Faculty engage in innovative research projects

by Abby Mihaly | 2/9/18 3:00am

2818buildingscaroylnsilverstein
by Carolyn Silverstein / The Dartmouth Staff

This article was featured in the 2018 Winter Carnival Issue.

Should Republicans pray for rain on election day? What do black holes tell us about the formation of galaxies? How can we predict future mass genocides? These are the kinds of questions Dartmouth professors, along with their colleagues and students, are working to answer, during the one or two terms a year they aren’t busy teaching.

The College recently ranked 20th out of 200 research institutions worldwide in the Nature Index 2017 Innovation ranking, based on patent contributions and researchers’ influence on their respective fields — and professors agree that Dartmouth’s culture encourages innovation.

When earth sciences professor Marisa Palucis arrived to teach at Dartmouth last July, she was surprised by the interdisciplinary nature of Dartmouth students’ interests.

“The breadth and the interest range of a lot of the students is very wide,” Palucis said.

Palucis noted that her planetary science students are also taking courses in disciplines like anthropology and history and applying their diverse knowledge to all areas of their work.

“If you only ever take geology classes, you’re going to end up always thinking like a geologist,” she remarked. “But if you’re learning principles from the humanities or other areas of science and you bring that to whatever your major, I think that can lead to a lot of innovation.”

Palucis’ students help her with her research on Mars — a planet environmentalists and science fiction enthusiasts alike are interested in. Palucis said scientists can use Mars’ unique surface to help understand how our entire solar system formed.

“On Earth, since we have plate tectonics, we lose a lot of information, because the surface of the Earth gets recycled,” she explained. “On Mars, there [are no] active tectonics, so we can study what the Earth would have been like early in its evolution.”

Palucis said she hopes to ascertain at what time Mars had water and how much there was on the planet. After a five-year stint working for NASA’s Mars Science Laboratory, she has continued to work in that field, even helping direct Mars rovers. She also analyzes data such as delta deposits, using laws and empirical relationships known on Earth to “back-calculate” how water may have existed on Mars.

Like Palucis, government professor Benjamin Valentino agreed that Dartmouth’s academic culture breeds innovation, in particular because of the connections that exist within and between departments.

“Because of our small size, faculty tend to know each other across disciplines and departments in a way that maybe doesn’t happen at much larger institutions,” said Valentino, adding that a single department at larger university may have upwards of 90 people.

But at Dartmouth, Valentino explained, not only do professors work together on campus, but they exist in the same community away from school.

“We know faculty from every corner of [the College], either because we meet them on campus, or … we meet them when our sons and daughters are playing soccer on the field and we get to talking,” he said, which leads to conversations about common interests, and eventually, interdisciplinary work.

Interdisciplinary idea-sharing has benefited Valentino’s research. He co-founded the Early Warning Project, a project that calculates countries’ risk of mass atrocity. Valentino said he utilizes students’ knowledge in other fields to help further his work on the project.

“Building these kinds of [risk] models has been part of my work for 20 years,” he explained. “But communicating the results of statistical models to the general public, high school students that come to the web page, that’s not something I’ve done before. Nor building a website.”

He added that students and researchers in other disciplines have helped him learn how to best communicate his results on the web page.

Valentino began performing risk-consulting work for the United States government in the 1990s, but he said he always thought it a shame that his forecasts weren’t accessible to the public.

“There are people besides the U.S. government that could benefit from knowing what countries are at highest risk,” he said.

So Valentino decided to make the risk data available to the public. In collaboration with the Simon-Skjdot Center for the Prevention of Genocide at the United States Holocaust Memorial Museum in Washington, D.C., Valentino launched the Early Warning Project. The museum later decided to host the project and hire full-time employees to support it.

Valentino said that he appreciates the connection with the public that the project encourages.

“I’ve started to get emails on a fairly regular basis from … concerned citizens, people who come from countries that are high risk,” he said.

Valentino added that he has been able to have constructive conversations with those who criticize him, as well as activist organizations looking for a way to make a difference in countries the project marks as high risk.

“That’s exactly what I had hoped would happen,” he said.

Computer science professor Lorenzo Torresani shares Valentino’s passion for spreading his work to far corners of the world. Torresani is currently in Ghana on a Fulbright Scholarship, where he is teaching machine learning at Ashesi University.

Torresani’s current research surrounds video analysis, which he said he first became involved in during his undergraduate education in Italy, where his thesis tackled lip-reading technology that aimed to improve speech recognition. After working at three different start-ups and pursuing a Ph.D. at Stanford University, Torresani came to Dartmouth to pursue an academic career.

At Dartmouth, Toressani works with “deep models” which consist of “many, many layers, and each layer extracts some features or representation from the preceding layer,” he sayid

Traditional artificial intelligence work, such as the technology Torresani had used for his lip-reading research, required step-by-step manual coding. Recent developments in deep modeling mean machines now have the capacity to decide what features are relevant from a given input.

