Abstract:The recently explosive growth of Super Apps brings great convenience to people's daily life by providing a wide variety of services through mini-programs, including online shopping, travel, finance, and so on. Due to the considerable gap between various scenarios, the restriction of effective information transfer and sharing severely blocks the efficient delivery of online services, potentially affecting the user's app experience. To deeply understand users' needs, we propose SupKG, a commonsense knowledge graph towards Super APP to help comprehensively characterize user behaviors across different business scenarios. In particular, our SupKG is carefully established from multiplex and heterogeneous data source in Alipay (a well-known Super App in China), which also emphasize abundant spatiotemporal relations and intent-related entities to answer the fundamental question in life service ''which service do users need at what time and where''. On the hand, the successful application of SupKG hinges on the effective form of network representation ie Knowledge Graph Embedding (KGE). However, a series of unsatisfying issues still need to be carefully considered in the industrial environment: i) bridging language representations with knowledge structure in a unified manner, ii) alleviating the skewed data distribution in SupKG, and iii) effectively characterizing hierarchical structures in SupKG. With these motivations, we develop a novel knowledge graph representation learning framework for SupKG, enabling various downstream applications to benefit from learned representations of entities and relations. Extensive experiments on the standard knowledge graph completion task demonstrate the consistent and significant performance improvement of our representation learning framework, which also greatly benefits the supplementation of potential knowledge of SupKG. Towards real-world applications in Alipay, our SupKG and learned representations show the potential superiority of integrating global behaviors in cold-start scenarios and providing high-quality knowledge for warming up the graph-based ranking.

An Industrial-Scale System For Heterogeneous Information Card Ranking In Alipay

An End-to-end Model of Predicting Diverse Ranking on Heterogeneous Feeds.

Large Scale App Recommendation in Ant Financial

KunPeng: Parameter Server Based Distributed Learning Systems and Its Applications in Alibaba and Ant Financial

COUPA: An Industrial Recommender System for Online to Offline Service Platforms

PIER: Permutation-Level Interest-Based End-to-End Re-ranking Framework in E-commerce

Preference Transfer Model in Collaborative Filtering for Implicit Data

Learning Personalized Page Content Ranking Using Customer Representation

Learning User Credibility for Product Ranking

Xorder: A Model Agnostic Post-Processing Framework for Achieving Ranking Fairness While Maintaining Algorithm Utility.

CREDIT CARD RATING BASED ON REPEAT CLUSTERING IN MASSIVE DATA

Credibility-based Product Ranking for C2C Transactions.

Bipartite Ranking Fairness through a Model Agnostic Ordering Adjustment

A Large-Scale Reviews-Driven Multi-Criteria Product Ranking Approach Based on User Credibility and Division Mechanism

An Interactive Personalized Recommendation System Using the Hybrid Algorithm Model

OMNIRank: Risk Quantification for P2P Platforms with Deep Learning

Commonsense Knowledge Graph towards Super APP and Its Applications in Alipay

Exploiting Shopping and Reviewing Behavior to Re-Score Online Evaluations

Hybrid Learning to Rank for Financial Event Ranking.

RLMixer: A Reinforcement Learning Approach for Integrated Ranking with Contrastive User Preference Modeling.

Learning Item Embedding with Heterogeneous Information for Collaborative Filtering