Program

Overall Program (Sunday August 15, 2021, time in PDT)

(all video recordings can be found at Document Intelligence Workshop @ KDD 2021 YouTube Channel)

Start TimeEnd TimeDurationTopicSession ChairSlidesVideo
8:008:100:10Workshop Opening RemarksBenjamin Hanpptx–
8:108:500:40Invited Talk 1: Cha Zhang - Visual Document Intelligence in the WildSandeep TatapdfVisual Document Intelligence in the Wild
8:509:501:00Presentation Session 1: OCR & Visual Document IntelligenceSandeep Tata––
9:5010:050:15COFFEE BREAK 1–––
10:0510:450:40Invited Talk 2: Kevyn Collins-Thompson - Enhancing Document Representations Using Analysis of Content Difficulty: Models, Applications, and InsightsYijuan (Lucy) LupdfEnhancing Document Representations Using Analysis of Content Difficulty: Models, Applications, and Insights
10:4511:350:50Presentation Session 2: Machine LearningYijuan (Lucy) Lu––
11:3512:050:30LUNCH BREAK–––
12:0512:450:40Invited Talk 3: Yunyao Li - Towards Deep Table UnderstandingDouglas BurdickpdfTowards Deep Table Understanding
12:4513:451:00Presentation Session 3: ApplicationsDouglas Burdick––
13:4514:000:15COFFEE BREAK 2–––
14:0014:400:40Invited Talk 4: Heng Ji - What’s in a Chemical Entity?Hamid MotaharipptxWhat’s in a Chemical Entity?
14:4015:200:40Invited Talk 5: Benjamin Van Durme - A Case for Statutory ReasoningHamid MotaharipptxA Case for Statutory Reasoning
15:2015:350:15COFFEE BREAK 3–––
15:3516:150:40Invited Talk 6: Don Metzler - Challenges in Enterprise Search and IntelligenceHamid Motahari–Challenges in Enterprise Search and Intelligence
16:1516:450:30DI 2021 Best Paper Presentation: HYCEDIS: HYbrid Confidence Engine for Deep Document Intelligence SystemDave LewispdfHYCEDIS: HYbrid Confidence Engine for Deep Document Intelligence System
16:4517:451:00Panel: DI Research Challenges & DirectionsDave Lewis–Panel: DI Research Challenges & Directions
17:4518:000:15Workshop Summary and Closing RemarksBenjamin Hanpptx–

Presentation Session 1: OCR & Visual Document Intelligence

PresentationTitleSlidesVideo
PaperCHARTER: heatmap-based multi-type chart data extractionpdfCHARTER: heatmap-based multi-type chart data extraction
PaperDetection Masking for Improved OCR on Noisy DocumentspptxDetection Masking for Improved OCR on Noisy Documents
PaperEfficient Document Image Classification Using Region-Based Graph Neural NetworkpptxEfficient Document Image Classification Using Region-Based Graph Neural Network
PaperLights, Camera, Action! A Framework to Improve NLP Accuracy over OCR documentspdfLights, Camera, Action! A Framework to Improve NLP Accuracy over OCR documents
PosterMulti-Stage Framework to Boost Optical Character Recognition Performance on Low Quality Document ImagespdfMulti-Stage Framework to Boost Optical Character Recognition Performance on Low Quality Document Images

Presentation Session 2: Machine Learning

PresentationTitleSlidesVideo
PaperData-Efficient Information Extraction from Form-Like DocumentspdfData-Efficient Information Extraction from Form-Like Documents
PaperPosition Masking for Improved Layout-Aware Document UnderstandingpptxPosition Masking for Improved Layout-Aware Document Understanding
PaperText Analysis via Binomial TailspptxText Analysis via Binomial Tails
PosterFew-Shot Learning for Structured Information Extraction From Form-Like Documents Using a Diff AlgorithmpdfFew-Shot Learning for Structured Information Extraction From Form-Like Documents Using a Diff Algorithm

Presentation Session 3: Applications

PresentationTitleSlidesVideo
PaperGenerating and evaluating simulated medical notes: Getting a Natural Language Generation model to give you what you wantpdf Generating and evaluating simulated medical notes: Getting a Natural Language Generation model to give you what you want
PaperSpecToSVA: Circuit Specification Document to SystemVerilog Assertion TranslationpdfSpecToSVA: Circuit Specification Document to SystemVerilog Assertion Translation
PosterMedical Report Generation with Multi-Attention for Abnormal Keyword Description and History ReportpptxMedical Report Generation with Multi-Attention for Abnormal Keyword Description and History Report
PosterThe Law of Large Documents: Understanding the Structure of Legal Contracts Using Visual CuespptxThe Law of Large Documents: Understanding the Structure of Legal Contracts Using Visual Cues
PosterTowards Semantic Search for Community Question Answering for Mortgage OfficerspdfTowards Semantic Search for Community Question Answering for Mortgage Officers