Computer Science > Information Retrieval
[Submitted on 15 May 2024 (this version), latest version 7 Jun 2024 (v3)]
Title:Diffusion-based Contrastive Learning for Sequential Recommendation
View PDF HTML (experimental)Abstract:Contrastive learning has been effectively applied to alleviate the data sparsity issue and enhance recommendation this http URL majority of existing methods employ random augmentation to generate augmented views of original sequences. The learning objective then aims to minimize the distance between representations of different views for the same user. However, these random augmentation strategies (e.g., mask or substitution) neglect the semantic consistency of different augmented views for the same user, leading to semantically inconsistent sequences with similar representations. Furthermore, most augmentation methods fail to utilize context information, which is critical for understanding sequence semantics. To address these limitations, we introduce a diffusion-based contrastive learning approach for sequential recommendation. Specifically, given a user sequence, we first select some positions and then leverage context information to guide the generation of alternative items via a guided diffusion model. By repeating this approach, we can get semantically consistent augmented views for the same user, which are used to improve the effectiveness of contrastive learning. To maintain cohesion between the representation spaces of both the diffusion model and the recommendation model, we train the entire framework in an end-to-end fashion with shared item embeddings. Extensive experiments on five benchmark datasets demonstrate the superiority of our proposed method.
Submission history
From: Ziqiang Cui [view email][v1] Wed, 15 May 2024 14:20:37 UTC (1,385 KB)
[v2] Wed, 22 May 2024 13:40:50 UTC (4,080 KB)
[v3] Fri, 7 Jun 2024 04:28:41 UTC (4,703 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.