Judea Pearl, Chancellor’s Professor of Computer Science at the UCLA Samueli School of Engineering, has been selected to join the Toyota Research Institute’s program on artificial intelligence. With more than $750,000 in funding over the next two years, his research will focus on transforming data science from its current paradigm, “data fitting,” to one that leads to robust and useful data interpretation.
Through this program, the Toyota Research Institute — the automotive giant’s U.S.-based research arm — will lead 35 joint research projects aimed at achieving breakthroughs in automated driving, robotics and machine-assisted cognition. The program will include an investment of more than $75 million in UCLA and 15 other academic institutions.
According to Pearl, “data fitting” is the term used to characterize the current thinking that dominates both statistics and machine-learning cultures. This view is driven by a belief that the secret to rational decisions lies within the data themselves and requires sophisticated data mining. Whereas, those who subscribe to the school of data interpretation view data as an auxiliary means for interpreting reality, the processes that generate the data.
Beyond analyzing data, Pearl said robust AI technologies that incorporate data interpretation can track data from multiple sources, while adjusting for data corrupted by noise, omissions, inadequate collection and even deliberate misreporting.
Pearl, who has received the highest honor in computing, the A.M. Turing Prize, outlined much of his research on this subject in the recently published “The Book of Why.”
In light of the many challenges caused by COVID-19, Pearl noted that such technology advancements could go a long way in preparing for future pandemics, enabling timely public policy directives and societal responses, as well as informing individual health care decisions.
