Tag: graduate research

With so many different career options in life, how do you know that you’ve found the right one for you?

Graduate students Edric Tam and Andrew McCormack, when asked what they hope to be doing in ten years, said they’d choose to do exactly the kind of work they’re doing right now – so clearly, they’ve found the right path. Tam, who obtained his undergraduate degree in Biomedical Engineering, Neuroscience and Applied Mathematics from Johns Hopkins University, and McCormack, who came from the University of Toronto with a degree in Statistics, are now 5^th-year PhD students in the Department of Statistical Sciences. Tam works with Professor David Dunson, while McCormack works with Professor Peter Hoff, and both hope to pursue research careers in statistics.

Research interests

For the past five years, McCormack has been doing more theoretical research, looking at how geometry can lend insights into statistical models. The example he gives is of the Fisher information matrix, a statistical model that many undergraduate statistics majors learn in their third or fourth year.

Tam, meanwhile, looks at data with unique graphical and connectivity structures that aren’t quite linear or easily modeled, such as a brain connectome or a social network. In doing so, he works on answering two questions – how can you model data like this, and how can you leverage the unique structure of the data in the process?

What both Tam and McCormack like about the field of statistics is that, as Tam puts it, “you get to play in everyone’s backyard.” Moreover, as McCormack says, the beauty of theoretical research is that, while it’s certainly more time-consuming and incremental, it is often timeless, giving insight into something previously unknown.

On walking the research path

What does it take to be a successful PhD student? Both McCormack and Tam agree that a PhD is just a degree – anyone can get one if you work hard. But what sustains both through a career in research is a passion for what they do. Tam says that “you need inherent motivation, curiosity, passion, and drive.” McCormack adds that it helps if you work on problems that are interesting to you.

Tam, who spent some time in a biomedical engineering lab during his undergraduate years, remembers reading about math and statistics the entire time he was there, which signaled to him that maybe, biomedical engineering wasn’t for him. McCormack’s defining moment occurred in the proof-based classes he took while as an undergraduate. He initially wanted to pursue a career in finance, but he quickly became enamored by the elegant precision of mathematical proofs – “even if all you’re proving is that 1+1=2!”

“You need inherent motivation, curiosity, passion, and drive.”
Edric tam, on what it takes to pursue a career in research

Even with passion for what you do, however, research can have its ups and downs. McCormack describes the rollercoaster of coming up with a new idea, convinced that “this is a paper right here”, and then a day later, after he’s had time to think about the idea, realizing that it isn’t quite up to the mark. Tam, who considers himself a pretty laidback person, sometimes finds the Type A personalities in research, as in any career field, too intense. Both McCormack and Tam prefer to not take themselves too seriously, and both exude a love for – and a trust of – the process.

Reflections on the past and the future

Upon graduation, McCormack will move to Germany to pursue a post-doc before beginning a job as Assistant Professor in Statistics at the University of Alberta. Tam will continue his research at Duke before applying to post-doc programs. In reflecting on their paths that have brought them till now, both feel content with the journey they’ve taken.

Tam sees the future in front of him – from PhD to post-doc to professorship – as “just a change in the title, with more responsibility”, and is excited to embark on his post-doc, where he gets to continue to do the research he loves. “It doesn’t get much better than this,” he laughs, and McCormack agrees. When McCormack joins the faculty at the University of Alberta, he’s looking forward to mentoring students in a much larger capacity, although he comments that the job will probably be challenging and he’s expecting to feel a little bit of imposter syndrome as he settles in.

When asked for parting thoughts, both Tam and McCormack emphasize that the best time to get into statistics and machine learning is right now. The advent of ChatGPT, for example, could replace jobs and transform education. But given their love for the field, this recommendation isn’t surprising. As Tam succinctly puts it, “given a choice between doing math and going out with friends, I would do math – unless that friend is Andy!”

Post by Meghna Datta, Class of 2023

Traveling With Friends Helps Even Mixed-Up Migrators Find Their Way

By Robin Smith

On December 6, 2022

In Animals, Behavior/Psychology, Biology, Climate/Global Change

North American monarch butterflies migrate each winter to just a few mountaintops in central Mexico, with help from an internal compass that guides them home. New computer modeling research offers clues to how migrating animals get to where they need to go, even when their magnetic compass leads them astray. Credit: Jesse Granger, Duke University

DURHAM, N.C. — Some of us live and die by our phone’s GPS. But if we can’t get a signal or lose battery power, we get lost on our way to the grocery store.

