Hello, I’m Nino!
I’m currently a second-year Ph.D. student in Computer Science at the University of Washington, where I am fortunate to be advised by Katharina Reinecke and René Just, and to collaborate closely with Spencer Wood. I truly admire each of them, both as researchers and as people, and I’m grateful for the opportunity to learn from them.
My research broadly explores the intersection of technology and society, with a particular focus on the role technology can play in addressing climate change. I am specifically interested in understanding how technology can contribute to climate change mitigation efforts, while also carefully considering the environmental impact of the technologies themselves.
Recent News
-
Selected to Serve on the Campus Sustainability Fund
November 2024
Chosen to serve on the Campus Sustainability Fund (CSF), a student-led initiative that allocates funding for projects aimed at enhancing sustainability and fostering cultural diversity and inclusion across the University of Washington campus.
-
Creating a Safer Space for Women in the Allen School
November 2024
Co-led efforts to create a more inclusive and supportive environment for women in the Allen School. Organized and facilitated a reading group and discussion sessions focused on how to be a supportive listener.
-
Research Presentation at the Annual Research Showcase + Open House
October 2024
Presented my research around forest management and HCI at the Annual Research Showcase + Open House hosted by the Paul G. Allen School of Computer Science & Engineering.
-
Science Policy Analyst at the Clean Energy Institute
September 2024
Selected as a Science Policy Analyst by the Clean Energy Institute to collaborate with the Washington State Academy of Sciences (WSAS), contributing to the development of clean energy policy in Washington state.
-
Co-Instructor for CSE390F: Tackling Climate Change with Technology
September 2024
Co-instructing the newly launched course, CSE390F: Tackling Climate Change with Technology, alongside Travis McCoy, to engage students in thinking about ways that technology might be deployed to address climate change.
About Me
I believe that the places where I’ve lived have distinctly shaped every part of who I am and am becoming. The map represents the journey of all the towns, cities and countries that have taken root within me.
I was originally born in Georgia, but also lived in Austria before moving to the U.S. I have a deep love for languages and culture — if I could have any superpower, it would be speaking all the languages of the world. It also instilled in me a tantalizing wanderlust and a yearn to constantly travel.
When I am not doing research, you can find me going on long walks or reading quietly with a hot mug of chamomile tea (even in the summer).
“To be whole is to be part; true voyage is return.”
Selected Previous Projects
Preventing Homelessness: Evidence-Based Methods to Screen Adults and Families at Risk of Homelessness in Los Angeles
-
The overall homeless population in Los Angeles County continues to grow as inflows into homelessness outpace exits to housing. The key to preventing homelessness is to ensure scarce prevention resources are going to people who will become homeless without those resources. In this study, we evaluate the surveys used to screen adults and families who self-identify as being at risk of homelessness. Specifically, we evaluate screening surveys called Prevention Targeting Tools (PTTs) currently used by homelessness prevention service providers in the City and County of Los Angeles. The PTTs are used to determine whether people are eligible for prevention services. Participants seeking prevention services must first meet two eligibility criteria: they must be at imminent risk of homelessness (i.e., will lose housing within 30 days) and have an income at or below 50% of the Area Median Income (AMI) for Los Angeles County. If they meet those criteria, they take the PTT, their answers are assigned points, and then a total score determines eligibility for services. There are separate versions of the PTTs for families, single adults, and transition-age youth (TAY). Those eligible for prevention typically receive short-term financial assistance (e.g., rental assistance, utility assistance) ranging on average between 1,000 to 5,000, legal assistance, and/or mediation with landlords or property managers.
Guided by the following research questions, we developed improved PTTs that can be used in a variety of different settings to determine eligibility for homelessness prevention programs among people who self-identify as being at risk.
Are there homelessness risk factors that are not currently captured on the PTTs that could be added to the PTTs to potentially improve their ability to predict future homelessness?
How can the wording and structure of the PTTs be improved to maximize the validity of responses?
What improvements can be made in the PTT administration process in order to more accurately capture information on at-risk individuals and families?
Can reweighting PTT questions and removing questions from the PTTs result in shorter, more accurate screening tools?
Parsimony and Machine Learning in Neuroimaging
-
The disparity between an individual’s brain age and their chronological age can be an indicator for various neurological disorders. A previous brain-age prediction study investigated the ability of multimodal brain imaging data to predict age, relying on anatomical and functional brain data to build a machine learning model with over 10,000 features. In our preregistered study, we used anatomical MRI data from the NIMH/NHLBI Data Sharing Project (NNDSP) dataset to compare accuracy in prediction of age for a complex machine learning model with a large number of features to a simple machine learning model with only four features: white matter fraction, grey matter fraction, CSF fraction and intracranial volume, chosen a priori. With samples from a large lifespan sample (N=441, age 5-77) as our training and test data we found that the predictive ability of the complex model was similar to the predictive ability of the simple model on out of sample data. We also tested the generalizability of each model to novel data from the Human Connectome Project (HCP) databank (N=895, age 22-37) and found that the complex model outperformed the simple model. Both the simple and complex model performed worse than chance in predicting age on the HCP data, which is likely attributable to the limited age range of the data and our stringent, preregistered definition of chance performance. In our exploratory analysis, we tested the generalizability of each model on the Nathan Kline Institute (NKI) dataset (N=907, age 6-85) and again found no significant difference between the predictive ability of the simple and complex models. The performance of the simple age prediction model illustrates some of the trade offs between parsimonious vs complex models for predicting brain age. Given the comparable performance of the approaches, the more rapid parsimonious approach using only a few features is generally advantageous.
Detecting and Harmonizing Scanner Differences in the ABCD Study-Annual Release 1.0
-
In order to obtain the sample sizes needed for robustly reproducible effects, it is often necessary to acquire data at multiple sites using different MRI scanners. This poses a challenge for investigators to account for the variance due to scanner, as balanced sampling is often not an option. Similarly, longitudinal studies must deal with known and unknown changes to scanner hardware and software over time. In this manuscript, we have explored scanner-related differences in the dataset recently released by the Adolescent Brain Cognitive Development (ABCD) project, a multi-site, longitudinal study of children age 9-10. We demonstrate that scanner manufacturer, model, as well as the individual scanner itself, are detectable in the resting and task-based fMRI results of the ABCD dataset. We further demonstrate that these differences can be harmonized using an empirical Bayes approach known as ComBat. We argue that accounting for scanner variance, including even minor differences in scanner hardware or software, is crucial for any analysis.