Read Aheads
- Judea Pearl’s Book of Why
- Mind over Data
- Ladder of Causation
Disrupting Wildlife Trafficking Networks Through Convergence of Physical and Virtual Ecosystems
This 18-month National Science Foundation-funded planning grant (Division of Information & Intelligent Systems Award IIS-2039951) is a new collaboration between University of Maryland, Worcester Polytechnic Institute and Focused Conservation Solutions.
Our goal is to build research capacity for creating an integrative team of scientists, policymakers and domain experts to disrupt wildlife trafficking networks through convergence of physical and virtual ecosystems. Check out our abstract!
The knowledge gaps we are interested in are as follows: (1) The evidentiary burden of wildlife trafficking in physical and virtual ecosystems is heavy, disparate datasets are growing, while sense making lags. (2) Causality and inference can be difficult to establish in a data universe of often indistinguishable licit and illicit activity.
Our planning mission is to address how can new teams of operations engineers, computer scientists and social scientists algorithmically resolve and validate data for disrupting wildlife trafficking to help/support responses of authorities and experts?
DATA4WILDLIFE CHALLENGE
The #DATA4WILDLIFE challenge will take place on 29/30 January 2022. Applications are now open and will close at 22.00 on 12 January 2021. The #Data4Wildlife Challenge is led by Bright-Tide UK.
We will hack THREE challenge questions:
Challenge Question #1 (Data for Deterrence): Social media platforms are one online location where wildlife products can be illegally bought, sold, advertised and marketed using images, text, and combinations of images/text. Right now, data scientists can’t use their tools to create new information about online wildlife crime, such as seeing new trend lines, drawing inference, or identifying counterintuitive correlations. Challenge #1 will create a benchmark dataset of online wildlife crime images, text, and emojis on three social media platforms (Instagram, TikTok, SnapChat). This database can be used by analysts and authorities to differentiate online wildlife crime from unproblematic depictions of wildlife.
Challenge Question #2 (Random Acts of Trade): Wildlife crime involves moving an animal or animal product from source to destination via a transit location. This illegal supply chain can cross both physical and virtual spaces. Right now, engineers can’t use their tools to create new information about the illegal supply chain structure or process. Challenge #2 will create a virtual and spatialized supply chain for a specific animal or animal product and identify the places and times where virtual and physical spaces touch. Understanding the intersections of physical and virtual supply chains will allow different organizations to track supply patterns and changes in supply chain patterns.
Challenge Question #3 (Media for the Masses): The media remains one of the most important witnesses of wildlife crime. Stories in the media tell us about animals and animal products, wildlife crime offenders and defenders, lawyers and judges, and how the wildlife crime happened. Right now, data scientists can’t use their tools to create new information about online wildlife crime, such as seeing new trend lines, drawing inference, or identifying counterintuitive correlations. Challenge #3 will create a wildlife crime media aggregator in multiple languages as well as a database of aggregated media that can be used to deter online wildlife crime.