Photo of Dr. Yan Liu, who received honors from both the AAAI and the IEEE.
It was the day before Thanksgiving, and Professor Yan Liu was at her desk, working as usual. While everyone else was prepping for turkey and time off, she was scrolling through email. But this year, she was about to get something to be truly thankful for.
What she found in her email wasn’t a holiday greeting. It was two pieces of news that most researchers spend an entire career hoping to receive just once, arriving within the same hour. Both the Association for the Advancement of Artificial Intelligence (AAAI) and the Institute of Electrical and Electronics Engineers (IEEE) had named her a fellow.
Liu is Professor of Computer Science, Electrical and Computer Engineering, and Biomedical Sciences and the Fletcher Jones Foundation Chair, with joint appointment at the Thomas Lord Department of Computer Science and the Ming Hsieh Department of Electrical and Computer Engineering at the USC Viterbi School of Engineering and the USC School of Advanced Computing.
Her AAAI citation spotlights her work on machine learning for “time series and spatio-temporal data analysis.” Machine learning and AI may be a common term these days, but those other terms certainly are not. Time series refers to data that changes over time. Traffic flow is one example of time series. The traffic flow at 10:40 a.m., compared with conditions at 10:15 a.m., helps researchers understand patterns as they unfold over the time and develop accurate forecasting models.
Spatio-temporal data analysis takes it a step further, analyzing both time and location. In the example of traffic flow, one considers both the traffic flow in the original Santa Monica location as well as relevant nearby areas, such as Culver City and Westwood. From there, machine learning models can put together much more accurate forecasts for time and/or space-dependent things like weather, traffic, and stock prices.
“Time series is an extremely important problem, because most of the previous machine learning models assume the data is actually independent,” Liu said.
When discussing the importance of time and space, she calls on a quote from Albert Einstein: “A human being is part of a whole, called by us the Universe—a part limited in time and space…” To paraphrase the legendary scientist: if we don’t consider time and space, we’re missing the whole picture of our world.”
Besides fundamental contributions in machine learning models, Liu’s work on time series and spatiotemporal data analysis have enabled a variety of applications, such as transportation, power, earth science, health science, biology and many others. This balance between academia and industry is something Liu thinks about a lot.
“Viterbi is excellent in technology transfer, in terms of how a faculty can actually have research from the premier stages in research to industry deployment,” Liu said. “Before joining Viterbi [in 2010], I was working in the industry at IBM Research. Coming to Viterbi, I was welcomed by this vibrant and collaborative community, where I would be able to conduct research on the most critical problems in science and engineering, and translate the research results to industry solutions. This culture also helped shape my career path.”
Liu’s IEEE citation is a more general recognition for her “contributions to the methodology and application of machine learning and data mining.” Of course, those are fields that have changed drastically in the last few years. Large language models and mainstream AI platforms like ChatGPT have made machine learning more accessible than ever, and Liu is keeping a watchful eye over these new developments.
Liu’s new work includes both continuing to advance machine learning and overseeing how it’s used. She co-chairs the USC Institute on Ethics & Trust in Computing (IETC).
“I do not think it’s a good idea for researchers in academia to only follow the industry trend,” Liu said. “There are a large number of new research frontiers and forward-looking topics we can work on in academia. One of the important directions the work we generated from is how to ensure these large language models will be aligned with our human ethics.”
Published on March 27th, 2026
Last updated on April 6th, 2026