Speaker: Hao Zhong（钟浩），ESCP Business School (Paris).

Time/Date: 2023年9月4日（周一），14：00-15：30

Classroom: 同济大厦A楼1922会议室

ABSTRACT:

The rapid acceleration of technology and the evolving global economy have led to a significant surge in high-potential startups, presenting immense opportunities for venture capital firms and investors to support and benefit from these innovative ventures. However, identifying startups with the highest likelihood of success remains a complex task, necessitating the examination of various information sources, including firm demographics, management team composition, and financial performance. The effectiveness of existing methodologies, such as feature-based and network-topological approaches, is limited for predicting high-potential startups. In response, we propose a novel Venture Graph Neural Network (VenGNN) model, leveraging Heterogeneous Information Networks (HIN) and Graph Neural Networks (GNN) techniques to address the prediction problem. Specifically, we construct a Heterogeneous Venture Information Network (HVIN) using raw business data and deem the prediction as a node classification task. Our model integrates theory-guided semantic meta-paths, firm demographics, a fused heterogeneous attention layer, sampling-based self-attention, and centrality encoding to overcome certain constraints of existing GNNs, such as over-smoothing and lack of interpretability. Our experimental analysis reveals that VenGNN outperforms state-of-the-art models by 15-20% across a wide range of performance metrics. Our study also includes a comprehensive interpretation analysis and presentation of case studies that illustrate the efficacy of VenGNN, providing stakeholders with the essential knowledge to make well-informed investment decisions.

GUEST BIO:

Dr. Hao Zhong is a tenure-track assistant professor in the Information and Operations Management (IOM) department and the scientific co-director of Masters in Big Data and Business Analytics program at ESCP Business School (Paris). He was a visiting professor at the AI Thrust of The Hong Kong University of Science and Technology (HKUST) – Guangzhou in 2023. He received his Ph.D. in Management (Information Technology) from Rutgers University in USA. His research interests include data mining, computational design science, venture data analyitics, and representation learning. His work has appeared in multiple refereed journals (TKDE, TMC, AOR) and peer-reviewed conference proceedings (ICIS, CIST, HICSS, WITS, KDD, ICDM, SDM). He has also served as Senior Program Committee (SPC) member at AAAI and Program Committee (PC) member at IJCAI, WWW, CIKM, KSEM, etc. He received the Best Paper Runner-Up award at WITS 2022. He has been serving as the co-chair of AI cluster at INFORMS Annual Meeting since 2021.