Federated learning has attracted attention in recent years for collaboratively training data on distributed devices with privacy-preservation. The limited network capacity of mobile and IoT devices has been seen as one of the major challenges for cross-device federated learning. Recent solutions have been focusing on threshold-based client selection schemes to guarantee the communication efficiency. However, we find this approach can cause biased client selection and results in deteriorated performance. Moreover, we find that the challenge of network limit may be overstated in some cases and the packet loss is not always harmful. In this paper, we explore the loss tolerant federated learning (LT-FL) in terms of aggregation, fairness, and personalization. We use ThrowRightAway (TRA) to accelerate the data uploading for low-bandwidth-devices by intentionally ignoring some packet losses. The results suggest that, with proper integration, TRA and other algorithms can together guarantee the personalization and fairness performance in the face of packet loss below a certain fraction (10%-30%).

Natural language processing (NLP) methods for analyzing legal text offer legal scholars and practitioners a range of tools allowing to empirically analyze law on a large scale. However, researchers seem to struggle when it comes to identifying ethical limits to using natural language processing (NLP) systems for acquiring genuine insights both about the law and the systems' predictive capacity. In this paper we set out a number of ways in which to think systematically about such issues. We place emphasis on three crucial normative parameters which have, to the best of our knowledge, been underestimated by current debates: (a) the importance of academic freedom, (b) the existence of a wide diversity of legal and ethical norms domestically but even more so internationally and (c) the threat of moralism in research related to computational law. For each of these three parameters we provide specific recommendations for the legal NLP community. Our discussion is structured around the study of a real-life scenario that has prompted recent debate in the legal NLP research community.

Surveillance cameras, like the one here in Boston, are used throughout Massachusetts. The state now regulates how police use facial recognition technology. Surveillance cameras, like the one here in Boston, are used throughout Massachusetts. The state now regulates how police use facial recognition technology. Massachusetts lawmakers passed one of the first state-wide restrictions of facial recognition as part of a sweeping police reform law.

The Commission proposes a risk–based approach based on the level of risk presented by the AI system, with different levels of risk attracting corresponding compliance requirements. The risk categories include (i) unacceptable risk (these AI systems are prohibited); (ii) high-risk; (iii) limited risk; and (iv) minimal risk.

For a couple of years, the EU Commission has worked on rules, regulations, and incentives around Artificial Intelligence. On April 21st, it released the long-awaited proposal for harmonized regulations on AI. The proposal to be adopted in national legislation is a far-reaching set of rules that will impact all organizations, at par and perhaps even more comprehensive, than GDPR. Understanding the basics of the new rules and the organizational impact is a requirement for all senior executives. An

We initiate the study of fairness for ordinal regression, or ordinal classification. We adapt two fairness notions previously considered in fair ranking and propose a strategy for training a predictor that is approximately fair according to either notion. Our predictor consists of a threshold model, composed of a scoring function and a set of thresholds, and our strategy is based on a reduction to fair binary classification for learning the scoring function and local search for choosing the thresholds. We can control the extent to which we care about the accuracy vs the fairness of the predictor via a parameter. In extensive experiments we show that our strategy allows us to effectively explore the accuracy-vs-fairness trade-off and that it often compares favorably to "unfair" state-of-the-art methods for ordinal regression in that it yields predictors that are only slightly less accurate, but significantly more fair.

This paper proposes a comprehensive analysis of existing concepts coming from different disciplines tackling the notion of intelligence, namely psychology and engineering, and from disciplines aiming to regulate AI innovations, namely AI ethics and law. The aim is to identify shared notions or discrepancies to consider for qualifying AI systems. Relevant concepts are integrated into a matrix intended to help defining more precisely when and how computing tools (programs or devices) may be qualified as AI while highlighting critical features to serve a specific technical, ethical and legal assessment of challenges in AI development. Some adaptations of existing notions of AI characteristics are proposed. The matrix is a risk-based conceptual model designed to allow an empirical, flexible and scalable qualification of AI technologies in the perspective of benefit-risk assessment practices, technological monitoring and regulatory compliance: it offers a structured reflection tool for stakeholders in AI development that are engaged in responsible research and innovation.Pre-print version (achieved on May 2020)

An institution, be it a body of government, commercial enterprise, or a service, cannot interact directly with a person. Instead, a model is created to represent us. We argue the existence of a new high-fidelity type of person model which we call a digital voodoo doll. We conceptualize it and compare its features with existing models of persons. Digital voodoo dolls are distinguished by existing completely beyond the influence and control of the person they represent. We discuss the ethical issues that such a lack of accountability creates and argue how these concerns can be mitigated.

Last year I wrote about how AI regulations will lead to the emergence of professional AI risk managers. This has already happened in the financial sector where regulations patterned after Basel rules have created a financial risk management profession to assess financial risks. Last week, the EU published a 108-page proposal to regulate AI systems. This will lead to the emergence of professional AI risk managers. The proposal doesn't cover all AI systems, just those deemed high-risk, and the regulation would vary depending on how risky the specific AI systems are: Since systems with unacceptable risks would be banned outright, most of the regulation is about high-risk AI systems.

The algorithms meant to make life easy could further divide us if we're not paying attention The current moment is asking us to evaluate almost every systemic issue in our society. While it might be uncomfortable, the only way through it is to look at the systems that control who wins and who loses and ask, "are they biased?" It's convenient to think that our work in technology is above these issues. A lot of folks in tech do this work to change the world or help communities in need. But just like the rest of society, we need to take this moment of self-reflection seriously, lest our failures are judged by the next generation as harshly as we're judging our elders now.