Pailitao-MMSearch: Building Native E-Commerce Multimodal Search Foundation

The evolution of e-commerce has fundamentally transformed how users search for products, shifting from simple text-based keyword queries to complex multimodal interactions that seamlessly combine product images, natural language descriptions, and mixed-intent instructions. However, existing approaches face a critical dilemma: single-modal specialist models, deployed independently for text retrieval, visual search, and voice recognition, operate in isolation and cannot handle cross-modal queries, while general-purpose vision-language models lack the domain-specific knowledge necessary for fine-grained product understanding, user behavior modeling, and commercial intent reasoning. In this work, we present Pailitao-MMSearch, one native e-commerce multimodal search foundation model designed to bridge this gap. Our approach introduces three key innovations: (1)HybSID (Hybrid Semantic ID);(2)a two-stage continual pre-training strategy; and (3)a hybrid reasoning post-training pipeline. Built upon Qwen and deployed on Taobao's Pailitao multimodal search platform, Pailitao-MMSearch achieves substantial improvements in online A/B testing, including up to +13.61\% in Gross Merchandise Volume (GMV) and +8.21\% in transaction volume compared to traditional multi-modal search pipeline, demonstrating the effectiveness of our native e-commerce multimodal search large language models.

Pailitao-VL: Unified Embedding and Reranker for Real-Time Multi-Modal Industrial Search

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 discriminative resolution while circumventing the prohibitive latency typically associated with conventional reranking methods. Extensive offline benchmarks and online A/B tests on Alibaba e-commerce platform confirm that Pailitao-VL achieves state-of-the-art performance and delivers substantial business impact. This work demonstrates a robust and scalable path for deploying advanced MLLM-based retrieval architectures in demanding, large-scale production environments.

Distributed Multisensor ISAC

Integrated Sensing and Communications (ISAC) will become a service in future mobile communication networks. It enables the detection and recognition of passive objects and environments using radar-like sensing. The ultimate advantage is the reuse of the mobile network and radio access resources for scene illumination, sensing, data transportation, computation, and fusion. It enables building a distributed, ubiquitous sensing network that can be adapted for a variety of radio sensing tasks and services. In this article, we develop the principles of multi-sensor ISAC (MS-ISAC). MS-ISAC corresponds to multi-user MIMO communication, which in radar terminology is known as distributed MIMO radar. \ First, we develop basic architectural principles for MS-ISAC and link them to example use cases. We then propose a generic MS-ISAC architecture. After a brief reference to multipath propagation and multistatic target reflectivity issues, we outline multilink access, coordination, precoding and link adaptation schemes for MS-ISAC. Moreover, we review model-based estimation and tracking of delay~/~Doppler from sparse OFDMA~/~TDMA frames. We emphasize Cooperative Passive Coherent Location (CPCL) for bistatic correlation and synchronization. Finally, issues of multisensor node synchronization and distributed data fusion are addressed.