Yet animals can find their way across vast distances with amazing accuracy.

Take monarch butterflies, for example. Millions of them fly up to 2,500 miles across the eastern half of North America to the same overwintering grounds each year, using the Earth’s magnetic field to help them reach a small region in central Mexico that’s about the size of Disney World.

Or sockeye salmon: starting out in the open ocean they head home each year to spawn. Using geomagnetic cues they manage to identify their home stream from among thousands of possibilities, often returning to within feet of their birthplace.

Now, new research offers clues to how migrating animals get to where they need to go, even when they lose the signal or their inner compass leads them astray. The key, said Duke Ph.D. student Jesse Granger: “they can get there faster and more efficiently if they travel with a friend.”

When their internal compasses go bad, migrating animals like these sockeye salmon don’t stop to ask directions. But they succeed if they stay with their fellow travelers. Credit: Jonny Armstrong, USGS

Many animals can sense the Earth’s magnetic field and use it as a compass. What has puzzled scientists, Granger said, is the magnetic sense is not fail-safe. These signals coming from the planet’s molten core are subtle at the surface. Phenomena such as solar storms and man-made electromagnetic noise can disrupt them or drown them out.

It’s as if the ‘needle’ of their inner compass sometimes gets thrown off or points in random directions, making it hard to get a reliable reading. How do some animals manage to chart a course with such a noisy sensory system and still get it right?

“This is the question that keeps me up at night,” said Granger, who did the work with her adviser, Duke Biology Professor Sönke Johnsen.

Multiple hypotheses have been put forward to explain how they do it. Perhaps, some scientists say, migrating animals average multiple measurements taken over time to get more accurate information.

Or maybe they switch from consulting their magnetic compass to using other ways of navigating as they near the end of their journey — such as smell, or landmarks — to narrow in on their goal.

In a paper published Nov. 16 in the journal Proceedings of the Royal Society B, the Duke team wanted to pit these ideas against a third possibility: That some animals still manage to find their way, even when their compass readings are unreliable, simply by sticking together.

To test the idea, they created a computer model to simulate virtual groups of migrating animals, and analyzed how different navigation tactics affected their performance.

The animals in the model begin their journey spread out over a wide area, encountering others along the route. The direction an animal takes at each step along the way is a balance between two competing impulses: to band together and stay with the group, or to head towards a specific destination, but with some degree of error in finding their bearings.

The scientists found that, even when the simulated animals started to make more mistakes in reading their magnetic map, the ones that stuck with their neighbors still reached their destination, whereas those that didn’t care about staying together didn’t make it.

“We showed that animals are better at navigating in a group than they are at navigating alone,” Granger said.

Even when their magnetic compass veered them off course, more than 70% of animals in the model still made it home, simply by joining with others and following their lead. Other ways of compensating didn’t measure up, or would need to guide them perfectly for most of the journey to accomplish the same feat.

But the strategy breaks down when species decline in number, the researchers found. The team showed that animals who need friends to find their way are more likely to get lost when their population shrinks below a certain density.

Prior to the 1950s, tens of thousands of Kemp’s ridley sea turtles could be seen nesting near Rancho Nuevo, Mexico on a single day. By the mid-1980s the number of nesting females had dropped to a few hundred.

“If the population density starts dropping, it takes them longer and longer along their migratory route before they find anyone else,” Granger said.

Previous studies have made similar predictions, but the Duke team’s model could help future researchers quantify the effect for different species. In some runs of the model, for example, they found that if a hypothetical population dropped by 50% — akin to what monarchs have experienced in the last decade, and some salmon in the last century — 37% fewer of the remaining individuals would make it to their destination.

“This may be an underappreciated aspect of concern when studying population loss,” Granger said.

This research was supported in part by the Air Force Office of Scientific Research (FA9550-20-1-0399) and by a National Defense Science & Engineering Graduate Fellowship to Jesse Granger.

CITATION: “Collective Movement as a Solution to Noisy Navigation and its Vulnerability to Population Loss,” Jesse Granger and Sönke Johnsen. Proceedings of the Royal Society B, Nov. 16, 2022. DOI: 10.1098/rspb.2022.1910

Following the people and events that make up the research community at Duke

Tag: graduate research

What does it take to be a successful PhD student? Two grad students in statistics weigh in

Traveling With Friends Helps Even Mixed-Up Migrators Find Their Way