For Torresani, the global impact of his work bears great importance.

“I’ve always had this dream of going somewhere I could teach my classes and potentially make a bigger impact than even at Dartmouth, because I would be able to really give opportunities to students who don’t have these opportunities,” Torresani said.

Torresani also emphasized the positive experience he has had teaching at Ashesi.

“It’s a great opportunity for me to be here, and to bring my expertise and inspire some of these amazing students and maybe convince a few of them to pursue graduate studies, maybe at Dartmouth,” he said.

After his work in Ghana, Torresani will spend six months working at Facebook. It’s a two-way exchange, Torresani explained: Facebook can offer him and his students as much as he can offer them. In particular, the company has access to powerful computational technology, which allows for larger and longer-running models and access to data not yet available in academia.

“I have good students who can do internships at these companies, and conversely, they have great resources in terms of computer resources, clusters [and] computational power,” he said.

All of the professors interviewed shared Torresani’s high regard for their student researchers. Government professor Yusaku Horiuchi said what he most likes about Dartmouth is the “excellent students.”

Horiuchi said he is working on upwards of 10 papers and three books at the moment, including a study on students’ preferences on college diversity. Students are involved in many of these projects and often contribute a great deal, he said. Sometimes, a student who began as a research assistant eventually becomes a collaborator due to substantial contribution.

Horiuchi recently co-authored an article titled, “Why Should the Republicans Pray for Rain? Electoral Consequences of Rainfall Revisited,” which was published late last year. The paper was written in response to an earlier study that argued that rainfall in the United States was advantageous to Republican candidates. The original paper explained the phenomenon based on voter turnout: when it rains, a higher number of Democrats decide to stay home from the polls, due to firmer party support among Republicans.

Horiuchi and his colleague, Woo Chang Kang of Australian National University, conducted a “replication and extension exercise” to examine the study’s findings, Horiuchi said. They repeated the same experiment, but changed the way they analyzed the data. The original study made a key assumption: that rainfall doesn’t change voter preference. But Horiuchi and his co-authors offered a different conclusion.

“We argue that … it’s possible that when it rains, people’s preference may change,” Horiuchi said.

Their analysis showed about 1 percent of voters likely changed their preference in rainy weather.

Horiuchi said their conclusion fell in line with previous findings in other disciplines: “When it rains, many, many studies in psychology and health science [found] that the rainfall affects people’s mood … When people are not in a good mood, they tend to be risk averse.”

In political science, Horiuchi said Democrats are often perceived as riskier choices for office. This risk aversion during poor weather supported Horiuchi’s findings that rain makes Americans more likely to vote for Republicans.

Like Horiuchi, physics and astronomy professor Ryan Hickox said that students are essential to his work at Dartmouth, and that undergraduate students, graduate students and postdoctoral scholars are all involved in his research.

“The work that I do would not be possible at all without these fabulous people,” he said.

Hickox is an extra-galactic astronomer, specializing in black holes and what they indicate about how galaxies formed and how they’ve changed over time. Hickox and his team observe the radiation that black holes produce as they grow, such as x-ray and infrared radiation, looking for “signatures” of these black holes in telescope images. His team uses data and images from other researchers, including data from telescopes in space, as well as collecting some data at the Southern African Large Telescope and MDM Observatory in Arizona.

If a galaxy is the size of Earth, Hickox said, the black hole at its center would be the size of a centimeter, and it would weigh anywhere from a million to over a billion times the mass of the sun. Hickox added that he is encouraged by the new information his team has uncovered regarding the location and behavior of these black holes, thanks to innovative techniques.

“It’s like an archaeological dig,” he said. “We’re finding stuff that no one had ever had the right tools to be able to see.”

Hickox explained that new x-ray observatories allow for better visibility through dense clouds of gas, and new techniques that combine data together help find hidden black holes in unprecedented ways.

Hickox’s research is just one example of Dartmouth professors making a splash in the greater academic world. Though Dartmouth’s smaller size may suggest Dartmouth professors are less visible among the academic community, Hickox said he has found quite the contrary.

“One of the things that I’ve loved about Dartmouth is that I’ve felt completely connected to the broader landscape of scientists working in my area,” he said. “For me, Dartmouth has been really great at allowing us to innovate on a large scale, as part of the whole community of researchers.”

Hickox expressed gratitude that the College’s environment encourages faculty and students to discuss their work, even across departments.

“It makes your mind flexible,” he said.

Though Valentino didn’t expect this culture of innovation when he chose to make Dartmouth home, he said it’s a “happy byproduct” of the kind of place Dartmouth is — small, close-knit and interdisciplinary.