Join our Folding@Home team:
Main F@H site
Our team page
Support us: Subscribe Here
and buy SoylentNews Swag
We always have a place for talented people, visit the Get Involved section on the wiki to see how you can make SoylentNews better.
This week Google released a report detailing the design and performance characteristics of the Tensor Processing Unit (TPU), its custom ASIC for the inference phase of neural networks (NN). Google has been using the machine learning accelerator in its datacenters since 2015, but hasn't said much about the hardware until now.
In a blog post published yesterday (April 5, 2017), Norm Jouppi, distinguished hardware engineer at Google, observes, "The need for TPUs really emerged about six years ago, when we started using computationally expensive deep learning models in more and more places throughout our products. The computational expense of using these models had us worried. If we considered a scenario where people use Google voice search for just three minutes a day and we ran deep neural nets for our speech recognition system on the processing units we were using, we would have had to double the number of Google data centers!"
The paper, "In-Datacenter Performance Analysis of a Tensor Processing Unit," (the joint effort of more than 70 authors) describes the TPU thusly:
"The heart of the TPU is a 65,536 8-bit MAC matrix multiply unit that offers a peak throughput of 92 TeraOps/second (TOPS) and a large (28 MiB) software-managed on-chip memory. The TPU's deterministic execution model is a better match to the 99th-percentile response-time requirement of our NN applications than are the time-varying optimizations of CPUs and GPUs (caches, out-of-order execution, multithreading, multiprocessing, prefetching, ...) that help average throughput more than guaranteed latency. The lack of such features helps explain why, despite having myriad MACs and a big memory, the TPU is relatively small and low power."
Tourists visiting the town of Klosterneuburg in eastern Austria often head for the 12th century monastery or the nearby memorial to author Franz Kafka. Virologists and evolutionary biologists, however, may one day pay homage to the town's sewage treatment plant, which has yielded a genome that appears to be from the most cell-like viruses yet [DOI: 10.1126/science.aal4657] [DX]. These oddities challenge the controversial hypothesis that so-called giant viruses are descendants of a vanished group of cellular organisms—a fourth domain of life. Instead, the study argues, these outsized viruses have more pedestrian origins.
"I found [the work] very convincing," says environmental virologist Matthias Fischer of the Max Planck Institute for Medical Research in Heidelberg, Germany. "Based on the data available now, I would not put my money on the fourth domain hypothesis."
Most viruses are much smaller than cells and need few genes because they replicate by co-opting the machinery of their hosts. Certain bird and pig viruses, for example, get by with just two genes, compared with nearly 4400 genes in a common strain of the intestinal bacterium Escherichia coli. Because viruses cannot reproduce independently and lack other hallmarks of cellular organisms, biologists have typically blackballed them from the club of life.
The first report of giant viruses, in Science in 2003, jolted researchers. Not only are these viruses larger than many microorganisms, but they can carry more than 2500 genes, surpassing many bacteria. These behemoths required revisions to the evolutionary tree of life, some scientists contended. The standard tree has three main groups, or domains—bacteria, archaea, and eukaryotes. But several researchers proposed that giant viruses are leftovers of a fourth domain of life. In this view, their ancestors were now-extinct cells that over time ditched many genes and became parasites.
On April 6, William Shatner had a Twitter spat with noted physician Dr. David H. Gorski, founder of Science-Based Medicine, a blog which takes a sceptical and scientific look at controversial medical claims. A few days earlier Shatner made pleas to support Autism Speaks, a controversial group which has has been criticised for an approach which stigmatises autism, and for only recently changing its position from suggesting a link between vaccines and autism to accepting the overwhelming science that such a link does not exist (and they still don't seem to have unequivocally rejected anti-vaccine views). Many people attempted to call Shatner out on this, including Dr. Gorski, and Shatner responded by doubling down and responding with hit pieces on Dr. Gorski from dubious pseudoscience sites critical of him, such as TruthWiki, Newstarget, and NaturalNews. Slate has an article about the incident:
With that, millions of followers were treated to a hit piece about Gorski hosted by TruthWiki. It's hard to overstate the unreliability of TruthWiki, a haphazard collection of conspiracy theories and pseudoscientific nonsense riddled with typos and bizarre assertions. The exercise section, for instance, includes only two entries: "Natural Help for Diabetes" and "Deepak Chopra's Eye Exercises."
