Month: October 2019

https://www.forbes.com/sites/ciocentral/2019/10/14/is-artificial-intelligence-good/#5ac0244d6f70

~ brucelandonblog ~ Leave a comment

Is Artificial Intelligence good?

Bruno Aziza Contributor

You’ve read it in the papers.  You’ve experienced it in life. Machines are taking over.

And they are doing it fast.  Siri turned 9 on October 4 (go ahead, ask her if you don’t believe me). Tesla’s first Autopilot program is also 9.  And Alexa is less than 5 years old.

Despite the fact that these AI-powered technologies haven’t graduated past their first decade, they seem to be running our lives. And they are apparently doing it better than we can.

I wrote about the future of driving 4 years ago.  I’m a big believer in the fact that autonomous cars will take over…and for the better.  However, the spectacular adoption of AI technologies in the last decade has many worried. Is this going too fast?  Should humans slow down this growth? Should you be worried about your job? Or should you be excited about the prospects of productivity and convenience?

How should we all be approaching the ubiquitous occurrence of Artificial Intelligence?

In this post, I’ll attempt to provide a few tools that hopefully help you approach what my friend and co-founder of Cloudera Amr Awadallah has labelled the “Sixth Wave: Automation of Decisions”.

Today In: Innovation

Will Human Jobs Go Away?

You will undoubtedly find a vast body of literature that dissects the potentially harmful consequences of Artificial Intelligence investments.  Just a week ago, AI and Analytics Futurist, Bernard Marr, was warning us about the “The 7 Most Dangerous Technology Trends In 2020 Everyone Should Know About”.  AI Cloning and Facial Recognition made the list.

No later than this past Friday, the McKinsey Global Institute published a research piece which showed that the $20B+ invested in artificial intelligence technologies are bound to impact African Americans the most…

If you’re starting to panic, take a breath and pick up a copy of Andrew McAfee and Erik Brynjolfsson’s book: “The Second Machine Age”.  

And if you don’t have time, watch this 14-min-TedTalk.  In it, Andrew lays out his thesis for Artificial Intelligence impact and adoption.  His points will help you rationalize this trend and will likely taper the creative headlines that predict the dawn of humanity as we know it.

Here are a couple essential take-aways:

  • Automation has occurred.  Don’t deny it.
  • For about 200 years, people have been saying that more technology drives fewer jobs. Data proves this wrong.

“Economies in the developed world have coasted along on something pretty close to full employment,” he says but continues: “this time it’s different” because “our machines have started demonstrating skills they have never, ever had before: understanding, speaking, hearing, seeing, answering, writing, and they’re still acquiring new skills”….

How To Automate Effectively

The short story of it, is that Artificial Intelligence is good.  And it can create incredible opportunities. But it does require inspection and it calls for a method.

When it comes to “automating tasks” at your work, I suggest that you break them up into 3 categories:

  • Automated: these are tasks that are binary, where involving people is counterproductive.  Manufacturing lines have been organized around these requirements for decades (look up Robotic Process Automation or RPA).
  • Augmented: these are tasks that are non-binary and require humans to check on machines OR for machines to check on humans.  This will most likely cover the majority of your tasks.
  • Collaborative: these are tasks that humans are uniquely positioned to initiate, advance and complete.

You might want to come up with your own model.  And there again, lots of literature is available to help you.  You can refer to my earlier post on the latest books in Artificial Intelligence for some key authors and reference books: I would start with Tom Davenport’s “Only Humans Need Apply”.

One thing for sure: the race for Artificial Intelligence domination is well under way.  Nothing will stop it. Slowing down is not in the cards. Competing is required.

If you live in the United States and are preparing to vote in the next 12 months, this is a topic that will become more relevant as the campaign progresses: DJ Patil (former Data Scientist for President Obama) just released the results of a report on America’s investment in Science and Technology.  The United States is at risk of losing its edge. If you don’t have time to read the report, be sure to at least watch this video: this is a topic which deserves attention.

https://www.engadget.com/2019/10/13/alexa-learned-spanish-through-new-tools/

~ brucelandonblog ~ Leave a comment

Here’s how Alexa learned to speak Spanish without your help

New tools helped create thousands of sentences.
Jon Fingas, @jonfingas

10.13.19 in AV
Nicole Lee/Engadget

Now that Alexa knows how to speak Spanish in the US, there’s a common question: how did it learn the language when it didn’t have the benefit of legions of users issuing commands? Through new tools, it seems. Amazon has revealed a pair of system that helped Alexa hone its español (and Hindi, and Brazilian Portugese) using just a tiny amount of reference material. Effectively, they gave the natural language machine learning model a jumpstart.

 

The first tool studies a handful of “golden utterances” (that is, reference commands suggested by the developers) to learn general syntax and semantics patterns. After that, it produces “rewrite expressions” that themselves create thousands of new yet similar sentences to work from. The system works quickly — you could move from 50 utterances to a fully operational linguistic set in less than two days.

