Our Work
We propose an end-to-end model which generates captions for images embedded in news articles. News images present two key challenges: they rely on real-world knowledge, especially about named entities, and they typically have linguistically rich captions that include uncommon words. We address both.
Read MoreJenny Davis was elected as incoming Chair-Elect for the Communication, Information, Technologies & Media Sociology section for the American Sociological Association
Read MoreThis day-long course brought together research fellows from HMI and the Gradient Institute to exchange knowledge. Claire Benn presented the basics of ethics and Dan Steinberg and Lachlan McCalman from the Gradient Institute introduced the core elements of machine learning.
Read MoreSeth Lazar and Atoosa Kasirzadeh are invited to be ethical reviewers for NeurIPS, as the world's leading machine learning conference explores new ways to build ethical thinking into AI research.
Read MoreThis study represents the first systematic, pre-registered attempt to establish whether and to what extent the YouTube recommender system tends to promote radical content. Our results are consistent with the radicalization hypothesis. We discuss our findings, as well as directions for future research and recommendations for users, industry, and policy-makers.
Read MoreMind Design III will update Haugeland's classic reader on philosophy of artificial intelligence for the modern era. It will contain a mix of classic and contemporary readings, along with a new introduction to contextualise the topic for students. Expected publication date Q1 2021.
Read MoreSeth Lazar and Will Bateman have been invited to be Gradient Institute Fellows, cementing an ongoing partnership with Australia's leading group of ethical machine learning researchers and practitioners.
Read MoreEdTech, data privacy, children, and children's rights.
Read MoreThis talk was given at a conference on Holly Smith’s book, Making Morality Work, held at Rutgers on October 18, 2019. I argued that Making Morality Work poses the problem that moral theories must be 'usable', but then offers a solution that only partly solves it. I offered a way to extend the solution, but argued that even that only partly solves the problem, and that we can’t stop there.
Read MoreIn this talk, I argue for the practical problems of a counterfactual theory of mathematical explanations in sciences.
Read MoreIn this talk, I discuss for the bridging role of mathematics in empirical sciences as a reliable connecting scheme in our explanatory reasoning from lower-level to higher-level phenomena. I support this discussion by analyzing two explanations in biology and physics.
Read MoreHumans and machines regularly interact as part of daily life in contemporary societies. It is critical to understand the nature of these relationships. This presentation addresses role-taking in human-AI teams. Role-taking is a process of putting the self in the shoes of another, understanding the world from the other's perspective. We use an experimental design to determine how actively humans role-take with AI as compared with role-taking activation when encountering other humans.
Read More
