Abstract:With the rise of e-commerce and short videos, online recommender systems that can capture users' interests and update new items in real-time play an increasingly important role. In both online and offline recommendation, the cold-start problem due to interaction sparsity has been affecting the recommendation effect of cold-start items, which is also known as the long-tail problem of item distribution. Many cold-start scheme based on fine-tuning or knowledge transferring shows excellent performance on offline recommendation. Yet, these schemes are infeasible for online recommendation on streaming data pipelines due to different training method, computational overhead and time constraints. Inspired by the above questions, we propose a model-agnostic recommendation algorithm called Popularity-Aware Meta-learning (PAM), to address the item cold-start problem under streaming data settings. PAM divides the incoming data into different meta-learning tasks by predefined item popularity thresholds. The model can distinguish and reweight behavior-related and content-related features in each task based on their different roles in different popularity levels, thus adapting to recommendations for cold-start samples. These task-fixing design significantly reduces additional computation and storage costs compared to offline methods. Furthermore, PAM also introduced data augmentation and an additional self-supervised loss specifically designed for low-popularity tasks, leveraging insights from high-popularity samples. This approach effectively mitigates the issue of inadequate supervision due to the scarcity of cold-start samples. Experimental results across multiple public datasets demonstrate the superiority of our approach over other baseline methods in addressing cold-start challenges in online streaming data scenarios.

AutoAssign: Automatic Shared Embedding Assignment in Streaming Recommendation.

A Cluster-Based Incremental Recommendation Algorithm on Stream Processing Architecture

Scalable Dynamic Embedding Size Search for Streaming Recommendation

Automated Embedding Dimensionality Searchin Streaming Recommendations

Automated Embedding Size Search in Deep Recommender Systems

Memory-efficient Embedding for Recommendations

Stratified and time-aware sampling based adaptive ensemble learning for streaming recommendations

Embedding in Recommender Systems: A Survey

Neural Memory Streaming Recommender Networks with Adversarial Training.

Online Item Cold-Start Recommendation with Popularity-Aware Meta-Learning

Embedding Factorization Models for Jointly Recommending Items and User Generated Lists

AutoDim: Field-aware Embedding Dimension Searchin Recommender Systems

Clustered Embedding Learning for Recommender Systems

Identifying Users Behind Shared Accounts in Online Streaming Services

RPAF: A Reinforcement Prediction-Allocation Framework for Cache Allocation in Large-Scale Recommender Systems

Embedding Factorization Models for Jointly Recommending User Generated Lists and Their Contained Items

AutoShard: Automated Embedding Table Sharding for Recommender Systems

AdaS&S: a One-Shot Supernet Approach for Automatic Embedding Size Search in Deep Recommender System

Collaborative Knowledge Base Embedding for Recommender Systems

A Distributed Real-Time Recommender System for Big Data Streams

AutoPooling: Automated Pooling Search for Multi-valued Features in Recommendations.