SessionRec: Next Session Prediction Paradigm For Generative Sequential Recommendation

TL;DR

SessionRec introduces a session-aware generative recommendation paradigm that aligns better with real-world user behavior, improves ranking effectiveness, and demonstrates scalability and efficiency in large-scale applications.

Contribution

The paper proposes a novel session-aware next-session prediction framework that addresses limitations of traditional methods, with a new hierarchical sequence aggregation and ranking loss to enhance performance.

Findings

SessionRec outperforms existing models on public datasets.

Incorporating rank loss improves ranking effectiveness.

Demonstrated scalability with power-law behavior similar to LLMs.

Abstract

We introduce SessionRec, a novel next-session prediction paradigm (NSPP) for generative sequential recommendation, addressing the fundamental misalignment between conventional next-item prediction paradigm (NIPP) and real-world recommendation scenarios. Unlike NIPP's item-level autoregressive generation that contradicts actual session-based user interactions, our framework introduces a session-aware representation learning through hierarchical sequence aggregation (intra/inter-session), reducing attention computation complexity while enabling implicit modeling of massive negative interactions, and a session-based prediction objective that better captures users' diverse interests through multi-item recommendation in next sessions. Moreover, we found that incorporating a rank loss for items within the session under the next session prediction paradigm can significantly improve the ranking effectiveness of generative sequence recommendation models. We also verified that SessionRec exhibits clear power-law scaling laws similar to those observed in LLMs. Extensive experiments conducted on public datasets and online A/B test in Meituan App demonstrate the effectiveness of SessionRec. The proposed paradigm establishes new foundations for developing industrial-scale generative recommendation systems through its model-agnostic architecture and computational efficiency.