When Science-Based Medicine objected to Shatner's tweet, he posted another set of links discrediting Gorski, this time to the websites Newstarget (motto: "Obliterating Your Safe Spaces With Truth Bombs") and NaturalNews, which is run by Mike Adams, aka "the Health Ranger"—who also founded TruthWiki.
"All on Google," he added after them, as if that certified their authenticity.
NaturalNews is like TruthWiki but without the veneer of reliability: It's a cesspool of pseudo-scientific insanity seasoned generously with political vitriol and outlandish conspiracy theories. That's not biased journalism—it's the only way to report accurately on the site.
[...]Shatner is a celebrity, which means that he has outsized influence. That he would use his platform to lend credibility to such sites, spreading them to 2.5 million followers, could have terrible consequences. Shatner has made his support of vaccination very clear, but NaturalNews has tons of "articles" demonizing vaccinations—an example is a video titled "Vaccine Cannibalism Exposed." But there's an upside: The real-time tweeting of his thought process provides a helpful window into the practices of everyday intelligent people trying to figure out the truth. They Google, they find a few articles that confirm their biases, and they're done. No matter that the articles are on websites that spread virulent misinformation. What look like dead giveaways of quackery for some go completely unnoticed. This is instructive, even if it's frustrating.
It's unfortunate that Shatner is now on the receiving end of a lot of outrage because of a few misinformed tweets. At the same time, celebrities wield tremendous power, and to be cliché about it, that comes with some responsibility. I can only hope this will serve as a lesson to him and to others about the importance of applying information literacy before tweeting. Ideally, he would have perhaps recognized his misstep and deleted the problematic tweets instead of doubling down, but human pride is a powerful thing.
As for the rest of us, well, let's just say that a part of information literacy is realizing that just because someone is famous doesn't mean they have it.
Dr. Gorski himself (writing under the pseudonym Orac) has written of the affair from his perspective.
Researchers have discovered that the solutions to a famous mathematical function called the Riemann zeta function correspond to the solutions of another, different kind of function that may make it easier to solve one of the biggest problems in mathematics: the Riemann hypothesis. If the results can be rigorously verified, then it would finally prove the Riemann hypothesis, which is worth a $1,000,000 Millennium Prize from the Clay Mathematics Institute.
While the Riemann hypothesis dates back to 1859, for the past 100 years or so mathematicians have been trying to find an operator function like the one discovered here, as it is considered a key step in the proof.
"To our knowledge, this is the first time that an explicit—and perhaps surprisingly relatively simple—operator has been identified whose eigenvalues ['solutions' in matrix terminology] correspond exactly to the nontrivial zeros of the Riemann zeta function," Dorje Brody, a mathematical physicist at Brunel University London and coauthor of the new study, told Phys.org.
What still remains to be proven is the second key step: that all of the eigenvalues are real numbers rather than imaginary ones. If future work can prove this, then it would finally prove the Riemann hypothesis.
[...] Riemann's hypothesis was that all of the nontrivial zeros lie along a single vertical line (½ + it) in the complex plane—meaning their real component is always ½, while their imaginary component i varies as you go up and down the line.
Over the past 150 years, mathematicians have found literally trillions of nontrivial zeros, and all of them have a real of component of ½, just as Riemann thought. It's widely believed that the Riemann hypothesis is true, and much work has been done based on this assumption. But despite intensive efforts, the Riemann hypothesis—that all of the infinitely many zeros lie on this single line—has not yet been proved.
More information:
Carl M. Bender, Dorje C. Brody, and Markus P. Müller. "Hamiltonian for the Zeros of the Riemann Zeta Function." Physical Review Letters. DOI: 10.1103/PhysRevLett.118.130201
Wikipedia: imaginary numbers, Eigenvalues and Eigenvectors, and the Riemann zeta function.
