Fall 2020

The number of scientific products, including scientific software, has been steadily growing in the last years. This growth makes it difficult for researchers to understand all the latest code and publications available. A great body of research has attempted at classifying similar papers and literature. However, there aren’t to date good approaches for finding similar or related code. In this project, the students will analyze different unsupervised methods to find scientific software similarities based on a) An automated analysis of their dependencies; b) By classifying the main functionality of software components.

Cancer remains the second leading cause of death in the US. However, recent advancements have increased cancer survivorship, now numbering tens of millions. Given this, there is tremendous interest in studying cancer-related cognitive impairment (CRCI) and CRCI due to chemotherapy or “chemobrain”, can afflict up to 78% of survivors. The neural substrates of CRCI are unknown and understanding this may improve survivors’ quality of life. The CRCI neuroimaging literature is still in its infancy and these studies have used small sample sizes from traditional research-dedicated nCE scans. Because conducting well-powered neuroimaging studies is very expensive, adapting clinical CE T1w scans could prove useful for CRCI and many other diseases like dementia. The promary objective of this project is to develop a novel deep learning method to generate nCE images from acquired CE T1w scans to allow accurate brain morphometry and be a plentiful source of inexpensive neuroimaging data.

Top climate scientists post their findings and views regularly on social media. These very scientists are met with tweets from those with opposing views, often containing vitriolic and false information. It is important that we can identify and characterize these tweets to understand the counter-narratives of climate change. We will address topics including false information, bot campaigns, and harassment.

College football recruiting is big business. This project aims to determine if there are biases in who and how college football coaches recruit players. By creating a comprehensive data set of college recruits and integrating relevant data with current socioeconomic markers (i.e. census data) we hope to determine if there are patterns in who and where football coaches recruit their players, regardless of talent.

Politicians in modern democracies across the world have eagerly adopted social media for engaging their constituents, entering into direct dialogs with citizens. From the perspective of political actors, there is a need to continuously gather, monitor, analyze, and visualize politically relevant information from online social media with the goal to improve communication with citizens and voters. The goal of this proposal is to create a tool that enhances interaction and dialogue between political actors and their followers. This will be achieved by creating compact and comprehensive summaries that aggregate and visualize common narratives, thus, reducing the cognitive load required to read all the messages and streamlining the dialogue experience.

As the requirements for success in the workforce become increasingly technical, there is a commensurate need for curricula that can engage and capture the imagination of students, especially those from traditionally underrepresented groups in STEM. One way to reach these groups is via curricula that appeals to contexts in which they’re familiar and engaged, such as sports. To that end, this project will explore the development of a sports-based introductory data science curriculum with the goal of engaging students who might otherwise not be interested in pursuing data science as a career. Students will work on generation of illustrative code examples/problem sets in Python using sports examples.

Cyber ontologies such as STIX and ATT&CK can represent complex relationships between cyber threat actors, attacks and infrastructure. While such representations are conducive to interoperability between systems, they are often unwieldy for human cyber analysts to deal with directly. Conversely, Natural language generation (NLG) frameworks like FrameNet represent language in a structured manner, but frame specifications are often not specific enough for specialized domains (such as cyber security). Leveraging and combining the semantic structure of both forms can create a tool that can translate cyber threat data in standard interoperable formats (such as STIX) to human-readable reports, via existing NLG frameworks. Working on a project such as this provides an opportunity for significant impact, as the fusion of these two structures could greatly increase both the adoption and the utility of cyber threat ontologies.

The Kaggle.com competition ecosystem is a rich and active community with a designed Progression System that uses performance medals to rank and differentiate users into tiers. However, winning performance medals in Kaggle is more complex than it appears. Users are bound by the available competitions, characteristics of the competition’s problem statement, the quality of their software submissions, and the quality of other competitors (including collaborators). With these factors, one user’s earned “Gold” medal from one competition may have required more effort and a higher quality solution than another user’s earned “Gold” medal in a different competition. This project has great potential to learn about open competitions in data science. Some example questions are: What features help predict whether a user will win a medal in a competition? How can users be clustered and differentiated from one another using their competition patterns and medal-winning solutions? How quickly (in days) will a user win their next competition medal? What is the probability that a user will assemble a team for a competition? What are features that predict high-performing teams? What features help generate teammate recommendations?

Students will analyze images from optical microscopes that reveal features of materials and microstructures using machine learning techniques. These images have been collected by geologists, who use them to study the rock samples that they collect in the field and determine their properties and origins. We have a baseline system already implemented, and the goal is to improve it with new machine learning techniques, guided by the insights of our collaborating geologists.

We are increasingly hyper-aware of the effects of climate change. But we are not always aware of how hyper-local those effects can be. Across the 2,500 square miles of Los Angeles County, the impact of climate change is playing out in different ways. For example, some areas are experiencing more frequent spikes in extreme temperatures, while others are not. We propose a project that would fuse together several different datasets in order to map how temperature changes and other variables are hitting some corners of Los Angeles harder than others. Often, these areas are inhabited by people facing numerous other inequities, such as poor healthcare access. By examining several years’ worth of hourly average temperatures from thousands of spots across Los Angeles County, and combining that with other datasets, such as tree cover, cases of asthma, and so forth, it is possible to create an interactive map that illustrates where the impacts of climate change are most acute. This project would be published by Annenberg’s Crosstown publishing outlet and would be distributed widely. The project would have immediate practical applications and could inform policy decisions on issues such as where to place parks and green spaces.