Lots of work is going into the apparently simple, but in fact complex, field of turning things off. As soon as you give any machine a goal, even as innocent as making daisy chains, you give it the subsidiary goal of staying alive, because it cannot make daisy chains if it has been turned off. So any reasonably intelligent system will seek a method to disable the off button. It is the hardest technical challenge AI fans face. But there is no answer yet.

Elon Musk has nightmares about A.I. (2016.1.11)

An industry-wide organisation including five Silicon Valley giants has been formed to promote the fair and ethical development of artificial intelligence technologies. The organisation aims to "conduct research, recommend best practices, and publish research" rather that directly lobby legislators, and will focus on "ethics, fairness and inclusivity; transparency, privacy, and interoperability; collaboration between people and AI systems; and the trustworthiness, reliability and robustness of the technology". The group - known as the Partnership on Artificial Intelligence to Benefit People and Society - includes Facebook, Google, Amazon, Microsoft and IBM, with academic and non-profit organisations expected to join in the near future. "Over the past five years, we've seen tremendous advances in the deployment of AI and cognitive computing technologies," said IBM AI ethics researcher, Francesca Rossi, "ranging from useful consumer apps to transforming some of the world's most complex industries, including healthcare, financial services, commerce and the Internet of Things." "This partnership will provide consumer and industrial users of cognitive systems a vital voice in the advancement of the defining technology of this century - one that will foster collaboration between people and machines to solve some of the world's most enduring problems - in a way that is both trustworthy and beneficial."

Volvo has decided to join the ranks of automakers with offices in Silicon Valley. The Swedish car company is in the process of opening a research center in Mountain View, Lex Kerssemakers, CEO of Volvo's U.S. division, said in an interview. The company is hiring some 70 engineers for the office, he said. Volvo, which is owned by Chinese automaker Geely but operates largely independently, has had an office in Camarillo for about 30 years that focused on car design, Kerssemakers said. Within the past three to four years, engineers based in that office also started to work on car infotainment systems, he said.

The earliest known recording of music produced by a computer - a machine operated by Alan Turing, no less - has finally been made to sound exactly as it did 65 years ago. The performance is halting and the tone reedy. It starts with a few bars of the national anthem, then a burst of Baa Baa Black Sheep, followed by a truncated rendition of Glenn Miller's swing hit In The Mood. ("The machine's obviously not in the mood," an engineer can be heard remarking when it stops mid-way.) But the rudimentary audio track is a landmark - the first time that music played on a computer is known to have been recorded. It was captured by the BBC in the Autumn of 1951 during a visit to the University of Manchester, where the Ferranti Mark 1 - the world's first commercially available general purpose computer - was based.

If you're a business executive (rather than a data scientist or machine learning expert), you've probably been exposed to the mainstream media coverage of artificial intelligence or AI. You've seen articles in The Economist and Vanity Fair, you've seen emotional stories about Tesla Autopilot and the threat of AI to mankind by such luminaries as Stephen Hawking, and you might even have seen Dilbert make jokes about Artificial Intelligence and Human Intelligence. So if you're an executive who cares about growing your business, all this AI media coverage may prompt two nagging questions. First, is the business potential of AI real or not? The answer to the first question is that the business potential of AI is real.

In the public cloud business, scale is everything – hyper, in fact – and having too many different kinds of compute, storage, or networking makes support more complex and investment in infrastructure more costly. So when a big public cloud like Amazon Web Services invests in a non-standard technology, that means something. In the case of Nvidia's Tesla accelerators, it means that GPU compute has gone mainstream. It may not be obvious, but AWS tends to hang back on some of the Intel Xeon compute on its cloud infrastructure, at least compared to the largest supercomputer centers and hyperscalers like Google and Facebook, who tend to get chips earlier in the Xeon product cycle. So it has been – and continues to be – with AWS and its GPU compute instances.

Amazon, Baidu, Facebook, Google and Microsoft, among other technology companies, have been investing heavily in artificial intelligence and related disciplines like machine learning because they see the technology enabling services that become a source of revenue. Consultancy Accenture earlier this week quantified this enthusiasm, predicting that AI "could double annual economic growth rates by 2035 by changing the nature of work and spawning a new relationship between man and machine" and by boosting labor productivity by 40 per cent. Certainly things could work out well for Accenture, which a day later announced a partnership with Google to help companies deploy Google technology like machine learning. It's as if the global services firm has a stake in the future it foresees. But the machine learning algorithms underpinning this harmonious union of people and circuits aren't secure. In a paper [PDF] presented in August at the 25th Annual Usenix Security Symposium, researchers at École Polytechnique Fédérale de Lausanne, Cornell University, and The University of North Carolina at Chapel Hill showed that machine learning models can be stolen and that basic security measures don't really mitigate attacks.

This paper proposes a general adaptive procedure for budget-limited predictor design in high dimensions called two-stage Sampling, Prediction and Adaptive Regression via Correlation Screening (SPARCS). SPARCS can be applied to high dimensional prediction problems in experimental science, medicine, finance, and engineering, as illustrated by the following. Suppose one wishes to run a sequence of experiments to learn a sparse multivariate predictor of a dependent variable $Y$ (disease prognosis for instance) based on a $p$ dimensional set of independent variables $\mathbf X=[X_1,\ldots, X_p]^T$ (assayed biomarkers). Assume that the cost of acquiring the full set of variables $\mathbf X$ increases linearly in its dimension. SPARCS breaks the data collection into two stages in order to achieve an optimal tradeoff between sampling cost and predictor performance. In the first stage we collect a few ($n$) expensive samples $\{y_i,\mathbf x_i\}_{i=1}^n$, at the full dimension $p\gg n$ of $\mathbf X$, winnowing the number of variables down to a smaller dimension $l < p$ using a type of cross-correlation or regression coefficient screening. In the second stage we collect a larger number $(t-n)$ of cheaper samples of the $l$ variables that passed the screening of the first stage. At the second stage, a low dimensional predictor is constructed by solving the standard regression problem using all $t$ samples of the selected variables. SPARCS is an adaptive online algorithm that implements false positive control on the selected variables, is well suited to small sample sizes, and is scalable to high dimensions. We establish asymptotic bounds for the Familywise Error Rate (FWER), specify high dimensional convergence rates for support recovery, and establish optimal sample allocation rules to the first and second stages.

Mathematics is everywhere, if you know where to look. A recently opened exhibition at the North Carolina Museum of Art (NCMA) is displaying the St. John Altarpiece, a 14th-century work by Francescuccio Ghissi. It has nine scenes in total: eight smaller pictures featuring St. John the Evangelist flanking a larger central Crucifixion. At the end of the 19th century, the altarpiece was separated into parts by a saw and eight of the nine resulting panels were sold to different collectors. One panel, the last of the smaller scenes, was lost.