Crops: Improving Dense Retrieval With Cross-perspective Positive Samples In Short-video Search

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,

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyxie2025crops

Related papers

• Towards Efficient And Robust Moment Retrieval System: A Unified Framework For Multi-granularity Models And Temporal Reranking (2025)2.26
• Use What You Have: Video Retrieval Using Representations From Collaborative Experts (2019)0.00
• Learning Video Retrieval Models With Relevance-aware Online Mining (2022)6.07
• Propy: Building Interactive Prompt Pyramids Upon CLIP For Partially Relevant Video Retrieval (2025)1.91
• Improving Video Retrieval By Adaptive Margin (2023)9.92
• Prototypes Are Balanced Units For Efficient And Effective Partially Relevant Video Retrieval (2025)0.00
• Multimodal Contextualized Support For Enhancing Video Retrieval System (2026)0.00
• Prompt-aware Of Frame Sampling For Efficient Text-video Retrieval (2025)0.95