Data Science newsletter – December 17, 2018

GROUP CURATION: N/A

In Philadelphia, Mayor Kenney’s administration has emphasized evidence-based decision-making. Last year, the city launched GovLabPHL, a multi-agency collaboration that uses studies of human behavior to shape how the city interacts with residents.

Municipal records are now digitized and easier to share among departments, so Philadelphia and other cities are centralizing data that agencies historically kept to themselves.

World’s first “academic” study on Santa Claus shows that many children discovered the truth behind jolly ol’ Saint Nick from missteps by their parents.

In the minds of artificial-life purists, researchers will only be able to claim success when they have produced a fully functioning cell out of chemically synthesized molecules. That cell will need to reproduce, sustain its own metabolism, and adapt to the environment. And scientists will need to understand what all the cell’s genes do.

If your goal is to make novel life forms from scratch, in other words, you need to do more than just reproduce what already exists. That’s basically just blind imitation, suggests Kate Adamala, a biochemist at the University of Minnesota, Twin Cities. By copying out all the words, she says, “I could say, ‘I wrote A Hundred Years of Solitude,’ and technically I did write it. But I didn’t understand how.”

Tornadoes may form from the ground up, rather than the top down.

That could sound counterintuitive. Many people may picture a funnel cloud emerging from the bottom of a dark mass of thunderstorms and then extending to the ground, atmospheric scientist Jana Houser said December 13 in a news conference at the American Geophysical Union meeting.

In September, the World Bank announced a new partnership with the United Nations, the Red Cross, Microsoft, Google, and Amazon to help end famine. The group will use artificial intelligence and machine learning to predict famine and mobilize early funding to mitigate its effects. The goal of this Famine Action Mechanism (FAM): Ready sufficient funding to respond promptly to food crises.

FAM is part of a new trend of donor-funded programs that try to avert emergencies by responding before crisis hits. But donors and aid organizations already have early warning systems in place. Humanitarian responses to food crises are chronically underfunded. Organizations delivering aid frequently ask donors to give more so they don’t have to reduce the amount of food they provide to the suffering. Warnings that are even earlier aren’t going to help boost funding.

So what will? My research suggests the biggest barriers to preventing famine are not technological, but governmental.

First-of-Its-Kind System Pairs Artificial Neural Networks with Infrared Imaging to Run Small-Scale Chemical Reactions with Big Impact

Google Brain cofounder and former Baidu AI chief scientist Andrew Ng today announced the launch of the “AI Transformation Playbook,” a prescriptive guide for business leaders looking to create AI-first companies.

The playbook is the latest initiative from Ng, who has been advising startups, developers, and legacy businesses on how to build successful ventures in the AI era.

Software is being developed using what is called “forensic linguistics”.

This analyses the composition of a document including vocabulary, punctuation and format.

On the cusp of arXiv’s move to Computing and Information Science (CIS) in January, members of Cornell University Library and CIS celebrated 17 years of the scientific research repository’s growth under library stewardship, and wished it continued success.

Carl A. Kroch University Librarian Gerald R. Beasley, arXiv program director Oya Rieger and CIS Dean Greg Morrisett spoke at a gathering Dec. 7 in the Physical Sciences Building.

“During this two-hour celebration, almost 40,000 [articles] will be downloaded by individuals from education and business organizations, from almost every country,” Rieger said.

The lab will be used to explore ways the IoT can be used in next-generation retail applications.

JD.com and Intel have teamed up to create a joint retail IoT lab to explore how the IoT can be used to create smart retail apps and solutions. The Digitalized Retail Joint Lab will develop IoT based vending machines and media and advertising technologies and applications to be used in stores. All will be built on Intel architecture.

The distribution of employment across industries determines the economic profiles of cities. But what drives the distribution of employment? We study a simple model for the probability that an individual in a city is employed in a given urban activity. The theory posits that three quantities drive this probability: the activity-specific complexity, individual-specific knowhow, and the city-specific collective knowhow. We use data on employment across industries and metropolitan statistical areas in the US, from 1990 to 2016, to show that these drivers can be measured and have measurable consequences. First, we analyze the functional form of the probability function proposed by the theory, and show its superiority when compared to competing alternatives. Second, we show that individual and collective knowhow correlate with measures of urban economic performance, suggesting the theory can provide testable implications for why some cities are more prosperous than others.

The University of Guelph is launching a new hub for artificial intelligence to grapple with ethical questions amidst growing concern around issues of privacy, bias and human-machine interaction.

Zeroing in on the moral side of everything from medical imaging to automated credit card approvals, the Centre for Advancing Responsible and Ethical Artificial Intelligence (CARE-AI) aims to bring together experts to study and teach humanist approaches to AI.

ClimaCell’s software, called HyperCast, works by tapping into communications network infrastructure. As devices wirelessly communicate, their signals are affected by precipitation. These effects are imperceptible when we send a text or stream a video, but ClimaCell uses sophisticated modeling algorithms to analyze the signal disruptions and classify the ground-level weather in that location.

“We’re saying something that is on one hand really quite simple, but on the other we think is quite brilliant,” says ClimaCell chief strategy officer Rei Goffer MBA ’17, who co-founded the company in 2015 with Itai Zlotnik MBA ’17 and Shimon Elkabetz. “A lot of the things around us actually sense the weather, just not by design. They’re affected by the weather, and by looking at how they’re affected you can actually reverse engineer things and see through that.”

For the last 20 years, every consumer-oriented application, service or content platform used “ads” as its default business model. Like always, the road to hell has been paved with good intentions: we assumed that the mostly innocuous billboards could be transposed to the web… but these banner ads quickly became endless trojan horses into our privacy, feelings and opinions.

Our naive “content-wants-to-be-free” approach failed to account for externalities, including the fact that hostile organizations are influencing us through our media consumption… Using ads to support content creation is not working anymore, whether it is from a purely technical aspect when more and more of the provisioned ads are not being displayed, from an economic standpoint where, outside of a shrinking number of larges platforms, nobody is able to even scrape cents per hour spent staring at ads.

Nashville and others are thriving thanks to a mix of luck, astute political choices and well-timed investments, while cities like Birmingham, Ala., fall behind.

Stanford, CA March 4, 2019. [sold out, waiting list available]

Knative provides a set of middleware components that are essential to build modern, source-centric, and container-based applications that can run anywhere: on premises, in the cloud, or even in a third-party data center. Knative components are built on Kubernetes and codify the best practices shared by successful real-world Kubernetes-based frameworks.

We have found that by measuring the gradient noise scale, a simple statistic that quantifies the signal-to-noise ratio of the network gradients[2], we can approximately predict the maximum useful batch size. Heuristically, the noise scale measures the variation in the data as seen by the model (at a given stage in training). When the noise scale is small, looking at a lot of data in parallel quickly becomes redundant, whereas when it is large, we can still learn a lot from huge batches of data.

To make it easier to build and deploy natural language processing (NLP) systems, we are open-sourcing PyText, a modeling framework that blurs the boundaries between experimentation and large-scale deployment. PyText is a library built on PyTorch, our unified, open source deep learning framework.

Many executives are beginning to realize the ethical dimension of AI. A 2018 survey by Deloitte of 1,400 U.S. executives knowledgeable about AI found that 32% ranked ethical issues as one of the top three risks of AI. However, most organizations don’t yet have specific approaches to deal with AI ethics. We’ve identified seven actions that leaders of AI-oriented companies — regardless of their industry — should consider taking as they walk the fine line between can and should.