LLNL’s Data Science Summer Institute hosts student interns from Japan

Lawrence Livermore National Laboratory’s Data Science Summer Institute (DSSI) hosted summer student interns from Japan on-site for the first time, where the students worked with Lab mentors on real-world projects in artificial intelligence (AI)-assisted bio-surveillance and automated 3D printing.

From June to September, the three students — Raiki Yoshimura, Shinnosuke Sawano and Taisei Saida — lived in rental apartments near the Lab and worked at the Lab on different data science projects using electronic health records and neural networks trained on experimental data.

Sponsored by Japan’s Agency for Medical Research and Development (AMED) and the Japan Science and Technology Agency (JST), the relationship between the Data Science Institute (DSI) and the Ministry of Japan stems from a series of agreements that arose in the wake of the Fukushima nuclear accident, to conduct academic exchanges and expand scientific collaboration. The DSSI opportunity began in 2019 but had to go all-virtual due to the COVID-19 pandemic. This summer was the first time that students from Japan were brought on-site.

DSI Director Brian Giera said the relationship is “part of a continuing trend of Livermore positioning itself to partner with Japan via scientific connections” and one the Lab hopes to expand in the coming years.

“The Data Science Institute has found itself an example of the U.S.’s posture in using science and technology to partner with geopolitically relevant allies,” Giera said. “Livermore is establishing itself as a global leader in data science, and it’s very clear that Japan has oriented in the direction of producing candidates that are highly attractive to us. The collaboration is showing that we are helping the students steer their curricular or scientific focus and sharing the messaging to academia [in Japan] that indeed, data science is relevant. Livermore can help be a partner in realizing that vision, and students are the boots on the ground of having that occur.”

Yoshimura, Sawano and Saida had never been to the U.S. before arriving in Livermore for their internships. The DSSI program helped the students find housing and matched them with mentors based on the students’ technical knowledge. Although the students had almost no knowledge of LLNL prior to their internships (outside of photos on the Internet) and were challenged with cultural and language barriers, they were able to acclimate to working in a national laboratory environment and had successful experiences.

While the students said it took some time to get used to the Lab’s security procedures and American food, they said they found working with their mentors rewarding and eye-opening.

Sawano, a Ph.D. student at the University of Tokyo with a clinical background in cardiology and treating patients with cardiovascular disease, said he thought his internship was a perfect opportunity to combine his knowledge of data science and medicine. His work is supported in Japan by AMED, and though he didn’t know what national labs did prior to his internship, he considers LLNL “a fascinating option” for a career after he finishes his Ph.D.

Working with his lead mentor, LLNL computer scientist Priyadip Ray, and Lab researchers Andre Goncalves and Jose Cadena-Pico, Sawano applied machine learning to electronic health records obtained from Kaiser Permanente, for a bio-surveillance project funded by the Department of Homeland Security. The Lab is developing AI tools to perform faster diagnostics to allow scientists to detect biological threats earlier, thus providing more time to develop possible countermeasures, according to Ray.

“As these tools advance, and if we are able to look at the clinical record of everybody in this country in real-time, then we can detect these kinds of anomalies much faster,” Ray said. “That would give us more time for developing countermeasures.”

Sawano said he hopes the research can someday make it into clinical practice.

“Collaborating with the professionals on the team is a really good experience for me because I’m usually analyzing data by myself and I check my data by myself, but in this team, Priyadip, Jose and Andre can check my coding and my results. As a result, our output is bigger than I do myself,” Sawano said. “We were really fortunate to get great mentors; they gave me a great deal of advice and support. It will take time to deliver our results to society, but I’m excited about the potential output of our research.”

Ray, who also worked with Yoshimura on a project using neural networks to evaluate and predict the impact of gene interactions on the viral load of the HIV virus, said the students did a “remarkable job” in contributing to advancing the research, which could have concrete impacts to public health in the future.

Yoshimura, who attends Nagoya University and studies biology, said he recently began using machine learning to predict clinical outcomes and was drawn to the DSSI internship to expand his skill set and apply his knowledge to large-scale datasets.

“This internship experience has been great because they have very big data sets here that I can apply deep learning to,” Yoshimura said. “In Japan, we have to pay for data and the data is very little so I can’t apply a graph or something like that.”

Saida, who studies civil engineering in Japan and wants to be a university professor, worked with his mentor — LLNL postdoctoral researcher Aldair Gongora — on a project on self-driving labs for additive manufacturing, where machine learning approaches are used to help decide which experiments to do next to speed up manufacturing processes with the goal of eventually deploying these approaches on fully automated robotic systems.

“There has been a big push toward autonomous experimentation or self-driving labs, and I think that our work really puts us in a position to continue contributing to those fields in a way that really adds value to the modeling and decision-making components of these systems,” said Gongora, who works in the Analytics for Advanced Manufacturing group in the Materials Engineering Division. “With all the tools that we now have in data science, it’s blurring the lines between data scientists, chemists, physicists and engineers.”

Gongora said he found Saida’s background in programming, machine learning and data science a perfect fit for the project and marveled at how his mentee was able to pick up algorithms and implement them at an extremely fast pace.

Saida said he learned how to use and program robots and integrate the hardware with the software on 3D printing machines, and that he found his internship valuable in expanding his knowledge of Bayesian optimization and mechanical engineering.

“It’s the first time for me researching outside of Japan, so it’s been a great experience to research with people in the U.S.,” Saida said. “I had a really positive experience working on these projects, especially 3D printing.”

During their summer in Livermore, the students explored the Bay Area, enjoying local pizza and ramen restaurants, seeing tourist attractions and even taking a trip to Yosemite National Park. They also attended regular DSSI community events, meet-ups and ice cream socials, where they got to know their fellow intern cohort.

The mentors said they found the experience just as valuable as the students did. Gongora, who came to the Lab as a foreign national himself, said he resonated with the cultural challenges the students faced and said the opportunity epitomized the strength of diversity in science.

“The benefit for me has really been being able work with someone from another country and learning more about Japan; learning about their lives, how their academic journey differs from education here in the U.S., and really finding the commonalities and differences,” Gongora said. “I'm taking away a lot of new perspective leveraging the expertise that [Saida] was able to bring to the project, both in terms of how very skilled he was at the data science concepts and the brainstorming of new ideas from the civil engineering perspective. Time and time again, the Lab teaches me that it's really through the diversity in thought that these interdisciplinary and multidisciplinary ideas emerge, and I think we’re stronger for it.”

Fellow mentor Ray added that the experience was “an incredible opportunity for the Lab to get some of the best students from Japan” and that he looks forward to mentoring more students from that country.

“At the Lab, we are trying to solve very impactful and challenging problems, and we want the best teams to work on these problems,” Ray said. “There are communication barriers and cultural barriers, because many of them are coming for the first time, but once they're on site, I see them working very hard to overcome all the challenges and really contribute to the mission and push things forward.”