I Am Data Science | David Danks

Faculty Feature Article Series: I Am Data Science

with Professor David Danks, Ph.D.

by Trista Sobeck & Bobby Gordon

David Danks, Ph.D. is not interested in drawing boundaries around any discipline. “I prefer to let a thousand flowers bloom,” he says when asked about what his definition of data science is. “If something counts as ‘data science’ for someone, then it counts for me,” he concludes.

This would be something you may expect from someone with a background in philosophy. And you would be right. His work lands him directly at the intersection of philosophy, cognitive science, and machine learning.

Danks received an A.B. in Philosophy from Princeton University and a Ph.D in Philosophy from University of California, San Diego. He came to the Data Science field because of his interest in how a ‘cognitive agent’ (human, machine, or animal) can understand and make decisions in an incredibly complicated world.

Welcome to Data Science

But how, exactly, did he land at data science—a field where philosophy seems an odd bedfellow? As computers become more and more sophisticated and artificial intelligence makes its way into the typical home (think “Alexa”), philosophy and the ability to make decisions becomes a priority.

“My research was on computational cognitive science and philosophy of science. I was doing some machine learning—though we usually called it ‘automated reasoning’ back in those days—but it was mostly to develop algorithms and frameworks that could be used to better understand human possibilities,” he explains.

“Over time, though, I began to see that I should have a broader focus on ‘cognitive agents’ in general, not just humans, and so became more interested in various forms of AI,” he explains of his career path.

He emphasizes that after 18 years at Carnegie Mellon, he’s excited to have the opportunity to pursue all these issues in an environment that prioritizes the importance of data and the understanding it can provide. That environment is indeed UC San Diego’s Halıcıoğlu Data Science Institute.

He admits that this does not completely describe what he was thinking as he moved to data science. He was thinking more about puzzles. “My research tends to be highly driven by puzzles, rather than any big themes. Much of my work is the result of looking at a puzzle, thinking ‘that seems interesting,’ and then spending some time trying to make progress on it, all without worrying about how the puzzle fits into some grander scheme,” he explains.

He continues to say that he can see common themes in retrospect. But his research has been more defined by unexpected twists and turns. This comes from either casual conversations or even a shift in interests.

Solving Problems

Perhaps it’s his own experience in making decisions that is reflected in his work. “I’m interested in how people and machines can make good decisions in a complicated world. In my view, one of the key tools is causal knowledge; if we want to change or control our world (either directly or in partnership with an AI), then we need to know what causes what,” he explains.

“As data science increasingly pervades every aspect of our lives and society, we need to ensure that it is used responsibly, and that the next generation recognizes that good data science is responsible data science,” he emphasizes.

“My work in this area ranges from the development of new data science practices, analyses of the ethical and societal impacts of data science, and close collaboration with companies and government agencies to help them do data science better,” he concludes.

Although data science can be very mathematical and computational, they are only tools to help with understanding. “And since data are about things in the world, it’s important to know about domain(s) beyond data science itself. While you must know the tools, frameworks, and methods of data science, it is also important to understand the questions that other people are trying to answer with data science.”

As Dr. Danks continues to understand the questions, there will always be someone asking them.

Connect with Professor David Danks

HDSI/UC San Diego Faculty Profile: https://datascience.ucsd.edu/directory/name/david-danks/

David Danks’ Website: https://www.daviddanks.org/