Amazon’s other tool uses guided resampling to replace terms that can be safely swapped, further improving the AI’s training. The technique draws both on data from existing Alexa languages as well as media sources like the Amazon Music catalog, and it’s capable enough to be aware of context (it won’t swap a musician’s name for an audiobook, for example).

This doesn’t mean that Alexa will master every known language in a matter of weeks. There are other factors to consider beyond the linguistic structure, such as accommodations for cultural differences and customer support systems. Still, it suggests that Amazon might have an easier time adding languages than it has in the past. It might just be a matter of which Alexa device you get, rather than whether you can get one in the first place.

https://www.express.co.uk/news/science/1189725/antarctica-nasa-stunned-creature-below-ice-south-pole-andrill-spt

~ brucelandonblog ~ Leave a comment

Antarctica shock: How NASA was stunned by ‘unexpected creature’ 600 feet below ice

NASA was left stunned after scientists captured snaps of an “unexpected creature” 600 feet below the ice, resurfaced footage from 2010 shows.

By CALLUM HOARE

Australia: Man spots 4.5-metre crocodile in Daly River

Pause

Unmute

Current Time 0:35
/
Duration 1:00
Loaded: 0%

Progress: 0%

FacebookTwitterShareFullscreen

Antarctica is Earth’s southernmost continent where the geographic South Pole is located. The frozen desert is home to some 1,000 scientists who live in the blistering conditions that reach -90C at times, as they attempt to understand more about the history of Earth. NASA too has a keen interest, both for space exploration and scientific research and, in 2010, a team from the space agency was left “surprised” after drilling into the ice.

They pushed a camera into a hole more than 600 feet below the West Antarctica Ice Sheet and spotted a small shrimp-like creature dancing around.

Bob Bindschadler of NASA’s Goddard Space Flight Centre remembered the moment well, revealing:  “We were like little kids huddling around, just oohing and aahing at this little creature swimming around and giving us a little show.

“It was the thrill of discovery that made us giddy, just totally unexpected.

“We thought we were just going into a deep, dark cold water hole and never anticipated we’d see any life.

NASA researchers captured a bizarre creature below the ice

NASA researchers captured a bizarre creature below the ice (Image: NASA/GETTY)

The team drilled a hole a put a camera down

The team drilled a hole a put a camera down (Image: GETTY)

It was the thrill of discovery that made us giddy, just totally unexpected

Bob Bindschadler

“The colour was what caught our eyes.”

The complex critter was identified as a Lyssianasid amphipod, about three inches in length.

Alberto Behar, of NASA’s Jet Propulsion Laboratory, was the brains behind the state-of-the-art camera used to capture the moment/

He remembered the day well, stating: “This was the first time we’d had a camera able to look back up at the ice.

“This probe is an upgrade to the original, it has three cameras – down, side and back-looking.

JUST IN:

“The back-looking camera saw the shrimp-like animal.

“The real benefit of these exploration programmes is that you go in not knowing what you’re going to find and you get surprised.

“It makes it worth all the trouble putting everything together when you find something new that you didn’t expect.”

Dr Behar designed the original NASA borehole camera apparatus in 1999.

It has since seen six deployments with British, Australian and American science teams in Antarctica, Greenland and Alaska.

DON’T MISS

The tiny creature clutching on to the camera

https://cleantechnica.com/2019/10/12/elon-musk-borrowed-problem-solving-skill-aristotle/

~ brucelandonblog ~ Leave a comment

Elon Musk Lifted A Problem-Solving Skill From Aristotle

Tesla CEO Elon Musk is considered a modern-day Steve Jobs. But does Musk get some of his management playbook from a more ancient source? Over 2300 years ago, Aristotle said that a first principle is the “first basis from which a thing is known” and that pursuing first principles is the key to doing any sort of systemic inquiry — whether in philosophy, as he did, or in business, as Musk does.

 Tesla Motors CEO Elon Musk. Image: screenshot of Tesla video, modified by CleanTechnica.

According to Inc., “First articulated and named by Aristotle, the First Principle has endured all these millennia as the basis for (Western) philosophical contemplation… A First Principle is a basic, essential, foundational truth that is ‘known by nature.’ It is not an assumption or deduction based on another theory or supposition. A key element of First Principle thinking is that just because something is ‘known by nature’ or true in the Universe does not mean it has ever been articulated and described by humans.”

It turns out that, “Physicists, scientists, and artists also engage the First Principle in order to dive into the unknown and surface with ideas totally new. A number of years ago, Elon Musk cited the ancient concept in an interview with Kevin Rose [see below], thereby adding the term to our shared entrepreneurial vocabulary, but most visionaries–whether in business, science, arts, or philosophy–would in fact tell you that First Principle thinking is essential to their process and work, even if the phrase itself is unfamiliar.”

