Vector Institute Co-Leads ACM FAccT Conference on AI Fairness and Transparency
The Vector Institute for Artificial Intelligence is actively preparing for the 2021 ACM Conference on Fairness, Accountability, and Transparency (ACM FAccT), scheduled to take place virtually from March 3 to March 10. Co-led by Vector Research Director Richard Zemel, the conference serves as a premier cross-disciplinary forum for researchers, policymakers, and practitioners addressing bias in socio-technical systems. With attendance expected to double from the previous year's 600 participants, the event highlights the critical need for checks and balances in AI to prevent the reinforcement of existing biases in areas like hiring and medical diagnosis. The conference features keynote speakers, tutorials, and diverse sessions involving journalists, activists, and educators. Notably, Vector researchers will present two significant papers examining the trade-offs between privacy, utility, and fairness in machine learning. One study investigates differentially private synthetic data in classification tasks, while the other explores privacy guarantees in healthcare predictions. These contributions underscore the ongoing challenges in balancing model performance with ethical considerations such as privacy protection and equitable representation for minoritized groups in high-risk AI applications.
Wire timeline
Vector Institute Co-Leads ACM FAccT Conference on AI Fairness and Transparency
The Vector Institute for Artificial Intelligence is actively preparing for the 2021 ACM Conference on Fairness, Accountability, and Transparency (ACM FAccT), scheduled to take place virtually from March 3 to March 10. Co-led by Vector Research Director Richard Zemel, the conference serves as a premier cross-disciplinary forum for researchers, policymakers, and practitioners addressing bias in socio-technical systems. With attendance expected to double from the previous year's 600 participants, the event highlights the critical need for checks and balances in AI to prevent the reinforcement of existing biases in areas like hiring and medical diagnosis. The conference features keynote speakers, tutorials, and diverse sessions involving journalists, activists, and educators. Notably, Vector researchers will present two significant papers examining the trade-offs between privacy, utility, and fairness in machine learning. One study investigates differentially private synthetic data in classification tasks, while the other explores privacy guarantees in healthcare predictions. These contributions underscore the ongoing challenges in balancing model performance with ethical considerations such as privacy protection and equitable representation for minoritized groups in high-risk AI applications.
Vector Institute for Artificial Intelligence