Didi's Participation in JSM 2019: Four Accepted Papers and AI‑Driven Insights

The Joint Statistical Meetings (JSM) 2019 was held in Denver, USA from July 27 to August 1. It is the largest and longest‑running annual conference in statistics, attracting over 6,000 scholars and practitioners.

Didi’s AI Labs participated for the second time after its 2018 debut in Vancouver. Representatives including chief statistician Zhu Hongtu, data science head Song Shijun, chief data scientist Xie Liang, researcher Luo Shikai, tech cooperation manager Ma Nan and algorithm intern Zhou Fan presented the company’s work.

Four Didi papers were accepted for oral or poster sessions:

“Equilibrium Metrics for Dynamic Supply‑Demand Networks” – a directed weighted‑graph metric that quantifies the real‑time distance between supply and demand, with applications to dispatch evaluation and platform‑level predictions.

“A Statistical and Machine Learning Framework for New Energy Vehicle Ride‑sharing System” – introduces an electricity‑consumption model for EVs and a multi‑objective deep reinforcement‑learning dispatch algorithm.

Two additional papers covering supply‑balance, user lifetime value, and dynamic network analysis were also selected.

In addition, chief statistician Zhu Hongtu gave a keynote “AI for Marketplace” at the Deep Learning for Data Science symposium, covering Didi’s AI strategy, causal models, equilibrium matrices, and applications such as demand forecasting, supply‑demand diagnosis, and strategy optimization.

Didi also hosted a technical salon during JSM, where senior data scientists shared case studies on data‑driven business development, experimental design, and economic modeling.

The event served as a recruiting platform; Didi’s booth attracted many visitors and the company conducted on‑site interviews for data‑science positions.

For more information, visit https://outreach.didichuxing.com/jsm2019/ or contact [email protected].