商品描述
As data lakes have become a prominent foundation for enterprise and scientific data management, organizations increasingly face the challenge of locating relevant datasets and building ad-hoc integration pipelines across heterogeneous, poorly documented, and rapidly evolving data collections. In this setting, data discovery becomes a critical capability for turning raw, distributed data assets into usable knowledge. This book examines data discovery and its evolution across industry and academia. It covers the principles, systems, and techniques that enable users to find, understand, and use relevant data across increasingly complex data ecosystems. The book discusses modern approaches to efficient and effective data discovery, including novel system architectures, search and matching methods, metadata use, dataset profiling, and human-in-the-loop techniques. Beyond core technical concepts, the book offers insight into how data discovery systems are evaluated and benchmarked. It highlights practical challenges faced in real-world deployments, compares emerging academic and industrial approaches, and identifies open research questions that continue to shape the field. The book is intended for researchers, practitioners, and students interested in data management, data integration, data lakes, and the future of intelligent data access.
商品描述(中文翻譯)
隨著資料湖（data lake）成為企業與科學資料管理的重要基礎，組織日益面臨一項挑戰：如何在異質、文件記錄不完整且快速演進的資料集合中，找到相關資料集，並建立即時且針對特定需求的整合管線。在此情境下，資料探索（data discovery）成為一項關鍵能力，能將原始且分散的資料資產轉化為可使用的知識。
本書探討資料探索及其在業界與學術界的演進。內容涵蓋相關原理、系統與技術，協助使用者在日益複雜的資料生態系統中，尋找、理解並使用相關資料。本書也討論現代化且高效的資料探索方法，包括新穎的系統架構、搜尋與比對方法、詮釋資料（metadata）的運用、資料集剖析，以及人在迴圈（human-in-the-loop）技術。
除了核心技術概念之外，本書也深入說明如何評估資料探索系統，以及如何建立基準測試。書中凸顯真實世界部署所面臨的實務挑戰，比較新興的學術與產業方法，並指出持續形塑此領域的開放研究問題。本書適合對資料管理、資料整合、資料湖，以及智慧資料存取的未來發展感興趣的研究人員、實務工作者與學生閱讀。
作者簡介
Ziawasch Abedjan is Chair of the Data Integration and Data Preparation group at TU Berlin and research group lead at Berlin Institute for Foundations of Learning and Data (BIFOLD). His research is focused on developing generalizable and scalable techniques for various data integration problems, such as data cleaning, data science pipeline generation and understanding, and data discovery. Previously, he chaired the Database and Information systems group at the Leibniz University Hannover. He held positions as Assistant Professor at TU Berlin, Postdoctoral Associate at MIT, Research Associate at QCRI, Senior Researcher at the German Center for Artificial Intelligence (DFKI), and Visiting Academic at Amazon Search. He received his PhD from the Hasso Plattner Institute in Potsdam and was awarded the University of Potsdam's best Dissertation Prize in 2014. He has co-authored more than 80 peer-reviewed papers in leading database and software engineering venues and received recognition with several academic awards. His research is supported by the German Research Council (DFG) and the German Federal Ministry of Science and Education. Mahdi Esmailoghli is a postdoctoral research associate at the David R. Cheriton School of Computer Science at the University of Waterloo. His research is centered on the challenges of data discovery within large-scale data lakes. He investigates how data systems can assist researchers and users in navigating vast table repositories to identify relevant datasets and suggest optimized pipelines, including index structures and algorithms. His work has been published in leading data management venues, including SIGMOD, VLDB, ICDE, EDBT, and CIDR. Mahdi earned his PhD in Computer Science with the highest honors from the Technical University of Berlin. Sainyam Galhotra is an assistant professor in the Department of Computer Science at Cornell University. His research focuses on trustworthy artificial intelligence, data-centric AI, causal reasoning, and reliable agentic systems. He studies how AI systems can reason over complex multimodal data while remaining transparent, robust, and adaptive in dynamic environments. His work has been published in leading venues across machine learning, data management, and AI systems, including ICLR, NeurIPS, SIGMOD, VLDB, and ICSE. He is the recipient of the Best Paper Award at FSE 2017, the Most Reproducible Paper Award at SIGMOD 2017 and 2018, and the Best Artifact Paper Honorable Mention Award at SIGMOD 2023. He has also been recognized as a Data Science Rising Star, a DAAD AInet Fellow, and the first recipient of the Krithi Ramamritham Award at the University of Massachusetts Amherst for contributions to database research.
作者簡介(中文翻譯)
Ziawasch Abedjan 是 TU Berlin 資料整合與資料準備（Data Integration and Data Preparation）團隊的主持人，也是 Berlin Institute for Foundations of Learning and Data（BIFOLD）的研究團隊主管。他的研究著重於為各種資料整合問題開發具一般化能力且可擴展的技術，例如資料清理、資料科學管線的生成與理解，以及資料探索。此前，他曾擔任 Leibniz University Hannover 資料庫與資訊系統團隊的主持人。他曾任職於 TU Berlin 的助理教授、MIT 的博士後研究員、QCRI 的研究助理、German Center for Artificial Intelligence（DFKI）的資深研究員，以及 Amazon Search 的訪問學者。他在 Potsdam 的 Hasso Plattner Institute 取得博士學位，並於 2014 年獲頒 University of Potsdam 最佳博士論文獎。他與他人共同撰寫了 80 多篇發表於頂尖資料庫與軟體工程學術會議及期刊、經同儕審查的論文，並獲得多項學術獎項的肯定。
他的研究獲得 German Research Council（DFG）及 German Federal Ministry of Science and Education 的支持。
Mahdi Esmailoghli 是 University of Waterloo David R. Cheriton School of Computer Science 的博士後研究員。他的研究聚焦於大規模資料湖（data lake）中的資料探索挑戰。他探討資料系統如何協助研究人員與使用者在龐大的資料表儲存庫中進行瀏覽，以找出相關資料集並建議最佳化的管線，其中包括索引結構與演算法。
他的研究成果已發表於資料管理領域的頂尖學術會議，包括 SIGMOD、VLDB、ICDE、EDBT 及 CIDR。Mahdi 以最高榮譽於 Technical University of Berlin 取得 Computer Science 博士學位。
Sainyam Galhotra 是 Cornell University Computer Science Department 的助理教授。他的研究聚焦於可信賴人工智慧（trustworthy artificial intelligence）、以資料為中心的 AI（data-centric AI）、因果推理（causal reasoning）及可靠的代理型系統（agentic systems）。他研究 AI 系統如何在保持透明、穩健，並能適應動態環境的同時，對複雜的多模態資料進行推理。他的研究成果已發表於機器學習、資料管理及 AI 系統領域的頂尖學術會議，包括 ICLR、NeurIPS、SIGMOD、VLDB 及 ICSE。
他曾獲頒 FSE 2017 最佳論文獎、SIGMOD 2017 及 2018 最具可重現性論文獎，以及 SIGMOD 2023 最佳研究成果論文榮譽提名獎。他也曾獲選為 Data Science Rising Star、DAAD AInet Fellow，並因其對資料庫研究的貢獻，成為 University of Massachusetts Amherst 首位獲頒 Krithi Ramamritham Award 的得主。
