Computer Science

Digital platforms, including social networks, are major sources of economic information. Evidence suggests that digital platforms display different socioeconomic opportunities to demographic groups. Our work addresses this issue by presenting a methodology and software to audit digital platforms for bias and discrimination. To demonstrate, an audit of the Facebook platform and advertising network was conducted. Between October 2019 and May 2020, we collected 141,063 ads from the Facebook Ad Library API. Using machine learning classifiers, each ad was automatically labeled by the primary marketing category (housing, employment, credit, political, other). For each of the categories, we analyzed the distribution of the ad content by age group and gender. From the audit findings, we considered and present the limitations, needs, infrastructure and policies that would enable researchers to conduct more systematic audits in the future and advocate for why this work must be done. We also discuss how biased distributions impact what socioeconomic opportunities people have, especially when on digital platforms some demographic groups are disproportionately excluded from the population(s) that receive(s) content regulated by law.

Presidential primaries are a critical part of the United States Presidential electoral process, since they are used to select the candidates in the Presidential election. While methods differ by state and party, many primaries involve proportional delegate allocation using the so-called Hamilton method. In this paper we show how to conduct risk-limiting audits for delegate allocation elections using variants of the Hamilton method where the viability of candidates is determined either by a plurality vote or using instant runoff voting. Experiments on real-world elections show that we can audit primary elections to high confidence (small risk limits) usually at low cost.

Media outlets play a key role in spreading scientific knowledge to the general public and raising the profile of researchers among their peers. Yet, given time and space constraints, not all scholars can receive equal media attention, and journalists' choices of whom to mention are poorly understood. In this study, we use a comprehensive dataset of 232,524 news stories from 288 U.S.-based outlets covering 100,208 research papers across all sciences to investigate the rates at which scientists of different ethnicities are mentioned by name. We find strong evidence of ethnic biases in author mentions, even after controlling for a wide range of possible confounds. Specifically, authors with non-British-origin names are significantly less likely to be mentioned or quoted than comparable British-origin named authors, even within the stories of a particular news outlet covering a particular scientific venue on a particular research topic. Instead, minority scholars are more likely to have their names substituted with their role at their institution. This ethnic bias is consistent across all types of media outlets, with even larger disparities in General-Interest outlets that tend to publish longer stories and have dedicated editorial teams for accurately reporting science. Our findings reveal that the perceived ethnicity can substantially shape scientists' media attention, and, by our estimation, this bias has affected thousands of scholars unfairly.

Advances in Artificial Intelligence (AI) and Machine Learning (ML) that are being applied to legal efforts have raised controversial questions about the existent restrictions imposed on the practice-of-law. Generally, the legal field has sought to define Authorized Practices of Law (APL) versus Unauthorized Practices of Law (UPL), though the boundaries are at times amorphous and some contend capricious and self-serving, rather than being devised holistically for the benefit of society all told. A missing ingredient in these arguments is the realization that impending legal profession disruptions due to AI can be more robustly discerned by examining the matter through the lens of a framework utilizing the autonomous levels of AI Legal Reasoning (AILR). This paper explores a newly derived instrumental grid depicting the key characteristics underlying APL and UPL as they apply to the AILR autonomous levels and offers key insights for the furtherance of these crucial practice-of-law debates.

In this paper, we describe the foundations for generating a chatbot out of a website equipped with simple, bot-specific HTML annotations. The approach is part of what we call conversational web browsing, i.e., a dialog-based, natural language interaction with websites. The goal is to enable users to use content and functionality accessible through rendered UIs by "talking to websites" instead of by operating the graphical UI using keyboard and mouse. The chatbot mediates between the user and the website, operates its graphical UI on behalf of the user, and informs the user about the state of interaction. We describe the conceptual vocabulary and annotation format, the supporting conversational middleware and techniques, and the implementation of a demo able to deliver conversational web browsing experiences through Amazon Alexa.