The first tests of Elon Musk's revolutionary high-speed transport system could begin soon after Hyperloop One, one of 12 companies competing to make the idea a reality, completed its test track. The company has finished work on its 500 metre long testing tunnel, which is situated in the Nevada desert, near Las Vegas, and has a diameter of 3.3 meters. It is expected to run initial trials on the near-supersonic speed train in the first half of this year.
The development follows last month's news that Hyperloop Transportation Technologies, another competing company, has started building the first passenger capsule. The pods will be able to carry 28 to 40 passengers at a time and depart every 40 seconds, the company said. They could be ready as early as next year.
The YouTube Partner Program (YPP) has changed its rules, and two Soylentils wrote in to tell us about it:
New channels will have to get 10,000 views before they can be considered for the YouTube Partner Program, the firm announced in a blog post.
YouTube will then evaluate whether the channel is adhering to its guidelines before letting it carry adverts.
It will help clamp down on content theft and fake channels, YouTube said.
"After a creator hits 10k lifetime views on their channel, we'll review their activity against our policies," wrote Ariel Bardin, vice president of product management at YouTube.
"If everything looks good, we'll bring this channel into YPP [YouTube Partner Program] and begin serving ads against their content. Together these new thresholds will help ensure revenue only flows to creators who are playing by the rules."
Stay on message, Citizen. Wrongthink is not allowed.
YouTube is making changes to the YouTube Partner Program. YouTube will make it easier to report a channel impersonating another channel. It will also stop serving ads on channels with less than 10,000 views:
Starting today, we will no longer serve ads on YPP videos until the channel reaches 10k lifetime views. This new threshold gives us enough information to determine the validity of a channel. It also allows us to confirm if a channel is following our community guidelines and advertiser policies.
[...] In a few weeks, we'll also be adding a review process for new creators who apply to be in the YouTube Partner Program. After a creator hits 10k lifetime views on their channel, we'll review their activity against our policies. If everything looks good, we'll bring this channel into YPP and begin serving ads against their content. Together these new thresholds will help ensure revenue only flows to creators who are playing by the rules.
At first, I thought the 10,000 view limit was per video. But it's actually the total amount of views on all videos on the channel. It remains to be seen whether the channel review that takes place after the 10,000 view threshold will be "hands on" enough to actually identify the content YouTube wants wiped away... before it can be used to scare advertisers away from the platform.
Also at The Verge.
Previously: Google Fails to Stop Major Brands From Pulling Ads From YouTube
Given the rising number of extreme weather events, . In an investigation recently published in Nature Climate Change, scientists looked into how quickly benefits of climate mitigation strategies—meaning dropping CO2 emissions—reduce the risk of heat waves.
The researchers answered these questions using climate-model simulations. These models can be run with different levels of emissions, some assuming a very aggressive mitigation scenario with lowered emissions, others assuming emissions that are unchecked, producing significant increases in emissions over time. By comparing model runs with different levels of emissions, the researchers were able to develop an understanding of the time required for effects of mitigation plans to be noticeable.
In particular, the team focused on extreme events that occurred on average once every 10 years when emissions continue to rise unchecked. They then introduced different levels of emissions mitigation until the probability of such an event is half as likely, occurring only once every 20 years. Using this method, the scientists determined that for many regions, it takes less than 20 years of emissions reductions to drop the probability of extreme hot weather by more than 50 percent after mitigation has begun.
Climate change is god's will, isn't it?
NVIDIA issued a press release for its new card, Titan Xp:
Introduced today [April 6], the Pascal-powered TITAN Xp pushes more cores, faster clocks, faster memory and more TFLOPS than its predecessor, the 2016 Pascal-powered TITAN X.
With the new TITAN Xp we're delivering a card to users who demand the very best NVIDIA GPU, directly from NVIDIA and supported by NVIDIA.
