Page 173 - AI for Good-Innovate for Impact Final Report 2024
P. 173

AI for Good-Innovate for Impact



               (continued)

                Domain         Good Health and Well-Being
                Model Train-   •  A continuous-time dynamic graph representation learning framework
                ing and           is proposed for OD(Origination and Destination) demand forecasting.               39-ChinaMobile
                fine-tuning       Different from the previous research, our method maintains continu-
                                  ous time node representations and updated them continuously once a
                                  number of transactions are available. As the evolutionary dynamics of
                                  stations could be learned in an extremely fine time scale, our method is
                                  promising to achieve higher prediction accuracy.
                               •  A hierarchical message passing module is proposed to model the spatial
                                  interactions of stations.
                               •  Extensive experiments have been conducted on two real-world datasets.
                Case Studies   The World Health Organization's Decade of Healthy Ageing 2021-
                               2030(SDGS 3, 10, 11).
                Testbeds or    The pilot is deployed on China Mobile's internal network.
                pilot deploy-
                ments



               39�2� Use case description


               39�2�1  Description

               Introduction: China Mobile is actively responding to the challenge of co-benefits arising from
               "working to improve the lives of older people, their families and the communities in which they
               live" within the National strategic plan advocated by the World Health Organization. To address
               these complex issues, China Mobile has developed the AI Enabling Jiutian Travel Model, which
               is associated with good health and well-being and can improve a country's performance in
               achieving the Sustainable Development Goals.

               Overview of the product: AI Enabling Jiutian Travel Model includes a large model of real-time
               positioning, identity determination, loss judgment and travel path prediction for elderly people
               living alone, aiming to accurately and quickly monitor the complex location trajectory dynamics
               of elderly people living alone. This product, based on 5G base station signaling data analysis,
               integrates AI-based trajectory analysis for elderly people living alone to achieve the co-benefits
               of sustainable development over the Decade of Healthy Ageing 2021-2030.

               Cooperative efforts: China Mobile, in cooperation with the Chinese government, has developed
               an AI Enabling Jiutian Travel Model, which covers functions such as real-time location, identity
               determination, loss of contact judgment, travel path prediction, and pilot applications to help
               family members of elderly people timely grasp the location information and health and safety
               status of elderly people living alone. On the basis of this work, through cooperation with the
               government, products and services are extended to more families, so that more elderly people
               can enjoy health care services. This product will enable the achievement of the SDG co-benefits,
               particularly in the realm of good health and well-being for older persons living alone, This is
               essential to promote equitable and sustainable development.

               Partner: Zhuhu Town People's Government, Sihong County, Suqian City, Jiangsu Province







                                                                                                    157
   168   169   170   171   172   173   174   175   176   177   178