Pailitao-vl: Unified Embedding And Reranker For Real-time Multi-modal Industrial Search
2026 · Lei Chen, Chen Ju, Xu Chen, et al.
Abstract
In this work, we presented Pailitao-VL, a comprehensive multi-modal retrieval system engineered for high-precision, real-time industrial search. We here address three critical challenges in the current SOTA solution: insufficient retrieval granularity, vulnerability to environmental noise, and prohibitive efficiency-performance gap. Our primary contribution lies in two fundamental paradigm shifts. First, we transitioned the embedding paradigm from traditional contrastive learning to an absolute ID-recognition task. Through anchoring instances to a globally consistent latent space defined by billions of semantic prototypes, we successfully overcome the stochasticity and granularity bottlenecks inherent in existing embedding solutions. Second, we evolved the generative reranker from isolated pointwise evaluation to the compare-and-calibrate listwise policy. By synergizing chunk-based comparative reasoning with calibrated absolute relevance scoring, the system achieves nuanced discriminat
Authors
(none)
Tags
Stats
Related papers
- Unified Interactive Multimodal Moment Retrieval Via Cascaded Embedding-reranking And Temporal-aware Score Fusion (2025)0.00
- Verve: Versatile Retrieval For Videos Via Unified Embeddings (2026)0.00
- Mine And Refine: Optimizing Graded Relevance In E-commerce Search Retrieval (2026)0.00
- Tevatron 2.0: Unified Document Retrieval Toolkit Across Scale, Language, And Modality (2025)3.58
- Towards Efficient And Robust Moment Retrieval System: A Unified Framework For Multi-granularity Models And Temporal Reranking (2025)2.26
- Retrieval-grpo: A Multi-objective Reinforcement Learning Framework For Dense Retrieval In Taobao Search (2025)0.00
- Visual Product Search Benchmark (2026)0.00
- Universal Vision-language Dense Retrieval: Learning A Unified Representation Space For Multi-modal Retrieval (2022)3.45