SessionRec: Next Session Prediction Paradigm For Generative Sequential Recommendation

TL;DR

SessionRec reframes generative sequential recommendation from next-item to next-session prediction, aligning model objectives with real-world session-based user behavior. It introduces a hierarchical, session-focused architecture (Embedding Module, ISE, SSE, Session Prediction Module) that reduces computational complexity and enables multi-item predictions in the next session, while integrating a rank loss to improve intra-session ordering. The approach is model-agnostic and demonstrates strong offline gains on public datasets and robust online impact in Meituan’s app, with clear scaling laws and efficiency Benefits from incorporating negative interactions. Overall, SessionRec offers a scalable, industrially viable foundation for generative recommendation systems that can outperform traditional cascaded retrieval/ranking pipelines.

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.

Figures (6)

• Figure 1: Next Item Prediction Paradigm (NIPP) vs. our proposed Next Session Prediction Paradigm (NSPP).
• Figure 2: Overall architecture of our proposed SessionRec.
• Figure 3: Performance and Traning Time per Epoch
• Figure 4: Impact of ranking loss weight on model performance improvement.
• Figure 5: Scaling laws of SessionRec with data volume.