Search for: All records

We call for a relational approach to constructing protest event data from news sources to provide tools for detecting and correcting errors and for capturing the relations among events and between events and the texts describing them. We address two problems with most protest event datasets: (1) inconsistencies and errors in identifying events and (2) disconnect between data structures and what is known about how protests and media accounts of protests are produced. Relational data structures can capture the theoretically important structuring of events into campaigns and episodes and media attention cascades and cycles. Relational data structures support richer theorizing about the interplay of protests and their representations in news media discourses. We present preliminary illustrative data about Black protests from these new procedures to demonstrate the value of this approach. 

Information access research (and development) sometimes makes use of gender, whether to report on the demographics of participants in a user study, as inputs to personalized results or recommendations, or to make systems gender-fair, amongst other purposes. This work makes a variety of assumptions about gender, however, that are not necessarily aligned with current understandings of what gender is, how it should be encoded, and how a gender variable should be ethically used. In this work, we present a systematic review of papers on information retrieval and recommender systems that mention gender in order to document how gender is currently being used in this field. We find that most papers mentioning gender do not use an explicit gender variable, but most of those that do either focus on contextualizing results of model performance, personalizing a system based on assumptions of user gender, or auditing a model’s behavior for fairness or other privacy-related issues. Moreover, most of the papers we review rely on a binary notion of gender, even if they acknowledge that gender cannot be split into two categories. We connect these findings with scholarship on gender theory and recent work on gender in human-computer interaction and natural language processing. We conclude by making recommendations for ethical and well-grounded use of gender in building and researching information access systems. 

Using novel data, we provide the first panoramic view of U.S. Black movement protest events as reported in U.S. newswires between 1994 and 2010 and put our quantitative data into dialogue with qualitative accounts. Struggles during these years presaged the Black Lives protest waves of 2014 to 2016 and 2020. Protests increased after the 1995 Million Man March into 2001 but dropped abruptly after the 9/11 attacks. Collective action increased again at the end of the 2000s. Protests in response to police violence and other criminal-legal issues were major arenas of struggle and news coverage. Also common were issues of national identity including celebrations of Black history and Black solidarity, protests about Confederate symbols, and protests about White hate groups and hate crimes. Although Black people protested about a wide variety of issues, newswires focused disproportionately on incidents of police violence and perceived threats of Black violence. There is substantial continuity in issues, organizations, and activism between this earlier period and the Black Lives Movement of 2014 to 2020. 

Benchmark datasets play a central role in the organization of machine learning research. They coordinate researchers around shared research problems and serve as a measure of progress towards shared goals. Despite the foundational role of benchmarking practices in this field, relatively little attention has been paid to the dynamics of benchmark dataset use and reuse, within or across machine learning subcommunities. In this paper, we dig into these dynamics. We study how dataset usage patterns differ across machine learning subcommunities and across time from 2015-2020. We find increasing concentration on fewer and fewer datasets within task communities, significant adoption of datasets from other tasks, and concentration across the field on datasets that have been introduced by researchers situated within a small number of elite institutions. Our results have implications for scientific evaluation, AI ethics, and equity/access within the field.