TRB+人工智能应用中的公平性（演讲PPT）-122页_5mb

TRB Webinar: Equity in Artificial Intelligence Applications

Introduction

The webinar explores equity in AI, focusing on transportation and logistics solutions, including responsible AI implementation, case studies, and strategies for addressing disparities.

Dr. Kofi Nyarko: Responsible AI for Inclusive and Sustainable Transportation

• Current transportation advancements: Electric vehicles, high-speed trains, and autonomous cars are integrated with AI.
• AI's role: Addresses equity issues like accessibility, affordability, and safety by optimizing routing and demand-responsive transit.
• Responsible AI: Emphasizes transparency, accountability, fairness, and human oversight to mitigate biases.
• Case studies: Examples from Google Maps and Via Transportation enhance transportation equity.

Dr. Ziping Wang: Equity in Logistics with Drones

• Focuses on equitable delivery solutions, particularly in rural areas during crises (e.g., pandemic).
• Identifies disparities: Limited access to delivery services in low-density regions.
• Proposed solution: Uses drones in conjunction with trucks for efficient, equity-based delivery routing via genetic algorithms.
• Case study: Implementation in Hereford Zone, Maryland, demonstrating reduced equity thresholds and balanced delivery outcomes.
• Challenges: Including biases in AI algorithms and ensuring equitable incentives.

AI Applications and Equity Challenges

• Non-iid data and adversarial settings can lead to segregation, polarization, and stable biased outcomes in AI systems.
• Strategies for responsible AI deployment involve community involvement, diverse datasets, and ethical guidelines.

Conclusion

The webinar highlights AI's potential to promote equity and sustainability in transportation and logistics, requiring careful implementation to address biases and ensure widespread access.