We have the distinct privilege of honoring our Top Ten Scholar-Athletes. These students are the five male and five female senior student-athletes with the highest cumulative grade point averages. The ten students being honored include Chancellor’s Award winners, undergraduate researchers, postgraduate scholarship winners, honors students and Phi Beta Kappa inductees. They will graduate with an average GPA of 3.91.
UNC Top 10 Scholar-Athletes
Jamie is a member of the gymnastics team and is from Park City, Utah. She is majoring in Biology with minors in Chemistry and Neuroscience. Jamie has been inducted into Phi Beta Kappa and has been on the Dean’s List each of her semesters at UNC. She has been conducting research on air pollution and lung health at the Alexis Lab in the Center for Environmental Medicine, Asthma and Lung Biology. After graduating this month, she will continue her work full time at the Alexis Lab while she is applying to medical school. In five years, she will be about to start her residency.
Andrew is from Ann Arbor, Michigan and is a member of the men’s swimming and diving team. He is a double major in Health Policy & Management and Economics. Andrew has been a member of our 4.0 Club and has also been named to the Dean’s List each semester. He has conducted research for the Center for Health Equity Research on the relationship between opioid dependency, incarceration and HIV/AIDS. After graduation, Andrew will be moving to Chicago to work for Guidehouse Consulting. In five years, he will be enjoying city life, establishing himself in his field and finding time to ski, swim and explore the outdoors.
Brian is from Allendale, New Jersey and is on the men’s fencing team. He is receiving a degree in Political Science with a minor in Information Systems. Brian has been named to the Dean’s List six semesters and has twice been the recipient of the Top GPA Award as a member of the men’s fencing team. He had an opportunity to study abroad in Paris in the Fall 2019 semester. After graduation, Brian will work at a law school in New York City and apply to law school in 2022. In five years, he hopes that he will be finishing up law school and will have found a creative outlet.
Jackie is on the women’s fencing team and is originally from Niskayuna, New York. She is double majoring in Public Policy and Environmental Sciences with the Quantitative Energy Systems track. She was inducted into Phi Beta Kappa, has contributed extensively to the APPLES Service-Learning program and also studied abroad in Israel. After graduation, she plans on working on a political campaign until the November elections and then finding a job in environmental/renewable energy policy. In five years, she would like to be working for a state government where she can design and ensure compliance with renewable energy and energy efficiency standards.
A.J. is from Batavia, Illinois and is a member of the wrestling team. He is majoring in Biomedical and Health Sciences Engineering and minoring in Chemistry. He has been inducted into Phi Beta Kappa and named the winner of the Wells Fargo Postgraduate Scholarship. A.J. is also on track to graduate with a perfect 4.0 GPA! After he finishes his degree at Carolina, he will be attending Duke School of Medicine in the fall. In five years, he will be starting a residency in orthopedic surgery.
Liz is on the women’s basketball team and from Chapel Hill. She is an Exercise and Sport Science major with a Sport Administration concentration and a minor in Coaching Education. Liz was inducted into Phi Beta Kappa, named the Athletic Director Scholar-Athlete for women’s basketball and has been on Dean’s List each semester. She also had an opportunity to study abroad in London last summer. After graduation, she will attend graduate school and pursue a master’s in Sport Administration. In five years, she will be on the staff of a women’s basketball program and giving back to future generations of female student-athletes.
Hunter Sheridan Hunter is originally from Charlotte and is on the football team. He is majoring in Business Administration. Hunter has been named to the Dean’s List each semester, was a member of the 4.0 Club and was the Athletic Director Scholar-Athlete for football. He also participated as both a student and leader in Kenan-Flagler’s Finance Trek where they visit 10 New York City investment banks in three days. After graduation, Hunter will work in investment banking with Bank of America Merrill Lynch in New York. In five years, he hopes to be still pursuing investment banking or private equity in either New York or Charlotte.
Ashley is on the women’s cross country and track and field teams and is originally from Clemmons, North Carolina. She will earn a Bachelor of Science in Computer Science with a Hispanic Studies minor. Ashley has been part of Honors Carolina and was inducted into Phi Beta Kappa. She also received a Chancellor’s Award this year after being named the winner of the Irene F. Lee Award which is presented to the woman in the senior class who is judged most outstanding in leadership, character and scholarship. After she graduates, Ashley will pursue a Master of Science in Computer Engineering at Virginia Tech. In five years, she will be reshaping and transforming the tech industry.
