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

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 code is available at https://github.com/USTC-StarTeam/RSIR.

Machine learning systems usually improve when they are trained with more high-quality examples. However, in recommender systems, such as those used for movies, music, or online shopping, useful user feedback is often very limited: most users interact with only a tiny fraction of all available items. This lack of data makes it hard for recommendation models to learn reliable user preferences. This paper introduces Recursive Self-Improving Recommendation (RSIR), a method that helps a recommender system improve itself without using extra outside data or a separate expert model. The key idea is simple: the current model creates possible user behavior examples, checks whether these examples are realistic and consistent with what users seem to like, and then trains a new version of the model using both the original data and the selected new examples. Repeating this process gradually improves the model. Our experiments show that RSIR consistently improves recommendation quality across different datasets and model types. Even smaller or weaker models can benefit from this process. Overall, RSIR offers a practical way to reduce the impact of limited user data and build better recommender systems.