David Williamson and Donna LaLonde

Make an impact! This was the theme for JSM 2019 and a challenge to the statistics and data science communities proposed by 2019 ASA President Karen Kafadar. To help meet this challenge, a working group was formed that focused on highlighting the contributions of the statistical sciences that advance science or technology and inform public policy.

A team led by Susan Paddock identified automated driving systems as an area in which statisticians have made an impact and conducted an interview with Feng Guo, Nidhi Kalra, and Maria Terres. The complete interview provides an overview of the relevant issues, assesses the current state of statistical research and practice in this domain, and identifies opportunities for statisticians. Here, we introduce the participants and share key takeaways.

Maria Terres joined Waymo as a data scientist in the fall of 2018. At Waymo, she has developed metrics to evaluate the quality of driving achieved by self-driving cars and leaned on her statistical expertise to ensure statistical rigor in uncertainty estimates. Prior to her time at Waymo, she spent three years at The Climate Corporation developing models to recommend fertilizer rates that optimize farmers’ yields while reducing excessive applications. Her educational background includes a PhD in statistical science from Duke University (2014), where she worked with Alan Gelfand on spatial and environmental modeling. She then held a postdoctoral position at North Carolina State University, working with Montserrat Fuentes.

Nidhi Kalra is a senior information scientist at the RAND Corporation. She previously served as director of RAND’s San Francisco Bay Area office and co-director of RAND’s Center for Decision Making Under Uncertainty. Her research focuses on autonomous vehicle policy, climate change adaptation, and tools and methods that help people and organizations make better decisions amid deep uncertainty. Kalra spearheads RAND’s autonomous vehicle policy work. In 2018, she served as senior technology policy adviser to Sen. Kamala D. Harris. She earned her PhD in robotics from Carnegie Mellon’s Robotics Institute.

Feng Guo is a professor in the department of statistics at Virginia Tech and a lead data scientist at the Virginia Tech Transportation Institute. With dual PhDs in transportation engineering and statistics, Guo has actively engaged in both methodology and practice research on quantitative transportation modeling, especially in traffic safety evaluation such as naturalistic driving studies, transportation infrastructure safety evaluation, advanced vehicle proactive safety device evaluation, and automated driving research.

One of the questions asked in the interview was, “What are the greatest unknowns related to self-driving cars?” Guo said that whether or not self-driving cars are able to provide the promised safety, security, and reliability at massive scale remains an open question. Kalra added that, beyond the technology, was the unknown of the impact of the technology on how we live, work, and engage with each other.

Terres reflected on areas in which statisticians can have the greatest impact on advancing self-driving car safety or technology. She shared that statisticians are advancing many aspects of self-driving technology, from computing metrics that help assess the safety of current software, to assessing subtle changes in self-driving motion planning, to optimizing fleet allocations. She emphasized that keeping the big picture in mind was critical.

We look forward to adding more topics to our impact series. If you have suggestions for topics and individuals, reach out to Donna LaLonde to share your ideas.