報告題目1: Knowledge Graphs and GeoAI: Uniting Symbolic and Subsymbolic Traditions
報告摘要：
Knowledge graphs are a novel paradigm for representing, retrieving, and integrating data from highly heterogeneous sources. Over the past years, they have attracted a substantial body of work combining symbolic and subsymbolic traditions, e.g., in representation learning. Places and geographic regions more broadly act as key nexus of these graphs connecting people, events, and objects. Hence, recent work also studies how these knowledge graphs can form the underpinning of geo-foundation models in GeoAI and even autonomous GIS agents. This presentation introduces KnowWhereGraph, a 15 billion graph statement strong knowledge graph of pre-integrated cross-domain data that provides geo-enrichment services to empower applications in supply chain management, humanitarian relief, and environmental intelligence. We will give an overview of the graph, its services, application areas, and research challenges and close with some broader thoughts about the future of knowledge engineering and GeoAI.
報告人簡介:
Krzysztof Janowicz is a geoinformatics and spatial data science full professor at the University of Vienna, Austria, and the University of California, Santa Barbara, USA. Before, he was an assistant professor at Pennsylvania State University. He is co-editor in chief of the Semantic Web journal, principal investigator of the KnowWhereGraph, former director of the Center for Spatial Studies at UCSB, and former program chair of UCSB’s cognitive science graduate program. Professor Janowicz’s research interests include knowledge graphs, geo-semantics, and GeoAI. Methodologically, his research combines theory-driven (e.g., semantics) and data-driven (e.g., data mining) techniques to improve data retrieval, integration, and reuse.
Krzysztof Janowicz, 奧地利維也納大學和美國加州大學圣巴巴拉分校（UCSB）教授。從事地理信息科學和空間數據科學研究。Janowicz教授還曾是美國賓夕法尼亞州立大學助理教授， 美國加州大學空間研究中心（Center for Spatial Studies @ UCSB）主任和UCSB 認知科學研究生項目主席。他現任Semantic Web Journal主編， KnowWhereGraph 項目 （美國NSF資助）主要負責人。Janowicz教授研究興趣包括知識圖譜，地理語義學，和地理人工智能（GeoAI）。Janowicz教授的研究方法結合理論驅動和數據綜和，并廣泛應用在信息檢索，信息整合和重用上。
報告題目2：KnowWhereGraph: A Geospatial Knowledge Graph to Support Cross-Disciplinary Knowledge Discovery
報告摘要：
Global challenges such as food supply chain disruptions, pandemics, and disaster responses require a more effective and rapid way of sharing and reusing data. However, most existing (geo)portals are stacked by separated or sparsely connected data “silos” impeding such a data consolidation. The KnowWhereGraph, a project consisting of not only the schema and data about human and environment, but also a series of supporting tools to access them, proposes to address such a challenge by introducing a large-scale, cross-domain, and AI-ready data warehouse that is rooted on knowledge graphs. In this talk, I highlight technical details underlying the core place-based ontology design of KnowWhereGraph, where place and time play an essential role of connecting and contextualising heterogenous and multimodal data. With the new data format as the backbone, I further introduce spatially explicit neural networks to reason over spatial and temporal information. Finally, I conclude the talk with challenges and outlook about geospatial knowledge graphs.
報告人簡介:
朱睿博士現就職于英國布里斯托大學講師（助理教授），博士生導師， 地理數據科學和空間分析研究生項目負責人。 朱睿博士畢業于美國加州大學圣巴巴拉分校（博士）和美國匹茲堡大學（碩士）。朱睿博士曾就職于美國加州大學圣巴巴拉分校空間研究中心（Center for Spatial Studies @ UCSB）和美國橡樹嶺國家實驗室。主要研究集中在理解和量化人以及機器對環境和空間的認知及交互。具體方向包括地理知識圖譜，空間顯性機器學習，和空間統計。
時 間：2023年8月9日（星期三）上午9:30
地 點：A0214會議室
主持 人：陸 鋒 研究員
