DenoiseRank: Learning to Rank by Diffusion Models

arXiv cs.AI / 4/25/2026

💬 OpinionIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper proposes DenoiseRank, a learning-to-rank (LTR) approach that reframes ranking as a diffusion-based generative modeling problem.
Instead of using a purely discriminative objective, DenoiseRank adds noise to the relevant labels during the forward diffusion process and then denoises them with respect to query documents in the reverse process to predict the label distribution.
The authors position DenoiseRank as the first method that tackles traditional LTR from a generative perspective and treats diffusion as the core mechanism for ranking.
Experiments on standard benchmark datasets show the method is effective, and the authors suggest it can serve as a benchmark for generative LTR research.

Abstract

Learning to rank (LTR) is one of the core tasks in Machine Learning. Traditional LTR models have made great progress, but nearly all of them are implemented from discriminative perspective. In this paper, we aim at addressing LTR from a novel perspective, i.e., by a deep generative model. Specifically, we propose a novel denoise rank model, DenoiseRank, which noises the relevant labels in the diffusion process and denoises them on the query documents in the reverse process to accurately predict their distribution. Our model is the first to address traditional LTR from generative perspective and is a diffusion method for LTR. Our extensive experiments on benchmark datasets demonstrated the effectiveness of DenoiseRank, and we believe it provides a benchmark for generative LTR task.