Multi-Modal Information Extraction from Text, Semi-Structured, and Tabular Data on the Web

KDD Attendees: The Zoom link to our tutorial is available on the chat page for our tutorial. Click on "Breakout Sessions" -> "Lecture Style Tutorials", then find our tutorial (#11) and click on the chat link.


  1. Introduction [pdf]
  2. Extraction from Unstructured Text Part 1 [pdf]
  3. Extraction from Unstructured Text Part 2 [pdf]
  4. Extraction from Semi-Structured Text [pdf]
  5. Extraction from Tabular Text [pdf]
  6. Multi-modal Extraction [pdf]
  7. Conclusion [pdf]


The World Wide Web contains vast quantities of textual information in several forms: unstructured text, template-based semi-structured webpages (which present data in key-value pairs and lists), and tables. Methods for extracting information from these sources and converting it to a structured form have been a target of research from the natural language processing (NLP), data mining, and database communities. While these researchers have largely separated extraction from web data into different problems based on the modality of the data, they have faced similar problems such as learning with limited labeled data, defining (or avoiding defining) ontologies, and making use of prior knowledge.

In this tutorial we take a holistic view toward information extraction, exploring the commonalities in the challenges and solutions developed to address these different forms of text. We will explore the approaches targeted at unstructured text that largely rely on learning syntactic or semantic textual patterns, approaches targeted at semi-structured documents that learn to identify structural patterns in the template, and approaches targeting web tables which rely heavily on entity linking and type information.

While these different data modalities have largely been considered separately in the past, recent research has started taking a more inclusive approach toward textual extraction, in which the multiple signals offered by textual, layout, and visual clues are combined into a single extraction model made possible by new deep learning approaches. At the same time, trends within purely textual extraction have shifted toward full-document understanding rather than considering sentences as independent units. With this in mind, it is worth considering the information extraction problem as a whole to motivate solutions that harness textual semantics along with visual and semi-structured layout information. We will discuss these approaches and suggest avenues for future work.


Xin Luna Dong

Xin Luna Dong is a Principal Scientist at Amazon, leading the efforts of constructing Amazon Product Knowledge Graph. She was one of the major contributors to the Google Knowledge Vault project, and has led the Knowledge-based Trust project, which was called the "Google Truth Machine" by the Washington Post. She has co-authored the book "Big Data Integration", was awarded ACM Distinguished Member, VLDB Early Career Research Contribution Award for "advancing the state of the art of knowledge fusion", and Best Demo award in Sigmod 2005. She serves in VLDB endowment and PVLDB advisory committee, and is a PC co-chair for VLDB 2021, ICDE Industry 2019, VLDB Tutorial 2019, Sigmod 2018 and WAIM 2015. She has given multiple tutorials on data integration, graph mining, and knowledge management in top-tier conferences.

Colin Lockard

Colin Lockard is an Applied Scientist at Amazon and a PhD candidate at the Paul G. Allen School of Computer Science & Engineering at the University of Washington, where he has published papers on knowledge extraction from both unstructured and semi-structured text.

Prashant Shiralkar

Prashant Shiralkar is an Applied Scientist in the Product Graph team at Amazon. He currently works on knowledge extraction from semi-structured data. Previously, he received a Ph.D. from Indiana University Bloomington where his dissertation work focused on devising computational approaches for fact checking by mining knowledge graphs. His research interests include machine learning, data mining, information extraction and NLP, and Semantic Web technologies.

Hannaneh Hajishirzi

Hannaneh Hajishirzi is an Assistant Professor at the Paul G. Allen School of Computer Science & Engineering at the University of Washington. She works on NLP, AI, and machine learning, particularly designing algorithms for semantic understanding, reasoning, question answering, and information extraction from multimodal data. She has earned numerous awards for her research, including an Allen Distinguished Investigator Award, a Google Faculty Research Award, a Bloomberg Data Science Award, an Amazon Research Award, and a SIGDIAL Best Paper Award.