New cohort of SRI graduate fellows expand research to digital labour, blockchain, morality, international security, and more
The Schwartz Reisman Institute welcomes 15 new graduate fellows for 2021-22. From left to right, top to bottom: Rawan Abulibdeh, Asmita Bhutani, Davide Gentile, Jessica Hall, Vinyas Harish, Lillio Mok, Morgan MacInnes, Reid McIlroy-Young, Shashank Motepalli, Victoria Oldemburgo de Mello, Lief Pagalan, Julian Posada, Aida Ramezani, Daniel J. Wilson, Mohammad Yaghini.
The Schwartz Reisman research community continues to grow with the appointment of 15 graduate fellows from across the University of Toronto.
Serving a one-year term, this year’s graduate fellows come from a wide variety of academic disciplines across the sciences, social sciences, and humanities and represent the best and brightest U of T graduate student researchers working on topics at the intersection of technology and society.
The 2021-22 graduate fellows will contribute to the intellectual pulse and diverse research community at SRI, fostering interdisciplinary connections and collaboration, participating in workshops, seminars, and conferences, and contributing their work to SRI’s goal to be a global leader in ensuring that artificial intelligence (AI) and other advanced technologies benefit all of humanity.
Our new graduate fellows join the Institute at a pivotal time, as we release our first strategic plan (PDF) to guide our work for the next four years.
Read on to learn more about our new graduate fellows and how they will shape the global impact of the Institute.
Graduate fellows in health and community development
SRI’s ongoing work on health data will be bolstered by the arrival of Graduate Fellow Vinyash Harish, who is pursuing a PhD in clinical epidemiology and healthcare research at U of T’s Institute of Health Policy, Management and Evaluation.
Harish’s work includes applications of machine learning (ML) in healthcare and medicine, and his SRI project will investigate barriers and hindrances to “adaptive governance”—a form of governance that coordinates resources in uncertain, complex, and rapidly changing situations. Drawing from perspectives across medicine, public health, ethics, governance, and the technology industry, Harish will examine how we can best facilitate partnerships between health organizations and private technology companies in responding to COVID-19.
Harish is joined at SRI by his colleague in epidemiology Lief Pagalan from U of T’s Dalla Lana School of Public Health. Pagalan’s PhD research develops predictive models for public health that combine health, social, and environmental data and perspectives.
Pagalan’s SRI research project will leverage new data infrastructures and technologies to predict and prevent health risks, develop new solutions towards planning and managing the health of whole populations, and delivering public health services in ways that are more precise, efficient, and equitable. He’s specifically working on models that predict and help prevent premature deaths—i.e., deaths before 75 years of age—in Canadian cities.
“I use a health equity paradigm in my work,” says Pagalan, “so I also evaluate how social biases affect the interplay of social-technological processes. The goal is to develop health decision-making systems that are safer, fairer, and lead to healthier and more equitable communities.”
SRI Graduate Fellow Asmita Bhutani is a PhD student at the Ontario Institute for Studies in Education (OISE), specializing in anti-racist feminist political economy and workplace studies. Examining data-driven AI chains as socio-technical systems, her work focuses on the gendered aspects of labour processes on “microwork” platforms—services that facilitate a series of small units of work as part of a larger project, often performed by many people over the internet (e.g., Amazon’s Mechanical Turk).
Bhutani believes we need to examine data-based service work in informal economies and the experiences of workers in those systems in order to better comprehend the production of AI systems. Her goal is to make connections between meaningful AI, equitable work, and gender justice.
Graduate fellows in engineering
SRI’s incoming graduate fellows in engineering span a wide spectrum of specializations, from medical data to blockchain technology to AI justification and beyond.
“Blockchain is technology for human coordination and I want to make it happen,” says Shashank Motepalli, who is pursuing his PhD in the Edward S. Rogers Sr. Department of Electrical & Computer Engineering (ECE) at U of T.
Motepalli’s research on blockchain—the technology that runs cryptocurrencies like Bitcoin—aims to study so-called decentralized autonomous organizations (DAO), which Motepalli believes can help us reimagine the future of human coordination and governance structures. Because blockchain technology allows participants to transact without any trusted third-party platforms, it relies on consensus instead. Motepalli’s research focuses on how to scale this kind of consensus in blockchain networks with the goals of building scalable, decentralized, and eco-friendly consensus protocols.
“As an SRI graduate fellow, I’m looking forward to delving into the social construction of technology and understanding the social, ethical, and regulatory consequences of disrupting existing governance models,” says Motepalli.
Like Motepalli, incoming SRI Graduate Fellows Mohammad Yaghini and Rawan Abulibdeh are also pursuing their PhDs at ECE. Yaghini studies ML model governance, including how we might protect the intellectual property of models, perform non-intrusive audits on them, or apply ownership and attribution rights to their deployment. Unlike the technical aspects of ML models, such as striving for accuracy, Yaghini’s research stretches further beyond to encompass crucial human values like privacy, equitability, and accountability.
Abulibdeh works on natural language processing (NLP), a type of AI that enables computers to better generate and respond to human language by taking in large amounts of data. Her research aims to use NLP techniques on a large repository of electronic medical records in Ontario to improve phenotyping, which is the analysis of expressed traits in an organism—e.g., how scientists can determine the physical appearance of a person solely based on their DNA. Abulibdeh’s project has significant implications for improving research in both mental and physical health.
