Skip to content

JayanaGunaweera01/EthAIAuditHub

Repository files navigation

EthAIAuditHub - An automated, collaborative bias auditing platform for ML models

Welcome to EthAIAuditHub - An Automated, Collaborative, ethical Bias Auditing Platform for ML models.This platform is designed to help ethically biased communities,developers, data scientists, policy makers and all the relavant stakeholders to assess and mitigate algorithmic biases in machine learning models effectively. By leveraging cutting-edge technologies and methodologies, this platform aims to uphold integrity, fairness in ML models which then automatically eliminate the ethical bias.

Workflow Status

Workflow Status

Workflow Status

Workflow Status

License

Objective

This repository contains a comprehensive Automated Auditing Framework for AI systems, designed to ensure transparency, fairness, and accountability in the development and deployment of ML models. By developing a meaningful automated platform with an AI assistant for addressing biases in algorithmic models, the developer aims at mitigating ethical bias which are not just technical glitches; but are societal challenges that necessitate collective attention and thoughtful solutions.

Stakeholder Involvement

Engage with key stakeholders, including developers, data scientists, legal experts, and end-users. Maintain open communication channels to gather feedback and ensure a collaborative approach.File sharing, discussion forums, communities and real-time collaboration tools promote inclusivity and diverse contributions.

Collaboration with IBM AIF 360

Utilization of IBM AIF 360, a framework with the highest accuracy in bias detection.Align the framework with relevant legal and ethical standards, emphasizing principles such as fairness, transparency, and accountability.

Automated ML model execution platform

Execute Bias Mitigation Automation Workflow and EthAIAuditHub Experiment Workflow to have an automated experience in executing ipynb files leveraging github actions

Chat with AI Assistant and do an audit with your ML model executions

Use in built AI assistance which is a GPT based platform that can streamline the process of bias detection, auditing your ipynb files and fixing bugs in executing ML models.

Log reports

Experience a better UX with automatically executed log reports

Getting started

visit: https://web-app-delta-wheat.vercel.app/