Newspapers are trustworthy media where people get the most reliable and credible information compared with other sources. On the other hand, social media often spread rumors and misleading news to get more traffic and attention. Careful characterization, evaluation, and interpretation of newspaper data can provide insight into intrigue and passionate social issues to monitor any big social incidence. This study analyzed a large set of spatio-temporal Bangladeshi newspaper data related to the COVID-19 pandemic. The methodology included volume analysis, topic analysis, automated classification, and sentiment analysis of news articles to get insight into the COVID-19 pandemic in different sectors and regions in Bangladesh over a period of time. This analysis will help the government and other organizations to figure out the challenges that have arisen in society due to this pandemic, what steps should be taken immediately and in the post-pandemic period, how the government and its allies can come together to address the crisis in the future, keeping these problems in mind.

When students write programs, their program structure provides insight into their learning process. However, analyzing program structure by hand is time-consuming, and teachers need better tools for computer-assisted exploration of student solutions. As a first step towards an education-oriented program analysis toolkit, we show how supervised machine learning methods can automatically classify student programs into a predetermined set of high-level structures. We evaluate two models on classifying student solutions to the Rainfall problem: a nearest-neighbors classifier using syntax tree edit distance and a recurrent neural network. We demonstrate that these models can achieve 91% classification accuracy when trained on 108 programs. We further explore the generality, trade-offs, and failure cases of each model.

Autonomous agents acting in the real-world often operate based on models that ignore certain aspects of the environment. The incompleteness of any given model---handcrafted or machine acquired---is inevitable due to practical limitations of any modeling technique for complex real-world settings. Due to the limited fidelity of its model, an agent's actions may have unexpected, undesirable consequences during execution. Learning to recognize and avoid such negative side effects of the agent's actions is critical to improving the safety and reliability of autonomous systems. This emerging research topic is attracting increased attention due to the increased deployment of AI systems and their broad societal impacts. This article provides a comprehensive overview of different forms of negative side effects and the recent research efforts to address them. We identify key characteristics of negative side effects, highlight the challenges in avoiding negative side effects, and discuss recently developed approaches, contrasting their benefits and limitations. We conclude with a discussion of open questions and suggestions for future research directions.

Countries across the world have instituted unprecedented restrictions on freedom of movement, privacy and individual rights to control the spread of COVID-19. These measures tend to have been derived from communally orientated East Asian cultures. The way that culturally relevant concepts of rights and freedoms underpin COVID restrictions in democratic and individually orientated countries remains unknown. This data memo addresses this issue through analysis of pro- and anti-restriction discourse on social media in the US. It finds that anti-social and economic restriction discourse more frequently articulates rights and freedoms, based on ideas of inviolable rights to freedom of movement or freedom of economic activity or a cost-benefit analysis that places economic activity over public health. Pro-social and economic restriction discourse less frequently mentions rights and freedoms, instead supporting restrictions as following state and medical advice and out of deference and respect to medical professionals. Discourse is highly polarised and divisive and articulated largely through established political identity positions. It is suggested that more attention is paid to discussions of balancing rights and freedoms in COVID control restrictions. To convince opposers of restrictions, supporters of restrictions should base arguments around communal rights and positive freedoms. It is also important to critically evaluate whether and how these perspectives need to be adapted to be appropriate and resonant in democratic and individualistic countries.

Space agencies and private companies prepare the beginning of human space exploration for the 2030s with missions to put the first human on the Mars surface. The absence of gravity and radiation, along with distance, isolation and hostile environments, are expected to increase medical events where previously unseen manifestations may arise. The current healthcare strategy based on telemedicine and the possibility to stabilize and transport the injured crewmember to a terrestrial definitive medical facility is not applicable in exploration class missions. Therefore, the need for deploying the full autonomous capability to solve medical emergencies may guide the design of future onboard healthcare systems. We present ten basic principles and concept design of a software suite to bring onboard decision support to help the crew dealing with medical emergencies taking into consideration physiological disturbances in space and spaceflight restrictions. 1) give real-time support for emergency medical decision making, 2) give patient-specific advice for executive problem-solving, 3) take into account available information from life support and monitoring of crewmembers, 4) be fully autonomous from remote facilities, 5) continuously adapt predictions to physiological disturbance and changing conditions, 6) optimize emergency medical decision making in terms of mission fundamental priorities, 7) take into account medical supplies and equipment on board, 8) apply health standards for the level of care V, 9) implement ethics responsibilities for spaceflights, and 10) apply ethical standards for artificial intelligence. Based on these principles, we propose an autonomous clinical decision support system (CDSS) to provide real-time advice for emergency medical interventions on board of space exploration missions.