Ezra is from Washington, D.C. and is a member of the fencing team. He is graduating with a double major in Women’s and Gender Studies and Hispanic Literature and Culture. Ezra was honored with two Chancellor’s Awards in 2020. He won the Mary Turner Lane Award in Women’s and Gender Studies and The Sterling A. Stoudemire Award for Excellence in Spanish. After graduation, Ezra will be working as a community organizer in Durham or Washington, D.C. In five years, he will be living in Ecuador and working in community organization and empowerment for the Ecuadorean people.
Caroline is originally from Sandy, Utah and is on the rowing team. She is double majoring in Business Administration and History with a Philosophy, Politics, and Economics minor. She has been inducted into Phi Beta Kappa and was a James M. Johnston Scholar. She completed a Burch Honors Research Seminar in London and Ireland and a Kenan-Flagler Global Immersion in Indonesia and Singapore. After graduation, she will be a financial analyst at the Walt Disney Company. In five years, she will still be a Tar Heel!
In a small corner of Sitterson Hall sits a fleet of pint-size cars that can see and navigate independently, winning races for the team of UNC computer science students that created them. While the stakes are low for these high-tech toys, it’s a completely different game when applied to full-size vehicles in the real-world — the application of professor Parasara Sridhar Duggirala’s research.
Abel Karimi and Charlotte Dorn stand huddled over a laptop, poring over the data streaming in from the toy-sized vehicle sitting next to them. An error pops up over and over somewhere in its circuitry, triggering its emergency brake system and preventing the car from moving forward for more than a few seconds.
The team isn’t sure whether it’s a short in the wiring or something wrong with the car’s wireless receivers, but they are determined to find the cause. Sometimes working with a car that can drive itself can seem like magic. Other times it can feel like a nuisance with a mind of its own.
On its surface, the team’s creation resembles a remote-controlled car with a small body and big, soft tires. It has been modified with custom parts, including a laser sensor that functions as the car’s “eyes” and allows it to drive itself around a track with the data it collects. These data are a series of distances gathered by a light detection and ranging sensor, or LiDAR, fed into an algorithm to tell the car if it is heading toward an obstacle and must turn.
UNC computer scientist Parasara Sridhar Duggirala oversees Dorn, a sophomore, and Karimi, a graduate student, in the autonomous vehicle lab. He acknowledges that there is a lot of hype surrounding self-driving cars. “The hype being that ‘oh, it’s going to come in three years or four years.’ The truth is, designing autonomous vehicles is a hard problem,” Duggirala says. “If you want some sort of safe autonomous vehicle, it is going to take time.”
For modified R/C cars navigating a track made out of cardboard walls, the stakes are low and the problems encountered by the algorithms are simpler, but as Duggirala points out, “when you think of deploying it in a mass scale, the game is completely different.”
The average American drives a little over 13,000 miles a year according to the US Department of Transportation. In 2014, Americans drove a total of 3.04 trillion miles — or half a light-year.
In the future, more of that total driving distance is expected to be covered by self-driving cars, Duggirala says. Even if the self-driving car makes one mistake every billion miles, though, that still amounts to 3,040 mistakes every year.
This risk factor is one of the major reasons why a nationwide rollout of the technology hasn’t come sooner.
That’s not to say that running autonomous vehicles at this scale is without its benefits. “Through the process of researching how to drive these small cars, you learn about how the real cars — [from] companies like Tesla — are operating in real-time,” Dorn says. “Even though a lot of it is small scale, you can kind of get a bigger picture.”
In 2019, the UNC computer science team behind these autonomous vehicles, which included Dorn and Karimi as well as fellow car designers Nathan Otterness and Tanya Amert, participated in an automated vehicle racing competition in Montreal. Titled F1Tenth, the competition is an annual event started by the University of Pennsylvania that brings together 1/10th scale cars from computer science programs around the world to compete against one another seeing who has the best build and the best code.
