Ctrip’s Deep Learning Recommendation System Paper Accepted at AAAI Conference

Domestic online travel companies in China are increasingly showcasing technological prowess. Last week, Ctrip displayed a series of new tech achievements in Shanghai, and its deep‑learning research paper was recently accepted by the top‑tier AAAI conference, indicating that Ctrip’s AI‑plus‑travel research has reached an international frontier.

AAAI is the premier international conference in artificial intelligence, representing the highest standards of AI research and serving as a primary venue for leading academic and industry breakthroughs from companies such as Google and Microsoft. Ctrip has now become a member of AAAI’s elite technology enterprises.

The accepted paper, titled “A Hybrid Collaborative Filtering Model with Deep Structure for Recommender Systems,” was written by Ctrip’s Business Intelligence (BI) team in the Basic Business Department, based on its internally developed universal recommendation system. The project previously received funding from the Shanghai Economic and Information Commission and, after one year of internal deployment, has delivered significant performance improvements.

According to Yu Lei, head of Ctrip’s Technical Center BI team, the recommendation system differs from traditional OTA products by leveraging extensive AI techniques, including deep neural network auto‑encoding as described in the paper. The system has been rolled out in more than 50 personalized scenarios, achieving dramatic gains—some scenarios saw conversion rates increase by up to 13 times—greatly enhancing user experience during product purchase and travel.

In 2023, Ctrip has achieved numerous AI breakthroughs. Its call‑center recently launched a next‑generation chatbot system, now handling nearly 40% of ticket‑booking customer service volume, making Ctrip the first large‑scale online travel company to adopt such a system. Additional innovations include smart door locks, hotel air‑quality monitoring, and dedicated customer‑service solutions.

According to its Q2 financial report, Ctrip’s R&D spending approaches 40% of revenue, surpassing peers such as Priceline and even tech‑heavy firms like Google and Amazon, underscoring its ambition to become a technology‑driven company. Wall Street analysts anticipate rapid revenue and profit growth, projecting a 20% operating margin by 2017.

Ctrip’s Technical Center currently runs five WeChat groups covering Architecture, Mobile, Big Data, Frontend, and Operations to facilitate peer exchange. Interested individuals can join by adding the assistant’s WeChat ID “ctirp_tech” and indicating the relevant domain (e.g., “Big Data”).