Joint Representation Learning for Top-N Recommendation with Heterogeneous Information Sources

By Yongfeng Zhang, Qingyao Ai, Xu Chen, W. Bruce Croft. 2017

332 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Information Systems

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Yongfeng Zhang
2. Xiangnan He
3. Xu Chen
4. Tat-seng Chua
5. Qingyao Ai
6. Xiang Wang
7. Yi Tay
8. Yixin Cao

In Computer Vision and Pattern Recognition

1. Yongfeng Zhang
2. Xiangnan He
3. Xu Chen
4. Tat-seng Chua
5. Yi Tay
6. Lei Zhu
7. Zhiyong Cheng
8. Mohan Kankanhalli

In Information Systems

1. Xiangnan He
2. Yongfeng Zhang
3. Xu Chen
4. Tat-seng Chua
5. Xiang Wang
6. Yixin Cao
7. Yi Tay
8. Qingyao Ai

Citations in the corpus

• Explainable Recommendation: A Survey and New Perspectives (2018, methodology): “These sentiment-enhanced modeling approach was also leveraged in point-of-interest recommendation Zhao et al. [2015] and social recommendation Ren et al.”
• Explainable Product Search with a Dynamic Relation Embedding Model (2019, background): “The Amazon product dataset 3 is a well-known benchmark for product search and recommendation [6, 57, 71].”
• Meta-learning on Heterogeneous Information Networks for Cold-start Recommendation (2020, methodology): “Some methods also integrate review texts, images or knowledge graphs to further enhance user and item representations [30, 34].”
• MMALFM (2018, methodology): “More recently, [65] integrate images with reviews and ratings in a multimodal deep learning framework for top-n recommendation.”
• Distilling Structured Knowledge into Embeddings for Explainable and Accurate Recommendation (2019, background): “tensions [5, 9, 10]. Their popularity is largely attributed to their superior performance, and also to their flexibility in incorporating both collaborative signals and content-related signals (e.g., [7, 36, 40, 42]). Nowadays, however, when not merely practitioners but also users are increasingly i”
• Dynamic Explainable Recommendation Based on Neural Attentive Models (2019, result): “This observation actually agrees with many previous studies (Li et al. 2016; Zhang et al. 2017; Chen et al. 2016a), and the reason can be that: in our dataset, a small number of parameters are enough for capturing user different behavior patterns, using redundant dimensions will increase the model…”
• Aspect-Aware Latent Factor Model: Rating Prediction with Ratings and Reviews (2018, methodology): “Textual reviews have also been used in deep learning models for recommendation [6, 40, 41, 43].”
• Recommendation as Language Processing (RLP): A Unified Pretrain, Personalized Prompt & Predict Paradigm (P5) (2022, background): “The main work mostly address four lines of research: 1) explainable recommendation [11, 16, 28–30, 71, 73] where NLP models help generating text explanations for a given recommendation; 2) sequential recommendation as language modeling [10, 56, 76] which considers user interaction histories as word ”
• Personalized Fashion Recommendation with Visual Explanations based on Multimodal Attention Network: Towards Visually Explainable Recommendation (2019, result): “This observation is consistent with many previous studies [14, 44], and the reasons can be that: the redundant parameters will increase the model complexity and may lead to the over-fitting problem, which will weaken the model generalization ability.”
• Deep Learning Based Recommender System (2017, background): “One recent work potentially paving the way toward models of this nature is the Joint Representation Learning framework [198].”
• Learning over Knowledge-Base Embeddings for Recommendation (2018, methodology): “ditional recommendation methods based on matrix factorization (BPR [23], BPR-HFT [17], and VBPR [12]), and the other three are deep models for product recommendation (DeepCoNN [33], CKE [29], and JRL [30]). •BPR: The bayesian personalized ranking [23] model is a popular method for top-N recommendati”
• Adversarial Distillation for Efficient Recommendation with External Knowledge (2018, background): “…review or product image, has been widely incorporated into modern recommender systems, because it can not only reveal more comprehensive user/item properties to enhance the recommendation performance, but also provides opportunities to make the recommendations more explainable [6, 7, 49, 50, 52].”
• Visually Explainable Recommendation (2018, background): “• JRL: The joint representation learning model [36] is a state-of-the-art neural recommender, which can leverage multi-model side information for Top-N recommendation.”
• Adversarial Personalized Ranking for Recommendation (2018, methodology): “Owing to its rationality and ease of optimization, BPR has been used in a wide variety of scenarios [6, 7, 37–39 , 41”
• λOpt: Learn to Regularize Recommender Models in Finer Levels (2019, background): “Top-K item recommendation from implicit feedback is a prevalent task in real-world recommender systems [6, 32, 33].”