For almost two decades, on-street direct fundraisers for the NGO sector have endured a lifetime's worth of criticism from the public. It appears we aren't all that enamoured by the presence of friendly but unfamiliar people fundraising in our cities and towns for various causes. The name "chuggers" – which, according to the Urban Dictionary, was coined in 2002 by a columnist for the London Metro newspaper – was as insulting and distasteful then as it is now. Still the "chuggers" gonna "chug", "chug", "chug", "chug" while the haters gonna, well, they're no longer as vocal as they once were. Many charities continue to target potential new donors through direct, public fundraising despite the negative press.

Computers are increasingly used to make decisions that have significant impact in people's lives. Often, these predictions can affect different population subgroups disproportionately. As a result, the issue of fairness has received much recent interest, and a number of fairness-enhanced classifiers and predictors have appeared in the literature. This paper seeks to study the following questions: how do these different techniques fundamentally compare to one another, and what accounts for the differences? Specifically, we seek to bring attention to many under-appreciated aspects of such fairness-enhancing interventions. Concretely, we present the results of an open benchmark we have developed that lets us compare a number of different algorithms under a variety of fairness measures, and a large number of existing datasets. We find that although different algorithms tend to prefer specific formulations of fairness preservations, many of these measures strongly correlate with one another. In addition, we find that fairness-preserving algorithms tend to be sensitive to fluctuations in dataset composition (simulated in our benchmark by varying training-test splits), indicating that fairness interventions might be more brittle than previously thought.

In 1976, philosopher Julian Jaynes issued the provocative theory that recent ancestors lacked self-awareness. Instead, they mistook their inner voices for outside sources–the voice of God, say, or the ghosts of their ancestors. Jaynes called his theory "bicameralism" (Westworld fans will recall an episode from the last season called "The Bicameral Mind") and, in his telling, it persisted in early humans until about 3,000 years ago. We are in a similar pre-conscious state now, but the voice we hear is not the other side of our brains. It's our digital self–a version of us that is quickly becoming inseparable from our physical self. I call this comingled digital and analog self our "Meta Me."

Khalifa University has filed a provisional patent application for a technology that can assess the general mood of a crowd. The researchers developed a computer-implemented method for crowd emotion recognition where a video feed of the crowd is processed through algorithms that analyse the movements and postures of the crowd, to determine if they are bored, aggressive, joyful, disgusted, fearful, surprised, sad, proud or angry. The research team focused their innovation on using machine learning and artificial intelligence-related technologies to analyse both the group's movement patterns, and the postures and behaviours of individuals in the crowd, to assess the crowd's general mood. The lead inventor on the patent filing, Sohailah Alyammahi, developed the solution with assistant professor of Electrical and Computer Engineering Dr. Harish Bhaskar, EBTIC Chief Researcher Dr. Dymitr Ruta, and Associate Professor of Professor of Electrical and Computer Engineering Dr. Andrzej Stefan Sluzek. The new technology has intended applications in the security and the video game industries.

The Agung volcano erupts, spewing magma and ash thousands of feet into the air on the island of Bali in Indonesia in November 2017.Josh Edelson/ZUMA This story was originally published by Grist and appears here as part of the Climate Desk collaboration. A trope of sci-fi movies these days, from Snowpiercer to Geostorm, is that our failure to tackle climate change will eventually force us to deploy an arsenal of unproven technologies to save the planet. Think sun-deflecting space mirrors or chemically altered clouds. And because these are sci-fi movies, it's assumed that these grand experiments in geoengineering will go horribly wrong. The fiction, new evidence suggests, may be much closer to reality than we thought.

DURHAM – A Durham-based financial management company has raised $770,000, according to a filing with the Securities and Exchange Commission. Trill, which uses artificial intelligence to manage clients' finances, plans to raise an additional $1.23 million, or a total of $2 million. The company did not indicate its intentions for the raised money. Previously, NC Biz News covered other fundraising by the finance company. In October 2017, the company raised $370,000 in a push to raise $1 million.

Five days into an epic court battle that was expected to last for weeks, Uber abruptly settled a lawsuit Friday brought by Waymo, part of Google parent Alphabet, over the alleged theft of trade secrets. As part of the deal, company officials said, Uber has agreed to pay 0.34 percent of the company's equity at a $72 billion valuation -- a sum that exceeds $244 million. The deal also includes a guarantee that Uber will not incorporate confidential Waymo self-driving-car technology into Uber's hardware and software, a Waymo spokesman said. The deal was announced in a San Francisco federal court before testimony began, shocking many in the audience and leading to a rapid succession of prepared statements from each side. Uber chief executive Dara Khosrowshahi released a statement saying he wants to "express regret" and commit Uber to taking steps to ensure that its self-driving technology "represents just our good work."

Artificial Intelligence (AI) is the broader concept of machines being able to carry out tasks in a way that we humans would consider'smart'. Machine Learning (ML), a subset of AI, refers to the idea of machines using data to make smart decisions -- i.e. decisions they are not necessarily programmed to make. Today, AI is growing faster than ever. A simple Google search lists the various industries and economies AI is threatening to disrupt -- finance, education, science, law… An AI system beats a Grandmaster at chess, and then at Go. Watson, an AI system designed by IBM, beats two Jeopardy champions. Achievements and milestones of these super-complex technologies are headlines of Techcrunch and Quartz articles, scoured by scholars, technology aficionados and VCs looking for'the next big thing'.

"I have read and understood…" has got to be one of the biggest lies people commit on a regular basis. It's the typical ending for the long-winded customer agreements or privacy policies attached to every online service, which few ever read. When humankind finds something difficult, we typically build a gizmo to do it for us -- and this case is no different. It turns out, reading lengthy fine-print is the expertise of a machine-learning artificial intelligence (AI) designed by researchers from the Federal Institute of Technology at Lausanne, Switzerland (EPFL), the University of Wisconsin, and the University of Michigan. Their research began with a question, said lead researcher Hamza Harkous from EPFL.

Though there is a growing body of literature on fairness for supervised learning, the problem of incorporating fairness into unsupervised learning has been less well-studied. This paper studies fairness in the context of principal component analysis (PCA). We first present a definition of fairness for dimensionality reduction, and our definition can be interpreted as saying that a reduction is fair if information about a protected class (e.g., race or gender) cannot be inferred from the dimensionality-reduced data points. Next, we develop convex optimization formulations that can improve the fairness (with respect to our definition) of PCA and kernel PCA. These formulations are semidefinite programs (SDP's), and we demonstrate the effectiveness of our formulations using several datasets. We conclude by showing how our approach can be used to perform a fair (with respect to age) clustering of health data that may be used to set health insurance rates.