EDMWebinar - Risk Analytics using Knowledge Graphs / FIBO with Deep Learning
Tell a Friend About This EventTell a Friend
EDMWebinar - Risk Analytics using Knowledge Graphs / FIBO with Deep Learning

When: Wednesday, October 21, 2020

« Go to Upcoming Event List  

Risk Analytics using Knowledge Graphs / FIBO with Deep Learning


Co-hosted by  and 

Featuring a guest speaker from 

Wednesday, 21 October 2020
11 AM EDT / 4 PM UK




Gartner reported that close to 50% of AI and machine learning projects fail in part due to challenges with "data complexity and systems integration". In the context of risk analytics, combining macroeconomic data with internal risk data, tracking data provenance, and staging analytics-ready data for constantly changing business requirements illustrate several of those data challenges. Using publicly available data for 40 million mortgage loans over 20 years, we built a knowledge graph leveraging the FIBO ontology to demonstrate a next-generation analytics platform that:

  • Reduces time spent on data engineering tasks by over 50%
  • Natively captures detailed data lineage for each reported output metric
  • Rapidly and flexibly generates input data for deep learning models
  • Increases speed to delivery of sensitivity analysis and stress testing by generating model predictions directly in the graph database
  • Enables a more holistic capture of model risk management data



The webinar will explore:

  • Challenges common to a risk analytics pipeline
  • Comparison of relational data model and graph approaches
  • Application of graph analytics to mortgage loan data
  • Deep neural networks and graph convolutional networks
  • Scalable graph computation with TensorFlow
  • Use cases in adjacent areas including customer service, collections, fraud and AML


The webinar will be led by industry experts including:

  • Greg Steck, Director of Data Engineering, FI Consulting
  • Thomas Cook, Director of Sales, Cambridge Semantics
  • Mike Meriton, Moderator, Co-Founder & COO, EDM Council



This website uses cookies to store information on your computer. Some of these cookies are used for visitor analysis, others are essential to making our site function properly and improve the user experience. By using this site, you consent to the placement of these cookies. Click Accept to consent and dismiss this message or Deny to leave this website. Read our Privacy Statement for more.