Newswise — Developing how the heart forms and brain works. How to analyze sarcasm computationally. Harnessing computers to develop campaign rhetoric that delivers voters across the political spectrum. The Halıcıoğlu Data Science Institute (HDSI) at the University of California San Diego has announced new funding for these and other innovative research projects helping develop the next generation in data science leadership.
HDSI has awarded research undergraduate scholarships to 18 students funding 15 projects total this spring quarter. In the inaugural year of the Undergraduate Scholarship Program, each project is granted a total $2,500-$6,000. The scholarship program goal is to provide hands-on training in real-world projects to emerging data science talent, noted Bradley Voytek, an HDSI Fellow and associate professor in the Department of Cognitive Science at UC San Diego.
More than 50 students competed for the 15 scholarship project awards for the spring academic term. “It’s impressive what a hunger our students have developed to use the newest data science tools to uncover answers to important questions,” said Voytek, who directs the Institute’s scholarship program.
This is the first full academic year for the Halıcıoğlu Data Science Institute (pronounced Haw-li-jyo-loo), which oversees one of the largest data science majors in the nation. The cross-disciplinary program educates more than 600 undergraduate students, and has an expanding graduate program. The newest funded undergraduate research projects drew students from more than 15 different academic disciplines – from mathematics to music, and data science to cognitive science. That diversity highlights the breadth of need and appeal of data science applications in society, noted Voytek.
Winning projects include data-science-based research exploring such subjects as: Quantum mechanics, how the brain works, how the heart forms, and even enhancing leisure activities such as fantasy sports games. The scholarships have proven to be a successful step in fulfilling the mission of the unique institute to serve as a data science hub drawing across academic boundaries, notes HDSI Director Rajesh Gupta.
“Our scholarships are a good way to develop research strengths in the undergraduate cohort, and for faculty and industry partners to work directly with our students on dynamic projects,” said Gupta, who is also a UC San Diego Distinguished Professor and the Qualcomm Endowed Chair in the Department of Computer Science and Engineering at the Jacobs School of Engineering.
Among the award-winning projects:
The Importance of Political Rhetoric, being researched by a team of Data Science majors: Sravya Voleti, Austin Le and Zahra Masood. The team plans to analyze how politicians can boost their approval ratings by optimizing language to appeal to voters outside their target demographics. The Data Science students to use statistical modeling and make predictions about how a politician’s word choice could affect campaign outcomes. “We hope to discover what kind of language makes politicians successful,” their proposal states, “to create a definitive model that will help future politicians optimize their success in elections.”
Sickly Greens: Visualizing Environmental Effects on Plants, conducted by Esther Wang, an undergraduate majoring in Interdisciplinary Computing and the Arts, a joint program from the departments of Visual Arts and Music. Wang plans on creating real-time 3-dimensional visualizations depicting the devastation of pollution on plant life on a global scale. Her goal: “To bring attention to the wide range of detrimental impacts that human activity has on the earth.” Her approach will be to collect, extract and analyze data on climate change impacts and transform the data into interactive 3-D imagery to illustrate current and future impacts on plant life. “The project aims to bring the investigation and further understanding of environmental data beyond scientists and academia to everyone through accessible and visually intriguing art,” Wang proposed. “By amplifying and playing on the visual… I hope to bring scientific understanding to more people.”
Deep Learning for Sarcasm Detection. Data Science major Takashi Yabuta will use his scholarship to develop algorithms that help machines better understand nuances of human language. Yabuta notes that the challenge of Natural Language Processing (NLP), which is basically talking to voice recognition computer applications like Siri and Alexa, has been in the classification of sarcasm. Sarcasm tends to confound computers because it’s negative speech that appears positive. He hopes to create new approaches to improve algorithm performance for NLP.
The HDSI Undergraduate Research Scholarship program is geared to producing results, with scholars submitting final reports highlighting accomplishments. Each awardee works with a mentor to develop analytical skills, data science portfolios, and foster novel data-driven approaches to problem solving. Mentors include: Virginia de Sa, professor of Cognitive Science and HDSI Associate Director; Pavel Pevzner, an HDSI faculty affiliate in Computer Science and Engineering; Federico Rossano, and Eran Mukamel, of the Department of Cognitive Science; Erin Rose Glass, Digital Scholarship Librarian, Geisel Library; Robert Twomey, lecturer in Data Science and postdoctoral scholar with the Arthur C. Clarke Center for Human Imagination; Julian McAuley, assistant professor of Computer Science and Engineering; Zhuowen Tu of the departments of Cognitive Science and Computer Science and Engineering; Sean Kross, Ph.D. student in Cognitive Science; Andreas Goetz, assistant research scientist, San Diego Supercomputer Center; Mohit Jain, Department of Medicine; and Matthew Daugherty of the Division of Biological Sciences.
The new scholarship winners and projects are:
- ● Mohamed Al Elew: Public Information in a Data-driven Society
- ● Ethan Armand: Interspecies Single-cell Neuronal Analysis through Multi-omic Data
- ● Ethan Bull-Vulpe: Database and Workflow Infrastructure for Machine Learning of Molecular Properties from Quantum Mechanical Data
- ● Yixin Chen: PreStage: Previews and Predictions for Landscape Photographers
- ● Fiona Cisternas: Using Ultra Wideband Technology to Improve Field and Experimental Tracking Methods
- ● Zheng Hao Tang: Multi-agent Optimization in Fantasy Soccer
- ● Chieko Imai: Devising Methods for Machine Learning Algorithm Transparency and Interpretability
- ● Chinmay Kalluraya: A Novel Method to Infer Horizontal Gene Transfer
- ● Gayatri Mainkar: Identifying Lipid Mediators of Pluripotency Maintenance and Cardiomyocyte Differentiation
- ● Bohan Ni: Estimation of Microbiome Complexity through Metagenomic Sequencing and Hybrid Statistical Approach
- ● Kevin Tan: Long-term Video Prediction with Self-supervised Spatiotemporal Learning
- ● Sravya Voleti, Austin Le and Zahra Masood: The Importance of Political Rhetoric
- ● Esther Wang: Sickly Greens: Visualizing Environmental Effects on Plants
- ● Takashi Yabuta: Deep Learning for Sarcasm Detection
- ● Yuxin Zou and Jiawei Tang: Exploration in Recurrent and Transformer Neural Networks for Text Summarization
HDSI plans to continue awarding the undergraduate scholarships annually, said Voytek. HDSI is also seeking mentors from across the campus community to lend their expertise and experience to guide student research. All UC San Diego undergraduate students are eligible to compete for scholarship funding. These awards follow the first quarter awards to 34 students who are currently mid-way through their year-long research projects. Learn more about the HDSI Undergraduate Research program.