A Simple Convolutional Generative Network for Next Item Recommendation

By Fajie Yuan, Alexandros Karatzoglou, Ioannis Arapakis, Joemon M. Jose, Xiangnan He. Web Search and Data Mining, 2018

670 citations recorded by the source corpus.

Convolutional Neural Networks (CNNs) have been recently introduced in the domain of session-based next item recommendation. An ordered collection of past items the user has interacted with in a session (or sequence) are embedded into a 2-dimensional latent matrix, and treated as an image. The convolution and pooling operations are then applied to the mapped item embeddings. In this paper, we first examine the typical session-based CNN recommender and show that both the generative model and network architecture are suboptimal when modeling long-range dependencies in the item sequence. To address the issues, we introduce a simple, but very effective generative model that is capable of learning high-level representation from both short- and long-range item dependencies. The network architecture of the proposed model is formed of a stack of holed convolutional layers, which can efficiently increase the receptive fields without relying on the pooling operation. Another contribution is the effective use of residual block structure in recommender systems, which can ease the optimization for much deeper networks. The proposed generative model attains state-of-the-art accuracy with less training time in the next item recommendation task. It accordingly can be used as a powerful recommendation baseline to beat in future, especially when there are long sequences of user feedback.

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

View Paper: https://www.semanticscholar.org/paper/5babbf2ed9f6e36b83ed246927b270db320fe866

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. Xiangnan He
2. Wayne Xin Zhao
3. Xiang Wang
4. Hongzhi Yin
5. Ji-rong Wen
6. Junliang Yu
7. Meng Wang
8. Yupeng Hou

In Computer Vision and Pattern Recognition

1. Xiangnan He
2. Xiang Wang
3. Wayne Xin Zhao
4. Ji-rong Wen
5. Fuli Feng
6. Meng Wang
7. Kun Zhou
8. Fajie Yuan

In Information Systems

1. Xiangnan He
2. Xiang Wang
3. Wayne Xin Zhao
4. Hongzhi Yin
5. Ji-rong Wen
6. Junliang Yu
7. Meng Wang
8. Xin Xia

Citations in the corpus

• Where to Go Next for Recommender Systems? ID- vs. Modality-based Recommender Models Revisited (2023, methodology): “Besides, most of them only discussed the DSSM architecture (or other two-tower variants) without considering the more powerful sequence-to-sequence (seq2seq) training approach (e.g., used in SASRec andNextItNet [78]).”
• A Survey on Accuracy-Oriented Neural Recommendation: From Collaborative Filtering to Information-Rich Recommendation (2021, methodology): “Another kind of models utilize the 3D convolutional networks for recommendation, which defines the recommendation problem as [51], [246]: ðSu t L; : .”
• TALLRec: An Effective and Efficient Tuning Framework to Align Large Language Model with Recommendation (2023, methodology): “Moving forward, we plan to explore more efficient methods to activate the recommendation ability of larger models and tune LLMs to handle multiple recommendation tasks simultaneously.”
• Handling Information Loss of Graph Neural Networks for Session-based Recommendation (2020, methodology): “• NextItNet [27] is a CNN-based method for next-item recommendation.”
• Sequential Recommender Systems: Challenges, Progress and Prospects (2019, background): “g SRSs built on Markov-chains, factorization machines or RNN can only handle the point-wise dependencies but are not good at modelling and capturing collective dependencies. Only quite few works like [17; 26] have attempted to address such a challenge by employing the strength of convulotional neura”
• Self-Supervised Hypergraph Convolutional Networks for Session-based Recommendation (2020, background): “2018) recently have drawn increasing attention and their applications in SBR also have shown promising results (Wang et al. 2020b,c; Yuan et al. 2019; Chen and Wong 2020).”
• One Person, One Model, One World: Learning Continual User Representation without Forgetting (2020, methodology): “dilated) convolutional network (TCN) NextItNet as the pretrained base model given its linear complexity and superb performance in literature [33, 36, 41, 43].”
• Text Is All You Need: Learning Language Representations for Sequential Recommendation (2023, methodology): “Various methods have been proposed to improve the performance of sequential recommendation, including Markov Chains [9, 25], RNN/CNN models [11, 17, 28, 34] and self-attentive models [14, 19, 27].”
• Filter-enhanced MLP is All You Need for Sequential Recommendation (2022, methodology): “It is widely used as assistant modules to couple with CNN [4], RNN [36] and selfattention network [57] to construct deep models.”
• RecBole: Towards a Unified, Comprehensive and Efficient Framework for Recommendation Algorithms (2020, neutral): “NextItnet [81] WSDM 2019 YOOCHOOSE, LastFM”
• AMER: Automatic Behavior Modeling and Interaction Exploration in Recommender System (2020, methodology): “The MLP model of all baseline methods is fixed to [200, 80], while the activation is set to Dice in DIN and DIEN and set to ReLU in other methods.”
• Parameter-Efficient Transfer from Sequential Behaviors for User Modeling and Recommendation (2020, methodology): “A large body of research has clearly shown that these interaction sequences can be used to model the item preferences of the users [8, 18, 31, 35–37].”
• Search-based User Interest Modeling with Lifelong Sequential Behavior Data for Click-Through Rate Prediction (2020, background): “Recently, a series of works, which we called User Interest Model, focus on learning the representation of latent user interest from historical behaviors, using different neural network architecture such as CNN[14, 17], RNN[5, 19], Transformer[3, 13] and Capsule[6], etc. DIN[20] emphasizes that user ”
• A Survey on Session-based Recommender Systems (2019, methodology): “, long-range dependencies [141] between two interactions that are far from each other in a session or high-order dependencies [119] across multiple interactions in a session.”
• Next-item Recommendation with Sequential Hypergraphs (2020, methodology): “good success in next-item recommendation which aims to predict a user’s next actions based on the sequential interactions in the past [22, 25, 30, 35, 37, 39, 44].”
• Hierarchical Gating Networks for Sequential Recommendation (2019, background): “• NextItNet, the next item recommendation net [42], applies dilated convolutional neural networks to increase the receptive fields without relying on the pooling operation.”
• Adapting Large Language Models by Integrating Collaborative Semantics for Recommendation (2023, methodology): “CNN [4], [5], GNN [6], [7], and Transformer [2], [8].”