Introducing Context Information in Lifelong Sequential Modeling using Temporal Convolutional Networks

2025-02-18

AI Summary

In lifelong sequential modeling (LSM) for social media recommendation, existing methods neglect local contextual information between adjacent items, resulting in coarse-grained interest representations. To address this, we propose a context-aware Temporal Convolutional Network (TCN)-enhanced framework. Our key contributions are: (1) the first integration of TCN into sequence-level attention to explicitly capture local temporal dependencies; (2) a multi-scale TCN-attention cascading module that fuses contextual information across varying receptive fields; and (3) a lightweight, user-profile-driven dynamic filter generation subnetwork for personalized context modulation. Extensive experiments on multiple public and private benchmarks demonstrate consistent and significant improvements over state-of-the-art methods. Online A/B testing further confirms substantial gains in both click-through rate (CTR) and average session dwell time.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Machine Learning: Time-Series/Data Streams

Application Category

• User Modeling, Personalization and Recommendation: Social recommender systems and personalization for the social Web
• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search
• Social Networks and Social Media: Social mining and social search on the Web

Abstract

The importance of lifelong sequential modeling (LSM) is growing in the realm of social media recommendation systems. A key component in this process is the attention module, which derives interest representations with respect to candidate items from the sequence. Typically, attention modules function in a point-wise fashion, concentrating only on the relevance of individual items in the sequence to the candidate item. However, the context information in the neighboring items that is useful for more accurately evaluating the significance of each item has not been taken into account. In this study, we introduce a novel network which employs the Temporal Convolutional Network (TCN) to generate context-aware representations for each item throughout the lifelong sequence. These improved representations are then utilized in the attention module to produce context-aware interest representations. Expanding on this TCN framework, we present a enhancement module which includes multiple TCN layers and their respective attention modules to capture interest representations across different context scopes. Additionally, we also incorporate a lightweight sub-network to create convolution filters based on users' basic profile features. These personalized filters are then applied in the TCN layers instead of the original global filters to produce more user-specific representations. We performed experiments on both a public dataset and a proprietary dataset. The findings indicate that the proposed network surpasses existing methods in terms of prediction accuracy and online performance metrics.

Problem

• Enhance lifelong sequential modeling accuracy
• Utilize context-aware representations in attention modules
• Generate user-specific representations using personalized filters

Innovation

• Temporal Convolutional Networks for context
• Multiple TCN layers enhance representations
• Personalized filters based on user profiles