Wu LIU

Wu Liu is one of the early pioneers in the field of all-scenario multi-modal intelligent perception for the intelligent supply chain. He originally proposed the supply-chain-oriented progressive search paradigm, along with the application framework for “human-product-scene” massive and all-scenario data. The paradigm contains three important search strategies: a) coarse-to-fine search in feature space; b) near-to-distant search in spatial-temporal space; c) low-to-high permission search in the security space. These strategies further improved the accuracy and efficiency of object search and retrieval.

He participated in building the unique National AI innovation platform for the smart supply chain, which is certified by the Chinese Ministry of Science and Technology. This AI platform released 86 APIs and 130 developer kits, with around 7.9 billion user call volumes per day, covering all supply chain scenarios.

On top of this platform, Liu’s team developed fine-grained product image-searching techniques, and applied them to the Jindong snap-and-shop and smart counter project. With the help of the multi-modal product search technique, the smart counter project, where he took charge, is able to recognize and check out multiple products with any viewing angle in one go, in the blink of an eye. This project has been deployed for commercial use now and can save 30% of the time for customers during rush hour.