Key stats:
- 12GB of GDDR5X memory running at 11.4 Gbps
- 3,840 CUDA cores running at 1.6GHz
- 12 TFLOPs of brute force
This is extreme performance for extreme users where every drop counts.
Open to Mac Community
Speaking of users, we're also making the new TITAN Xp open to the Mac community with new Pascal drivers, coming this month. For the first time, this gives Mac users access to the immense horsepower delivered by our award-winning Pascal-powered GPUs.
TITAN Xp is available now for $1,200 direct from nvidia.com, and select system builders soon.
Don't shoot the messenger.
[More details can be found on the TITAN Xp product page where you can also place an order (Limit 2 per customer). --Ed.]
Germany has followed through on its proposal to make social networks remove slanderous hate speech and fake news or face massive fines.
The nation's Bundesministerium der Justiz und für Verbraucherschutz (Federal Ministry of Justice and Consumer Protection) has announced that cabinet approved a plan to force social network operators to create a complaints mechanism allowing members of the public to report content that online translate-o-tronic services categorise as "insults, libel, slander, public prosecutions, crimes, and threats."
The Bill approved by Cabinet proposes that social networks be required to establish complaints officer who is subject to local law and gets the job of removing obviously criminal content 24 hours after receiving a complaint. A seven-day deadline will apply to content that's not immediately identifiable as infringing. Social networks will also be required to inform complainants of the outcome of their takedown requests and to provide quarterly summaries of their activities.
A 'Historic Book Odour Wheel' which has been developed to document and archive the aroma associated with old books, is being presented in a study in the open access journal Heritage Science. Researchers at UCL Institute for Sustainable Heritage created the wheel as part of an experiment in which they asked visitors to St Paul's Cathedral's Dean and Chapter library in London to characterize its smell.
The visitors most frequently described the aroma of the library as 'woody' (selected by 100% of the visitors who were asked), followed by 'smoky' (86%), 'earthy' (71%) and 'vanilla' (41%). The intensity of the smells was assessed as between 'strong odor' and 'very strong odor'. Over 70% of the visitors described the smell as pleasant, 14% as 'mildly pleasant' and 14% as 'neutral'.
In a separate experiment, the researchers presented visitors to the Birmingham Museum and Art Gallery with an unlabelled historic book smell - sampled from a 1928 book they obtained from a second-hand bookshop in London - and collected the terms used to describe the smell. The word 'chocolate' - or variations such as 'cocoa' or 'chocolatey' - was used most often, followed by 'coffee', 'old', 'wood' and 'burnt'. Participants also mentioned smells including 'fish', 'body odour', 'rotten socks' and 'mothballs'.
Cecilia Bembibre, heritage scientist at UCL and corresponding author of the study said: "Our odour wheel provides an example of how scientists and historians could begin to identify, analyze and document smells that have cultural significance, such as the aroma of old books in historic libraries. The role of smells in how we perceive heritage has not been systematically explored until now."
Will our grandchildren recognize the smell of diesel, or oil?
Are some wolves being 'redomesticated' into dogs?
It happened thousands of years ago, and it may be happening again: Wolves in various parts of the world may have started on the path to becoming dogs. That's the conclusion of a new study, which finds that the animals are increasingly dining on livestock and human garbage instead of their wild prey, inching closer and closer to the human world in some places. But given today's industrialized societies, this closeness might also bring humans and wolves into more conflict, with disastrous consequences for both.
[...] Not everyone is convinced. "I doubt if we're domesticating wolves that eat human-sourced food," says Robert Wayne, an evolutionary biologist and expert on canine genetics at the University of California, Los Angeles. "That diet is more likely to get them killed." Unlike the trash-picking dingoes, which reduced their territories, wolves still range so widely that garbage-eaters are less likely to become genetically isolated from the rest of their population, he says. Bobcats, coyotes, and other animals that are already well-integrated in our neighborhoods are more likely to become domesticated, he adds.