“We put it on the track for the first time and immediately it just rushed all the other teams and it was much to our surprise. We hadn’t tested it that thoroughly, so I think those first few laps when we realized how well it was doing was really an amazing experience,” Dorn says.
The competition is all in good fun and involves plenty of conversation with other schools to gain insights into problem-solving and development. After the races ended, teams shared the code which ran each car on an open-source platform so they could see where their programming differed.
Because of the program’s success, Duggirala envisions a platoon of autonomous vehicles based on their current designs. The team has built three of the F1Tenth cars and plans to build two more prototypes from the University of Washington. The goal will be to have the cars communicate with each other and travel as a unit.
Beyond that, he says, they can scale up the project and bring the algorithms to full-scale electric vehicles that can drive around campus. Universities with engineering programs like those at MIT and Stanford already offer full scale autonomous vehicles with which graduate students and researchers can test their ideas.
For miniature autonomous vehicles like those built by Duggirala’s team, the LiDAR scanners the cars use to “see” are only getting cheaper, which means more schools should soon be able to join UNC among the handful of schools nationwide with similar programs.
“I had done some robotics before but not in a research environment, Dorn says. “Being introduced to this lab and being able to do these hands-on components is just honestly what I love to do.”
Charlotte Dorn is a sophomore undergraduate student in the Department of Computer Science within the UNC College of Arts & Sciences.
Parasara Sridhar Duggirala is an assistant professor in the Department of Computer Science within the UNC College of Arts & Sciences.
Abel Karimi is a research assistant in the Department of Computer Science within the UNC College of Arts & Sciences.
As a Tar Heel, graduating senior Sweta Karlekar had the opportunities to intern at some of the country’s most influential companies, conduct research and present at international conferences. Now, she’s heading to Facebook to begin her career.
By Kim Spurr, College of Arts & Sciences, Monday, May 4th, 2020
To say that Sweta Karlekar’s four years at Carolina have been packed is an understatement. She had internships at Disney, Yelp and Facebook. She was a part of Google’s AI Research Mentorship Program. She presented her research at international and domestic conferences. She spent a semester in Silicon Valley.
Still, despite the amazing experiences, one of the things she said she’ll miss the most is a familiar routine.
“I’ll miss that feeling of walking out of Davis Library at 2 a.m. with your friends, feeling that sense of accomplishment after studying for hours — and being on a campus that echoes that back to you,” said Karlekar, a senior computer science major and entrepreneurship minor.
The Ashburn, Virginia, native came to UNC-Chapel Hill as a Chancellor’s Science Scholar. It became an integral part of her Carolina experience from day one.
“It’s the reason I came to UNC. Just meeting the scholars and the caliber of people in the program — they have been supportive of me the entire time,” she said. “Coming in as an out-of-state student, it made campus feel smaller.”
Karlekar got involved with undergraduate research early on; during her first year she applied for a position with the UNC School of Education working on developing science visualizations for children. During her sophomore year, she reached out to Mohit Bansal, assistant professor of computer science, and began diving into natural language processing and artificial intelligence research. She became very interested in the idea of applying artificial intelligence to improve people’s lives. She examined how to use AI to find linguistics characteristics in the speech of Alzheimer’s patients.
“Research gave me the chance to prove myself in ways I never would have imagined; it allowed for freedom, creativity and mobility,” she said. “It’s an amazing opportunity to tell other people your idea, communicate, work in teams and develop something that is truly yours.”
Through working in Bansal’s lab, she got to present research at a conference in New Orleans, and it was there that she met an employee at Disney and learned about the company’s efforts to combat cyberbullying. That conversation eventually led to an interesting twist in her academic journey.
Before beginning her junior year, she emailed the contact at Disney on a whim to ask about internship opportunities for the next summer. She was told they had a great fall internship opportunity, but, unfortunately, nothing for the summer.
So Karlekar made the difficult decision to drop out of Carolina for the fall semester and headed to Seattle to work for Disney. She then reapplied, came back to campus for about a week in the spring of her junior year, then left for the Silicon Valley semester through the Shuford Program in Entrepreneurship and Honors Carolina, completing an internship at Yelp. She stayed on in San Francisco through the summer and interned at Facebook, working on AI for recruiting.