Rounding out SRI’s graduate fellows from the field of engineering is Davide Gentile from U of T’s Department of Mechanical & Industrial Engineering. Gentile’s research focuses on human interactions with AI, especially in “safety-critical domains”—situations in which the malfunction or failure of systems can result in serious consequences like severe harm to people, equipment, or the environment. An example of a safety-critical domain is the control room of a nuclear power plant.
Gentile’s goal is to create explanation and justification techniques that allow non-expert users to understand, trust, and effectively manage AI systems in safety-critical domains. As a mixed-methods researcher, Gentile says “the most stimulating part of joining SRI will be drawing on different areas of expertise to solve specific problems.”
Graduate fellows in the social sciences and humanities
We know that machine learning requires large amounts of data, but the labour required for producing this data is sometimes overlooked. At U of T’s Faculty of Information, Julian Posada’s PhD research examines the labour implications of outsourced data production for ML. His background in sociology, with specializations in digital labour and platform studies, will inform his research at SRI on Latin American crowdsourced workers who annotate data for ML and verify its algorithmic outputs.
“Data production and human labour are interrelated,” says Posada. “Without dignified working conditions, datasets—and algorithms trained with them—cannot be ethical and reliable.”
Posada is joined by two SRI graduate fellows from U of T’s Department of Psychology: Victória Oldemburgo de Mello and Daniel J. Wilson, both pursuing PhDs.
Wilson works in “decision neuroscience”—the study of what the architecture of the brain can tell us about human thought, feeling, and behaviour—with a particular focus on the “intention-behaviour gap,” which is the difference between what we intend to do and what we actually do. Using computational modeling, Wilson aims to measure, predict, and understand this gap. His work will contribute to the SRI’s ongoing engagement with the alignment problem.
Oldemburgo de Mello’s area of expertise is social psychology, and she uses computational methods to study social cognition and the psychological effects of different forms of social media use.
“I investigate how social media use is linked to changes in well-being and political polarization,” says Oldemburgo de Mello. “We know that billions of people worldwide use social media to share their opinions, status, and emotions. There are claims that social media harms our well-being and makes us more polarized, but the evidence is ambiguous at best. The questions in my research are directly linked to the SRI’s goal of deepening our knowledge of how technologies transform human lives.”
Jessica Hall’s PhD research at U of T’s Institute for the History & Philosophy of Science & Technology is in the areas of philosophy of science and philosophy of mind, with a particular focus on AI and the foundations of computing. As an SRI graduate fellow, Hall will explore what we mean by the terms “computation” and “AI.”
“Given that the range of technologies we call AI are many and varied, I am interested in what currently qualifies as AI,” says Hall. “And even though AI is usually framed in terms of computation, there is surprisingly little conceptual agreement on what it is for something to be ‘computational.’ My work aims to gain a deeper understanding of these foundational issues.”
Hall’s work will contribute to an understanding of the ways in which the characteristics of AI affect the social, ethical, and pragmatic issues associated with it.
Rounding out the graduate fellows from the social sciences and humanities is Morgan MacInnes from U of T’s Department of Political Science. In his PhD, MacInnes works on international relations, geopolitics, and the relationship between technology and global security. As a political science scholar, he says he is “particularly interested in the opportunities at SRI to work alongside and learn from those with technical expertise in machine learning.”
MacInnes’s SRI project will focus on the use of AI in military technology, its implications for a new kind of arms race, and its effects on the international security environment. He’ll also examine what kinds of governance mechanisms might be employed to mitigate the adverse effects of AI in international relations.
Graduate fellows in computer science
SRI’s three graduate fellows in U of T’s Department of Computer Science are Lillio Mok, Reid McIlroy-Young, and Aida Ramezani, all currently pursuing PhDs.
Mok’s ongoing work is in computational social science, an area of research in which computers are used to simulate, model, and analyze social phenomena such as human behaviour and communication. Using a mix of behavioural data science and human-computer interaction (HCI) methods, Mok’s research looks at human behaviours, preferences, and their (mis)alignment online. He’s particularly interested in online news consumption and understanding whether news is shared differently between and within communities with partisan affiliations.
McIlroy-Young also works in computational social science, specializing in machine learning, reinforcement learning, behavioural modelling, and social networks. As an SRI graduate fellow, he will look at “sponsorship reads” on YouTube—when a person in a video vocally promotes a product or service, rather than a standalone ad interrupting the video itself. The aim of this research is to see how advertisers’ content overlaps with the video in which it appears, what the norms are for sponsorship, and what this suggests about the kind of information that is collected about a video’s audience.
Ramezani’s research in computer science, cognitive science, and NLP aims to extract so-called “moral inference” from textual data. Moral inference refers to the logical relationships between moral claims—for example, if it’s morally wrong to do one thing, does that extend to another related thing? Ramezani’s work has, for example, focused on public moral judgments about particular social issues. Her research at SRI will explore the question of whether computational methods can be used to detect moral stances from the speech that humans produce.
The Schwartz Reisman Institute extends a warm welcome to our new graduate fellows, who will no doubt leave a mark on our research community by connecting with fellow scholars from other academic disciplines on common questions, contributing to ongoing dialogues at the intersection of technology and society, and helping SRI achieve its goal to become the world’s leading institute ensuring that AI and other advanced technologies benefit all of humanity.
