FDD
The Fraud Detection & Defense (FDD) working group, which includes members from GitcoinDAO, is focused on identifying and defending against economic attacks such as collusion strategies by bots and fraudulent accounts. The group analyzes Gitcoin data, explores mechanisms related to Civil resistance, and considers process and policy design solutions. The FDD-WG utilizes Grant round data and machine learning pipelines to define, detect, evaluate, and sanction fraud and other attacks on the Gitcoin platform, while also expanding the monitoring capabilities of human jurors.
The FDD working group is divided into several workstreams, with DisruptionJoe.icon and jimcook.icon being part of it. There are likely more members due to the large size of the organization.
The Fraud Detection & Defense Working Group Mapping - 🧙 🧙‍♀️ Ideas and Open Discussion - Gitcoin Governance post provides an overview of the FDD-WG. It explains that the group aims to identify and defend against economic attacks, and discusses the analysis of Gitcoin data, mechanisms related to Civil resistance, and the exploration of process and policy design solutions. The post also includes a diagram that can be found in the Gitcoin Grants Round 11 Anti-Fraud Evaluation & Results | by BlockScience | BlockScience | Medium article.
The FDD working group consists of various entities, including Gitcoin Holdings (the core team of Gitcoin), BlockScience as paid contractors/subject matter experts/advisors, and unpaid volunteers from the Gitcoin community, Token Engineering community, and other ecosystems.
Further information about the FDD working group can be found in the Build and Fund the Open Web Together | Gitcoin page on Gitcoin Grants. The post explains that the Anti-Fraud & Collusion Detection working group analyzes Gitcoin data, explores mechanism design solutions for anti-fraud, and gradually decentralizes machine learning algorithm governance and other policies within the Gitcoin community. It also mentions the collaboration between Gitcoin, BlockScience, Token Engineering Academy & Token Engineering Commons Labs, and the 8-week research on grant data. The post provides an overview of the five substreams within the FDD working group, including verification of grants, ML Sybil detection, flag evaluation, appeals and sanctions, and workstream operations, education, and communication.
The FDD working group requires significant effort, with over 80 hours per month dedicated to workstream operations alone. This includes creating and mapping processes, onboarding, education and training, flag evaluation, and the development and maintenance of appeals and sanctions processes. The group also takes on the responsibility of decentralizing this large workstream and plays a crucial role in infrastructure and security.
For the current grants, support is requested for the 10th round until further plans for the long-term decentralization of anti-fraud activities are made. The funds will support several volunteer groups (and potentially more) actively participating in the work.
Funding is managed by a multisig consisting of two community members, two Gitcoin team members, and one BlockScience team member, who volunteer to oversee the operations and coordination of this workstream. The allocation of funds is determined by these core multisig stewards, who have a comprehensive understanding and tracking of all the work.
In the future, the FDD working group plans to explore longer-term plans, contributor reward mechanisms, and systems. However, since the group is currently small, it is relatively easy to allocate funds to support the volunteers.
For more details about the workstream, questions and greetings can be posted in the Workstream: Fraud Detection and Defense Working Group Assemble! - 🧙 🧙‍♀️ Ideas and Open Discussion - Gitcoin Governance thread.