She has spent her senior year on campus at Carolina, and she’ll go to work for Facebook as a software engineer after graduation.
Because she spent an entire year away from Carolina, Karlekar said she has been able to adjust to the abrupt transition to remote learning pretty well. She had already been focusing on making her senior year memorable. She reflected on the diversity of her four-year experiences, which outside of research included being involved in the Indian classical dance club, the makerspace community and mentoring refugees in Chapel Hill.
“Every moment of my senior year, I’ve felt more appreciative of Carolina. I took the time to reflect and made sure every time I spent with my friends was really special,” she said. “I’d love to see my friends again, but friendships don’t end just because we’re not in the same place right now.”
In fact, after everyone is done with finals, she and her friends are planning a Zoom party to celebrate.
Don’t give up on your dreams, she urges her fellow Carolina graduates.
“Carolina students have always been good at pushing through challenges and embracing new opportunities. Continue to do so.”
Han Guo is a senior double-majoring in computer science and statistics and analytics within the UNC College of Arts & Sciences. He researches how to build language technologies like automated image captioning.
Q: When you were a child, what was your response to this question: “What do you want to be when you grow up?”
A: I remember Thomas Edison making a significant impact on me when I was pretty young, so I thought I would become an inventor like him. Later, I came across books on Bill Gates, Paul Allen, and Mark Zuckerberg, and it became clearer to me that I wanted to do something related to technology. Actually, one night in primary school, after reading the book on Zuckerberg, I dreamt of myself coding — and that’s why I learned to code. The single most significant impact, though, came from Steve Jobs. His biography, which I bought right after he passed away, was life-changing. To this day, I have read or listened to the book about five times, and I reread it every year.
Q: Share the pivotal moment in your life that helped you choose your field of study.
A: When I was in high school, I was pretty interested in behavioral economics and statistics. This helped me develop a passion for using numbers to understand the fundamentally random nature of the world. In my first year of college, I worked on a project with my research advisor. I was tasked with using numeric representations of words to understand the evolution of conversation in the Wikipedia talk pages. At the same time, Cornell University professor Lillian Lee gave a talk at UNC on the use of machine learning and statistics to study sentiments and natural language. I was so hooked by the presentation that I fell in love with machine learning and language understanding.
Guo (first row, left) hangs out with members of the UNC Natural Language Processing Lab at a local brewery.
Q: Tell us about a time you encountered a tricky problem. How did you handle it and what did you learn from it?
A: One of the more recent technical problems I had was measuring the uncertainties in modern machine learning systems. My usual go-to approach is to read loads of papers and, if possible, talk to people who are experts. In this case, I ended up spending half of the winter break going through the PhD thesis of a professor at Oxford.
Q: Describe your research in 5 words.
A: Teaching machines to understand language.
Q: What are your passions outside of research?
A: Reading papers, books, and the news. Some of the books are directly related to my research, but I enjoy reading things on the economy and mathematics, as well as biographies. Books help me connect my knowledge to other disciplines.
Research UNCovered delves into the lives of UNC researchers from all disciplines and career levels, showcasing not only their research prowess but personal experiences in academia and beyond. Know someone we should feature? Nominate a researcher.
Associate Professor Jasleen Kaur received a 2019 Google Faculty Research Award from Google AI. Kaur’s proposal was one of only 150 to be funded out of 917 applications from 50 countries.
Google Faculty Research Awards provide unrestricted gifts to support world-class technical research in computer science, engineering, and related fields. The awards are structured as seed funding to support one graduate student for one year. The awards are very competitive; on average, around 15 percent of applications receive funding, and Kaur’s research proposal was one of only eight to be funded in the area of networking. The review process for 2019 involved 1,100 reviewers from across Google.
Many enterprises like Google, Facebook, and Amazon maintain large-scale, private networks to connect datacenter clusters. Those networks are equipped with instruments that monitor the health and performance of the network, but when anomalies and outages occur, manually diagnosing the problem can require much time and effort.
Kaur’s proposal will use unsupervised learning methods to analyze performance data collected on Google’s B4 network. Unsupervised learning is a branch of machine learning where a computer attempts to draw inferences from input data without any output data. In other words, rather than being given raw data and corresponding conclusions and determining how to classify the raw data, the computer will be asked to analyze only the raw data in search of patterns that will help in manual diagnosis.
