RecRec: Recursive Refinement for Sequential Recommendation

2026-07-11

AI Summary

This work addresses the limitations of conventional sequential recommendation models, which rely on single-pass encoding and deep architectures yet lack iterative refinement mechanisms. To overcome this, the authors propose RecRec, a novel model that employs a shared recursive module to iteratively update compact latent states. Crucially, RecRec incorporates an evidence-anchored correction mechanism grounded in the original interaction context, effectively mitigating semantic drift during recursion and thereby enhancing both model stability and representational capacity. Despite its lightweight architecture, RecRec achieves state-of-the-art performance across three benchmark datasets, matching or surpassing existing approaches—including advanced sequential models, graph neural networks, and reasoning-augmented recommenders—while utilizing only 3.9M to 14M parameters.

Abstract

Sequential recommender systems typically infer user preferences through single-pass encoding of interaction histories without iterative refinement, relying on increasingly deep architectures to capture complex patterns. In this work, we revisit sequential recommendation from a recursive inference perspective: can user preferences be modeled as a persistent latent state that is recursively refined? We propose RecRec (Recursive Recommendation), a lightweight model that maintains a compact latent state and updates it through a shared recursive module conditioned on interaction evidence. Unlike prior recursive models, RecRec introduces an evidence-anchored correction mechanism that stabilizes refinement by grounding each update in the original interaction context, preventing semantic drift during deep recursive reasoning. Experiments on three benchmark datasets under standard evaluation protocols show that RecRec matches or outperforms state-of-the-art sequential, graph-based, and reasoning-enhanced recommenders while using only 3.9M to 14M parameters. Ablation studies demonstrate that both recursive refinement and the evidence-anchored correction gate contribute significantly to performance, highlighting the effectiveness of recursive latent inference as a scalable alternative to deeper or language-based architectures. Code is available at https://anonymous.4open.science/r/RecRec-6B67/README.md.

Problem

• sequential recommendation
• recursive refinement
• latent state
• user preference modeling
• semantic drift

Innovation

• recursive refinement
• sequential recommendation
• latent state modeling
• evidence-anchored correction
• lightweight architecture