S-Walk: Accurate and Scalable Session-based Recommendation with Random Walks

  title={S-Walk: Accurate and Scalable Session-based Recommendation with Random Walks},
  author={Minjin Choi and Jinhong Kim and Joo-Yeon Lee and Hyunjung Shim and Jongwuk Lee},
  journal={Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining},
  • Minjin Choi, Jinhong Kim, Jongwuk Lee
  • Published 4 January 2022
  • Computer Science
  • Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining
Session-based recommendation (SR) predicts the next items from a sequence of previous items consumed by an anonymous user. Most existing SR models focus only on modeling intra-session characteristics but pay less attention to inter-session relationships of items, which has the potential to improve accuracy. Another critical aspect of recommender systems is computational efficiency and scalability, considering practical feasibility in commercial applications. To account for both accuracy andโ€ฆย 

Figures and Tables from this paper

Space-Efficient Random Walks on Streaming Graphs

This work presents Wharf, a system that efficiently stores and updates random walks on streaming graphs and achieves the succinct representation by coupling compressed purely functional binary trees and pairing functions for storing the walks, and efficient walk updates by effectively pruning the walk search space.

Item Similarity Mining for Multi-Market Recommendation

This paper proposes a novel model called M$^3Rec to improve all markets recommendation simultaneously, and argues that mining the similarities among items is the key point of MMR.

Graph Neural Networks for Recommender Systems: Challenges, Methods, and Directions

A comprehensive review of the literature on graph neural network-based recommender systems, following the taxonomy above, and systematically analyze the challenges in graph construction, embedding propagation/aggregation, model optimization, and computation efficiency.

Bilateral Self-unbiased Learning from Biased Implicit Feedback

This paper proposes a novel unbiased recommender learning model, namely BIlateral SElf-unbiased Recommender (BISER), to eliminate the exposure bias of items caused by recommender models and shows that BISER consistently outperforms state-of-the-art unbiasedRecommender models over several datasets.



Session-aware Linear Item-Item Models for Session-based Recommendation

The comprehensive nature of the proposed simple-yet-effective linear models helps improve the quality of session-based recommendation and provides a generalized framework for reflecting different perspectives of session data.

Session-based Recommendation with Graph Neural Networks

In the proposed method, session sequences are modeled as graph-structured data and GNN can capture complex transitions of items, which are difficult to be revealed by previous conventional sequential methods.

Boosting Item-based Collaborative Filtering via Nearly Uncoupled Random Walks

RecWalk is introduced; a novel random walk-based method that leverages the spectral properties of nearly uncoupled Markov chains to provably lift this limitation and prolong the influence of usersโ€™ past preferences on the successive steps of the walkโ€”thereby allowing the walker to explore the underlying network more fruitfully.

Collaborative Self-Attention Network for Session-based Recommendation

A novel solution, Collaborative Self-Attention Network (CoSAN) for session-based recommendation, to learn the session representation and predict the intent of the current session by investigating neighborhood sessions and applies self-attention to learn long-range dependencies between collaborative items and generate collaborative session representation.

Rethinking Item Importance in Session-based Recommendation

A Session-based Recommendation approach with an Importance Extraction Module, i.e., SR-IEM, that considers both a user's long-term and recent behavior in an ongoing session and has a reduced computational complexity.

Fusing Similarity Models with Markov Chains for Sparse Sequential Recommendation

Fossil is proposed, a similarity-based method that outperforms alternative algorithms, especially on sparse datasets, and qualitatively that it captures personalized dynamics and is able to make meaningful recommendations.

A random walk method for alleviating the sparsity problem in collaborative filtering

A novel item-oriented algorithm that first infers transition probabilities between items based on their similarities and models finite length random walks on the item space to compute predictions, and suggests a method to enhance similarity matrices under sparse data as well.

Global Context Enhanced Graph Neural Networks for Session-based Recommendation

A novel approach to exploit item transitions over all sessions in a more subtle manner for better inferring the user preference of the current session, called GCE-GNN, which outperforms the state-of-the-art methods consistently.

Sequence and Time Aware Neighborhood for Session-based Recommendations: STAN

The results suggest that STAN can be considered as a strong baseline for evaluating session-based recommendation algorithms in future, and is even comparable to the recently proposed state-of-the-art deep learning approaches.

Star Graph Neural Networks for Session-based Recommendation

Experiments on two public benchmark datasets show that SGNN-HN can outperform state-of-the-art models in terms of P@20 and MRR@20 for session-based recommendation.