The goal of Kaur’s project is to derive models that can help identify the causes of anomalies in monitored data. If successful, the research will enable more efficient and accurate diagnosis of the root causes of anomalies in wide-area enterprise networks like Google’s.
On March 18, Professor Gary Bishop of the Department of Computer Science logged in to check on Tar Heel Reader, an online library of beginner-level books that he has maintained since 2008. In the midst of social distancing to slow the spread of COVID-19 and with many of the site’s users out of school, Bishop expected to see a significant decline in usage.
Instead, Bishop found that overall usage had slightly increased. With students no longer accessing the site together in classrooms, peak usage has decreased. But, Bishop says, students are spreading out their visits throughout the day, using it later into the evening and even logging in on weekends. For Bishop, this is the latest instance in more than a decade of Tar Heel Reader exceeding his most ambitious hopes for the platform.
Tar Heel Reader is an online, open-source library of free, easy-to-read, accessible books. In addition to reading, users can log in and create their own books using public domain photos from Flickr, enabling teachers and caregivers to write books on topics that their students will find interesting. The site, developed by Bishop and Professor Karen Erickson of the Center for Literary and Disability Studies in the Department of Allied Health Sciences, grew out of a need for books for beginning readers with disabilities. It quickly grew popular with unexpected demographics, including students without disabilities, adults learning second languages, and even students of Latin. The site now includes books in 27 different languages, and users come from more than 200 countries and territories. Tar Heel Reader hit 10 million books read in January 2017, less than a decade after its launch.
Tar Heel Reader users are able to navigate the site and read books using a wide variety of input devices, including a single switch. Users with limited internet access can collect books for offline reading. The site includes instructional resources, and users have created their own tutorials on YouTube and other sites.
In a time of social distancing and stay-at-home orders, Tar Heel Reader exhibits a different type of virality, spreading among educators through positive reviews and conference testimonials. Erickson attributes the current surge in usage to Tar Heel Reader’s promotion by educators as a solution for students with disabilities who cannot otherwise engage in “virtual” learning. She also credits positive reviews for Tar Heel Shared Reader, a shared reading interface launched in 2019 that enables adults to help engage students more actively in reading and gain more meaning from texts.
Bishop says he has received several messages from Tar Heel Reader users citing COVID-19 distancing measures, including positive reviews for the site and requests to become book creators from countries hit hard by the pandemic. Users from the United States, Italy, Spain and the Netherlands have made up more than 80 percent of the site’s usage since March 16.
Pearl Hacks will soon host over 400 participants on campus for UNC’s only hackathon designed specifically for female and non-binary students.
Founded in 2014 by UNC alumna and current Google employee Maegan Clawges, Pearl Hacks is a beginner-friendly hackathon open to female and non-binary students 18 or older and will take place from Feb. 21 to 23.
“Pearl Hacks is sort of a cross between a competition and a conference,” Director of External Workshops and UNC junior Tylar Watson said.
The three-day event offers a number of experiences for students. During the 36-hour project-building hacking competition, participants also have the opportunity to attend workshops led by mentors, local community members and corporate sponsors like Twitter and Google.
“There’s like two things happening that you can participate in,” Watson said. “So if you are working on a project and you get stuck, you can say ‘OK, I’ll go do a workshop for the next hour.’”
Pearl Hacks hosts workshops and sponsor fairs that give participants the chance to meet with recruiters and professionals in the technology field, said Prasiddhi Jain, a junior biostatistics and computer science major and director of hacker experience for Pearl Hacks.
“We have a lot of mentors from companies like Google and Microsoft, so a lot of big name companies,” Jain said. “So you develop relationships, and you develop skills to talk to a lot of different kinds of people.”
While competition and networking are large components of Pearl Hacks, Jain said the goal of the weekend is to create a positive environment for those involved.
“It’s like a safe, inclusive space for people to just explore their interests in technology,” junior and Vice President of Logistics Hannah Cao said. “A lot of people associate hackathons with making a project and being really experienced. With Pearl Hacks, it’s just to create a space safe space for women and non-binaries, which is not very present in the computer science community.”
