Cross-Border Threat Screening and Supply Chain Defense

A Department of Homeland Security Center of Excellence

Automated AI Capabilities for Entity Risk Screening to Highlight Supply Chain Vulnerabilities in Pharmaceutical, Chemical and ESG-Sensitive Industries

Issue:

This project examines how an artificial intelligence platform designed to deliver actionable intelligence and identify institutional risks in relation to human rights violations, environmental abuses, and foreign malign influence can be used to assess risks across selected supply chains. This project provides an opportunity to test the capacity of investigatory processes and resolution assistance in assessing risks and provides a basis for identifying supply chain risks on a global scale.

Objectives:

  1. Define the necessary new risk typologies and data sets required for discovering supply chain risk.
  2. Establish clear standards, definitions, and training data examples for target risk factors and supply chain vulnerabilities.
  3. Train risk models and prove the ability of discovered features to identify threats optimizing for balance between minimal false positives and false negatives (precision and recall).
  4. Demonstrate the effectiveness of Quantifind’s models and discovered features in identifying emerging supply chain risks.
  5. Train interested and relevant stakeholders, academia, and private industry on a curated environment of Quantifind’s GraphyteSearch UI and GraphyteBatch.
  6. Demonstrate time and cost savings, as well as enhanced coverage by automating supply chain risk identification.
technical illustration from Quantifind, Inc.

Value Proposition:

Analyze the pharmaceutical and chemical industry supply chains and develop the capability to evaluate vendor typologies for fraud and criminal risk, financial risk, foreign malign influence risk and ESG risk (including labor rights violations and sustainability issues). Train risk models on examples from supply chains with known counterfeit and fraudulent products, or products from organizations with known human rights or environmental violations or with links to entities with the same risk factors to automatically discover predictive features of fraud or human rights or sustainability violations.

Project LeadQuantifind, Inc.
Research TeamPI: John Stockton, Ph.D., Quantifind Inc.
Co-PI: Ari Tuchman, Ph.D., Quantifind Inc.
Co-PI: Jennifer Roy, Quantifind Inc.
Budget$729, 762
DurationJan. 2023 – July 2024