CroPS: Improving Dense Retrieval with Cross-Perspective Positive Samples in Short-Video Search

TL;DR

CroPS introduces a novel training approach for dense retrieval in short-video search by incorporating diverse positive samples from multiple perspectives, reducing bias and improving retrieval performance.

Contribution

The paper proposes CroPS, a data engine that leverages multi-perspective positive samples and a hierarchical label strategy to enhance dense retrieval models.

Findings

Significantly improves retrieval accuracy on Kuaishou Search

Reduces query reformulation rates in live deployment

Achieves superior offline and online performance

Abstract

Dense retrieval has become a foundational paradigm in modern search systems, especially on short-video platforms. However, most industrial systems adopt a self-reinforcing training pipeline that relies on historically exposed user interactions for supervision. This paradigm inevitably leads to a filter bubble effect, where potentially relevant but previously unseen content is excluded from the training signal, biasing the model toward narrow and conservative retrieval. In this paper, we present CroPS (Cross-Perspective Positive Samples), a novel retrieval data engine designed to alleviate this problem by introducing diverse and semantically meaningful positive examples from multiple perspectives. CroPS enhances training with positive signals derived from user query reformulation behavior (query-level), engagement data in recommendation streams (system-level), and world knowledge synthesized by large language models (knowledge-level). To effectively utilize these heterogeneous signals, we introduce a Hierarchical Label Assignment (HLA) strategy and a corresponding H-InfoNCE loss that together enable fine-grained, relevance-aware optimization. Extensive experiments conducted on Kuaishou Search, a large-scale commercial short-video search platform, demonstrate that CroPS significantly outperforms strong baselines both offline and in live A/B tests, achieving superior retrieval performance and reducing query reformulation rates. CroPS is now fully deployed in Kuaishou Search, serving hundreds of millions of users daily.