This year’s event will be the seventh annual hackathon, and Jain said the Pearl Hacks team has been working on making the weekend as inclusive as possible.
“We are trying to frame a lot of our workshops, our panel discussions, projects and events throughout Pearl Hacks toward being more beginner-friendly,” Jain said. “So we’re hoping that a lot of people that are attending — that haven’t had any coding experience or do identify themselves as beginner-friendly — really learn a lot from it and have a positive experience.”
In anticipation for the main event, Pearl Hacks is also hosting a number of “first look” beginner workshops from Feb. 10 to 13. These events are open to all genders.
As the weekend approaches, Cao said the Pearl Hacks team intends to strengthen and share their message that tech is for everyone.
“Tech is such a buzzword right now, and Pearl Hacks creates that safe, inclusive, welcoming space for people to just explore their interests,” Cao said. “There’s no commitment, there’s no pressure.”
Carolina was one of more than 900 Global Game Jam host locations last weekend, with nearly 90 students developing prototypes of video games over the span of two days.
By Rob Holliday, University Communications, Friday, February 7th, 2020
With USB cables and power strips spread through Sitterson Hall last weekend, dozens of Carolina students settled in for a long weekend in front of their computer screens for work that would never result in a letter grade.
Gathered in groups around tables, the students were on a mission to develop video games — and they only had 48 hours to do it. The students’ efforts were part of an international event called Global Game Jam.
Carolina was one of more than 900 Global Game Jam host locations across 119 countries. Nearly 90 people — most of them Carolina students — participated in Chapel Hill event, which was hosted by the UNC-Chapel Hill Game Development Club.
Because of time limitations, the groups only designed prototypes of their games, which are often turned into more fully fleshed-out versions later. The event encourages teams to collaborate, learn from each other and develop a new set of skills.
“The purpose of a ‘hackathon,’ a game jam or any sort of timed event like this is to learn. It’s to get better,” said Scott Amaranto, a Carolina sophomore and Global Game Jam participant. “That’s much more apparent and helpful with a team. You can share knowledge. You can push each other to be better.”
Four exceptional doctoral students, who are working in broadly-construed data science, including natural language processing (NLP), vision-and-language tasks, machine learning, and artificial intelligence, visited Bloomberg’s Global Headquarters in New York City in 2019 as part of the Bloomberg Data Science Ph.D. Fellowship. Bloomberg has benefited significantly from its first class of Fellows in 2018-2019, as well as its Data Science Research Grant Program, which builds relationships with academic researchers around the globe.
The goal of this fellowship is to engage professionals early in their careers and to provide support and encouragement for groundbreaking publications in both academic journals and conference proceedings. Today, we are pleased to announce the second class of Bloomberg Data Science Ph.D. Fellows.
A committee of Bloomberg’s data scientists from across the organization selected the Fellows based on their proposals’ technical resiliency and strengths, and recommendation letters from their academic advisors, some of whom accompanied the Fellows on their visit to Bloomberg. The committee’s decisions were based in part on the candidate’s diverse academic focus, with priority given to machine learning, NLP, information retrieval, knowledge graph, and quantitative finance; the quality of the ideas presented in the proposal; the candidate’s achievements and experience; and the idea’s potential business impact.
“Each project will advance the state of the art in their respective academic areas,” explained Songyun Duan, Head of Machine Learning Incubation in Bloomberg’s Office of the CTO. “The results will be published in top-tier conferences.”
During their Fellowship, each of the Fellows will work to advance their research and explore real-world applications that contribute to the innovative work leveraging data science and machine learning in Bloomberg’s products and services. The Fellows will also participate in an internship during the summer of 2020, during which they will collaborate with a Bloomberg team under the guidance of their research advisor to implement their research into one of the company’s applications to solve real-world problems and refine workflows.
As an introduction to Bloomberg, the Fellows traveled to New York City for three days to meet with their mentors and the company’s data science and AI engineering teams. During their visit, they learned more about the variety and depth of the company’s data science research and how the organization operates.
During her fellowship, Keith will further her research focused on improving natural language processing methods for computational social science applications. “I’m interested in improving social measurements of text, such as event extraction and extracting semantic and pragmatic signals from text, and improving methods that use these measurements in descriptive and causal inferences,” Keith said.
