unscrupulous parties to spoof RS for profits. In this paper, we study the shilling attack: a subsistent and profitable attack where an adversarial party injects a number of user profiles to promote or demote a target item. Conventional shilling attack models are based on simple heuristics that can be easily detected, or directly adopt adversarial attack methods without a special design for RS. Moreover, the study on the attack impact on deep learning based RS is missing in the literature, making the effects of shilling attack against real RS doubtful. We present a novel Augmented Shilling Attack framework (AUSH) and implement it with the idea of Generative Adversarial Network. AUSH is capable of tailoring attacks against RS according to budget and complex attack goals, such as targeting a specific user group. We experimentally show that the attack impact of AUSH is noticeable on a wide range of RS including both classic and modern deep learning based RS, while it is virtually undetectable by the state-ofthe-art attack detection model.

show abstract

Section: Related Work 21 Recommender Systems (Rs)mentioning

confidence: 99%

Attacking Recommender Systems with Augmented User Profiles

Lin

Chen

et al. 2020

Proceedings of the 29th ACM International Conference on Information &Amp; Knowledge Management

View full text Add to dashboard Cite

show abstract

“…Attention mechanism is widely used in many neural network based tasks, such as machine translation [4] and image captioning [53]. Recently, the attention mechanism is also widely used for recommender systems [19], [52], [43], [41]. Given the classical collaborative filtering scenario with user-item interaction behavior, NAIS extended the classical item based recommendation models by distinguishing the importance of different historical items in a user profile [19].…”

Section: Related Workmentioning

confidence: 99%

A Hierarchical Attention Model for Social Contextual Image Recommendation

Chen

Hong

et al. 2020

IEEE Trans. Knowl. Data Eng.

Self Cite

View full text Add to dashboard Cite

Image based social networks are among the most popular social networking services in recent years. With tremendous images uploaded everyday, understanding users' preferences on user-generated images and making recommendations have become an urgent need. In fact, many hybrid models have been proposed to fuse various kinds of side information (e.g., image visual representation, social network) and user-item historical behavior for enhancing recommendation performance. However, due to the unique characteristics of the user generated images in social image platforms, the previous studies failed to capture the complex aspects that influence users' preferences in a unified framework. Moreover, most of these hybrid models relied on predefined weights in combining different kinds of information, which usually resulted in sub-optimal recommendation performance. To this end, in this paper, we develop a hierarchical attention model for social contextual image recommendation. In addition to basic latent user interest modeling in the popular matrix factorization based recommendation, we identify three key aspects (i.e., upload history, social influence, and owner admiration) that affect each user's latent preferences, where each aspect summarizes a contextual factor from the complex relationships between users and images. After that, we design a hierarchical attention network that naturally mirrors the hierarchical relationship (elements in each aspects level, and the aspect level) of users' latent interests with the identified key aspects. Specifically, by taking embeddings from state-of-the-art deep learning models that are tailored for each kind of data, the hierarchical attention network could learn to attend differently to more or less content. Finally, extensive experimental results on real-world datasets clearly show the superiority of our proposed model. !

show abstract

“…For example, the dualflow attention mechanism [21] is used to perform feature refinement modeling on both dynamic and static aspects. The Transformer can even replace the RNN for feature capture [41]. High-order attention is the new attention-grabbing mechanism proposed by Chen et al [42] , which is also a new method for high-order expression in fine-grained image features.…”

Section: Related Workmentioning

confidence: 99%

HHA: An Attentive Prediction Model for Academic Abnormality

et al. 2020

View full text Add to dashboard Cite

Warning students with poor performance in advance based on historical academic data, namely, the academic abnormality prediction is important task in education. The majority of existing methods focus on digging out abnormal complex clues from historical data, while ignoring two basic considerations:(1)these works fail to handle unrecorded/missing data when this part is sparse; (2)these works ignore the complex relationship between courses. The different courses are used as the attention weight vector for abnormality prediction, but they do not notice the mutual influence between courses. To this end, we contribute a Hybrid Neural Network Model based on High-Order Attention Mechanism, called HHA, to address the academic abnormality prediction problem. Specifically, we first exploit Generative Adversarial Network(GAN) to mine hidden factors in the unrecorded/missing data reasonably by simulating student behavior. Thereafter, a high-order attention mechanism is proposed to measure the importance of course and course combination. Lastly, a multi-layer projection abstracts feature and classifies whether the student is abnormal. By experimenting on real-world dataset, we demonstrate the effectiveness and rationality of our proposed model. INDEX TERMS academic abnormality prediction, high-order attention mechanism, hybrid neural network, generative adversarial network This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.

show abstract

Attentive Recurrent Social Recommendation

Cited by 115 publications

References 37 publications

Attacking Recommender Systems with Augmented User Profiles

Attacking Recommender Systems with Augmented User Profiles

A Hierarchical Attention Model for Social Contextual Image Recommendation

HHA: An Attentive Prediction Model for Academic Abnormality

Contact Info

Product

Resources

About