“Serving client-side computation on a static server: WebAssembly“

Abstract: WebAssembly is a new tool that offers the opportunity to deliver efficient, high-performance programs to diverse hardware (like your laptop) consistently. This results in a program that is delivered via a web page and then runs on a local machine. As in, whoever writes the code can use a static server to deliver dynamic results. WebAssembly is an instruction format for computers (i.e., the thing that translates a computer programming into something that is machine readable) that uses resources provided by the browser to execute code. There are a number of reasons we might want to do this including economics (you don’t have to pay for a server), philosophy (all programs should always be accessible), and user experience (cross-compatibility). This comes with problems, though, namely security, which the community has worked hard to counteract. One of the primary benefits of WebAssembly is that, as a binary format, it is able to glue together many different software languages, each of which has been optimized in different ways and has communities with different interests. This allows us to use highly performant libraries and packages, no matter the language in which they are written, using JavaScript as the glue. Such programs are also web native with little adjustments. WebAssembly is a highly flexible tool that all developers and data scientists should know to efficiently and effectively deliver to a variety of stakeholders.

Bio: Curtis Atkisson joined the eScience Institute in Feb 2024. He received his PhD in Evolutionary Anthropology from UC Davis with a designated emphasis in Computational Social Science. His dissertation was on how changes in people’s complex social networks impact their cooperative behavior. This work involved ethnography, participant observation, and surveys done amongst the Makushi (and some Wapashana) in southern Guyana. The analysis of those data required complex networks, designing new measures of information in those networks, and modeling those changes with custom-built Bayesian statistical models. His postdoctoral work applied his methodological expertise to understanding Open Source Software communities and how they persist, as well as expanding his tools to include Text As Data/Text Mining, machine learning, and AI approaches to understanding text (e.g., GPT as a zero-shot translator). Substantively, Curtis is interested in why people cooperate with others, or, perhaps more accurately, why we do not always cooperate with others. This ranges from small-scale cooperation such as sharing food with a neighbor up to large-scale cooperation such as building computer software that we give away for free that makes other people rich. He is also interested in using open-source software as a model system to study many processes in the social sciences–due to the nature of open-source software, almost all of the data regarding interactions within and participation in the system are publicly available.

Jihyeon Bae is a PhD candidate in the Political Science Department at the University of Washington. She is interested in comparing design choices around international organizations and how they influence cooperation among states. She is also passionate about applying NLP models to explore how rhetoric changes in international forums like the United Nations General Assembly. During the UW DSSG program, she will work on a project assessing water reuse patterns, based on substantive knowledge in actor-based institutional design. Born and raised in South Korea, she received her B.A. in International Studies from Kyung Hee University in 2019, with additional training from the Applied Mathematics Department.

Daniel Vogler currently pursuing a Master’s degree in Data Science at the University of Washington. Before graduate school he worked as a management consultant, primarily supporting clients in healthcare and retail. Daniel graduated with a B.A. from Princeton University in 2021. He is deeply interested in the intersection between data science, machine learning, and energy, especially in data-driven approaches to energy policy.

The 2024-2025 seminars will be held in person, and are free and open to the public.