Contagion risk exists within the financial system when market participants utilize contracts to interact with each other, and the complex financial products that bind these networks create new ‘spiral’ risks that can potentially be measured and controlled with algorithms. Applying financial network analysis to real-world problems has been a challenge though, and Ariah Klages-Mundt, a Ph.D. candidate in applied math at Cornell University’s Center for Applied Mathematics, hopes to overcome hurdles by developing numerical methods and machine learning tools for sensitivity analysis and probabilistic measure of network risks. “I look forward to working with Bloomberg to see first-hand how these tools could realistically be used within the finance industry,” said Klages-Mundt.
Hao Tan, a fourth-year Ph.D. student in the University of North Carolina at Chapel Hill’s NLP Research Group (which is led by Assistant Professor Mohit Bansal), has worked on building the mapping from words and phrases to visual concepts, like objects and relationships, by designing tasks, building neural models, and developing training methods to learn this connection. At Bloomberg, where billions of data points are processed daily, Tan plans to explore the possibility of including visual information like images, videos and data plots by adapting methods developed for natural images to structural figures.
Machine learning algorithms can be used to solve sequential decision-making problems, as most interactive systems – like search engines – are improved through a recurrent loop with various stages involving learning from new data, improving the features and the model, and then testing the new system.
“My goal is to fundamentally speed up this process through new counterfactual inference techniques that move both learning and evaluation from ‘online’ to ‘offline’,” said
The 2018-2019 Ph.D. Fellows have all completed their internships and have had their fellowships renewed. Bloomberg has already benefited from the academic collaboration with strong Ph.D. candidates with interests similar to the company’s Data Science and AI Engineering teams. This included facilitating publications in top-tier conferences and improving the quality of the company’s products, where applicable, said Duan.
Out of 79 applications from doctorate students at universities in the United States, European Union and United Kingdom, a committee of Bloomberg’s data scientists from across the organization selected these four Fellows for the 2019-2020 academic year:
Yi Su (Cornell University) Off-Policy Evaluation and Learning for Interactive Systems Search engines, recommender systems, and most other user interactive systems go through a recurrent loop of improvement. This loop typically involves learning from newly collected data, making improvements to the features and the model, and then testing the new system in an online A/B test. My goal is to fundamentally speed up this process through new counterfactual inference techniques that move both learning and evaluation from “online” to “offline.”
Katherine Keith (University of Massachusetts Amherst) Constructing Subjective Knowledge Bases The field of information extraction (IE) has made great strides in constructing knowledge bases by extracting facts from large collections of unstructured text. IE methods have been used in many applied settings, including my recent work building a database of police fatality victims. However, extracting facts implies discerning between different realities in order to determine what is “true”. Social scientists, journalists, policy makers, and financial investors may be more interested in understanding a populations’ shared and conflicting subjective beliefs and how they vary temporally and spatially. This leads to the following research questions:
Can we extract propositions representing authors’ stated beliefs in order to construct subjective knowledge bases?
Once we have extracted subjective propositions for individual authors, can we infer belief communities by clustering authors with similar beliefs?
Hao Tan (UNC Chapel Hill) Summarizing Salient Content in Structured Documents with Figures Describing content in structured images, i.e., plots, charts, and diagrams, is crucial when summarizing or searching over, for example, complex financial or legal news documents. It helps a layperson to understand the salient information inside these complex figures and also enables visually-impaired people to “see” the figure. It also enhances pure-text paragraph summarization systems by providing additional valuable information from figures inside the news/legal document and allows search/retrieval over documents containing such structured figures. Hence, we propose models that learn to generate informative and comprehensive summaries for such structured figures in complex documents, that capture salient and logically entailed (verified) information.
Ariah Klages-Mundt (Cornell University) Learning Cascade Risks in Complex Economic Networks: Methods to Make Financial Network Analysis Practical for Application Computational and sensitivity problems currently present a barrier to using network models to quantify risks in networks of interacting firms. For instance, sensitivity from parameter uncertainty is not currently well understood and network computations can be algorithmically hard. I am developing numerical methods and machine learning tools to address these problems. This work will help make financial network analysis practical for application in industry.