Can Recommender Systems Teach Themselves? A Recursive Self-Improving Framework with Fidelity Control

TL;DR

The paper introduces RSIR, a recursive self-improving framework for recommendation systems that enhances model performance by self-generated data filtering and augmentation, reducing reliance on external data.

Contribution

It presents a novel, model-agnostic recursive self-improvement approach with fidelity control, demonstrating consistent gains across benchmarks and architectures.

Findings

RSIR acts as an implicit regularizer, smoothing the optimization landscape.

RSIR improves performance across multiple benchmarks and architectures.

Weak models can generate effective curricula for stronger models.

Abstract

The scarcity of high-quality training data presents a fundamental bottleneck to scaling machine learning models. This challenge is particularly acute in recommendation systems, where extreme sparsity in user interactions leads to rugged optimization landscapes and poor generalization. We propose the Recursive Self-Improving Recommendation (RSIR) framework, a paradigm in which a model bootstraps its own performance without reliance on external data or teacher models. RSIR operates in a closed loop: the current model generates plausible user interaction sequences, a fidelity-based quality control mechanism filters them for consistency with user's approximate preference manifold, and a successor model is augmented on the enriched dataset. Our theoretical analysis shows that RSIR acts as a data-driven implicit regularizer, smoothing the optimization landscape and guiding models toward more robust solutions. Empirically, RSIR yields consistent, cumulative gains across multiple benchmarks and architectures. Notably, even smaller models benefit, and weak models can generate effective training curricula for stronger ones. These results demonstrate that recursive self-improvement is a general, model-agnostic approach to overcoming data sparsity, suggesting a scalable path forward for recommender systems and beyond. Our anonymized code is available at https://github.com/USTC-StarTeam/RSIR.