Making a New Dog? (DOI: 10.1093/biosci/bix022) (DX)
Dietary niche overlap of free-roaming dingoes and domestic dogs: the role of human-provided food (open, DOI: 10.1644/13-MAMM-A-145.1) (DX)
Now free: citation data from 14 million papers, and more might come
Consider this: A scientist publishes a study citing other papers. Those cited papers, in turn, cite studies that came before them. But much of that citation information—which is often of great interest to scientists tracking research trends and hot topics—has not been available freely.
Enter the Initiative for Open Citations (I4OC), a project aiming to make citation data free to all, formally announced today by six organizations, including the Wikimedia Foundation, publisher Public Library of Science, and the open-access journal eLife. So far, the initiative has partnered with 29 journal publishers to enable anyone to access citation data from about 14 million papers indexed by Crossref, a nonprofit collaboration that promotes the sharing of scholarly information. And more publishers are likely to sign on, says Mark Patterson, executive director of eLife, in Cambridge, U.K.
Dog attacks on mail carriers hit 6,755 as online sales boom
Booming Online Sales Mean More Dog Bites for Mail Carriers
Booming online retail sales are good news for the U.S. Postal Service, but its carriers are incurring a cost: more dog bites.
Dog attacks on postal workers rose last year to 6,755, up 206 from the previous year and the highest in three decades, as internet shopping booms and consumers increasingly demand seven-day-a-week package delivery and groceries dropped at their doorstep. The high for attacks dated back to the 1980s, at more than 7,000, before maulings by pit bulls and other potentially aggressive dogs became a public issue.
Los Angeles topped the 2016 list with 80 attacks on postal workers, followed by Houston with 62 and Cleveland with 60.
The Postal Service released its annual figures Thursday as part of National Dog Bite Prevention Week, which begins Sunday.
Next week Microsoft will begin the slowish rollout of its big update to Windows 10, the Creators Update.
Right now, it's doing a little damage control, and preempting complaints about privacy, by listing the types of information its operating system will automatically and silently leak from PCs, slabs, and laptops back to Redmond.
When Windows 10 came out, Reg readers were alarmed by the volume of information the software was collecting and sending back to base. Ever since then, Microsoft has been fighting a PR battle to reassure people that such data slurping isn't all bad – it's "just" telemetry and diagnostics and potentially your files.
Now Redmond's had a little rethink for the Creators Update, and decided to come clean on exactly what the software will phone home – even insisting the closed-source operating system will scoop up less surveillance this time.
What makes you think it's your data?
Phys.org reports:
The idea of a life lived modestly is gaining traction. Ten years ago, Samantha Weinberg, a mother of two young children, spent a year not shopping. Her aim was to reduce her environmental impact. The next year, Mark Boyle, founder of the online Freeconomy community, embarked on a life without money in order to sever his connection with it. Since then, others have joined this "Not Spending" movement.
Frugality has its limitations. Not everyone is able-bodied enough to cycle, and if we all started foraging for wild food it would deprive non-human species of nutrients and disrupt local ecosystems. While minimalism has found new converts, especially in Japan, this extreme approach is unlikely to go mainstream.
Perhaps a more realistic hope is for a steady rise in the number of people who discover that pursuing non-material riches brings greater happiness than the getting and spending of money. In fact, significant numbers of "voluntary simplifiers" have been choosing and enjoying lives of material simplicity for decades.
Have Soylentils found greater happiness through simplification?
Goldman Sachs says mining platinum from asteroids is a way for bankers to earn billions and is a realistic goal. Goldman Sachs is bullish on space mining with "asteroid-grabbing spacecraft." In a 98-page note for clients, analyst Noah Poponak and his team argue that platinum mining in space is getting cheaper and easier, and the rewards are becoming greater as time goes by.
"Prospecting probes can likely be built for tens of millions of dollars each and Caltech has suggested an asteroid-grabbing spacecraft could cost $2.6bn." It is believed an asteroid the size of a football field could be worth up to £40 billion. However, bringing that much platinum back to Earth is likely to crash the precious metal market - and probably the rest of the economy with it.