Above: Musk discusses his approach to critical thinking using a “First Principles” framework (Source: innomind)

“Someone could–and people do–say battery packs are really expensive and that’s just the way they will always be because that’s the way they have been in the past. They would say, ‘It’s going to cost $600 per kilowatt-hour. It’s not going to be much better than that in the future,” Musk said in his interview with Rose.

But in First Principle thinking, you forget what has been, you erase what is assumed, and ask questions based on your desire for what is possible. In Musk’s words: “What are the material constituents of the batteries? What is the spot market value of the material constituents? It has carbon, nickel, aluminum, and some polymers for separation, and a steel can. Break that down on a materials basis, if we bought that on a London Metal Exchange, what would each of these things cost?”

Samsung SDI battery cells. Photo by Kyle Field | CleanTechnica.

By using First Principle thinking, “Musk’s findings were that he could get those materials for $80 per kilowatt-hour, combine them into a battery cell shape of his choosing, and model modern innovation within the energy industry… By definition, true innovation can only occur if we start with the First Principle. When we want to make the leap from what is to what is possible, we can’t get to what doesn’t exist by creating an iteration of what already exists.”

About the Author

 is all about Tesla. He’s a TSLA investor, pre-ordered the Model 3, and loves driving the family’s Model S and Model X company cars. As co-founder of EVANNEX, a family business specializing in aftermarket Tesla accessories, he’s served as a contributor/editor of Electric Vehicle University (EVU) and the Owning Model S and Getting Ready for Model 3 books. He writes daily about Tesla and you can follow his work on the EVANNEX blog.

https://www.forbes.com/sites/fernandezelizabeth/2019/10/12/minibrains-grown-in-the-laboratory-produce-brainwaves–now-what/#251f43ba9ac7

~ brucelandonblog ~ Leave a comment

Minibrains Grown In The Laboratory Produce Brainwaves. Now What?

Elizabeth FernandezContributor 
I write about the philosophy and ethics of science and technology

It’s hard to study the human brain. It is the most complex in the animal kingdom with its massive collection of neurons, 80-100 billion to be exact, three times more than chimpanzees. Research relating our brains to the brains of mice and monkeys can only go so far. And because of this complexity, scientists often came up short when studying diseases such as schizophrenia, autism, and Alzheimer’s in the brains of monkeys and mice.

Enter minibrains.

Minibrains are small clusters of human brain cells that can be grown in a Petri dish. Floating through the agar, these small gray lumps don’t look particularly impressive, but they are allowing scientists to study actual living human brain tissue in ways they couldn’t before.

Growing these minibrains gives scientists a chance to study a host of psychological issues and diseases, and perhaps make advancements that they would not have made previously. Minibrains will even be sent to space to study how the human brain develops in zero-G.

But then came the surprise. These lab-grown brains started producing brainwaves.

Today In: Innovation

These brainwaves, equivalent to brain wave patterns in a pre-term infant, were seen by a group of researchers at the University of California San Diego. They reported in a recent paper in Cell Stem Cell that these minibrains began showing neural activity after two months, and in four to six months, they reached levels of neural activity never before seen in a lab. At ten months, they were equivalent to pre-term babies, complete with lulls and flutters of activity.

Minibrains are created by using stem cells, in this case, human skin cells. When stem cells are placed in a conducive environment, they can develop into any organ.

But minibrains are still a far cry from a full human brain. To develop into a mature brain, these minibrains would need to communicate with other areas of a larger brain and have some sort of connection with the outside world. But this might not be far off. Already, scientists have given minibrains retinal cells so they can sense light.

While some note that these minibrains are nowhere near real human brains, others begin to feel uneasy at seeing this neural activity. What does it mean? In this quickly developing field, how soon will these minibrains develop even further? There is an ethical code when dealing with animals in the lab – should this code apply to minibrains too? Could they one day feel pain, have memories, or even become self-aware?

“There is now a need for clear guidelines for research,” says Dr. Nita Farahany and collaborators in a 2018 Letter to Nature. They point out that as research develops and these minibrains become more advanced, it is less far-fetched to believe that one day these minibrains might have some sort of sentience or feelings such as pleasure or pain. The benefits of minibrain research are promising, but they caution, “to ensure the success and social acceptance of this research long term, an ethical framework must be forged now, while brain surrogates remain in the early stages of development.”

Elizabeth Fernandez

Dr. Elizabeth Fernandez is the host of SparkDialog Podcasts, which covers the intersection of science and society.

https://www.teslarati.com/tesla-compact-battery-patent-easy-production/

~ brucelandonblog ~ Leave a comment

Tesla patent paves way for compact battery systems that are easier to produce

Joey Klender
By
Posted on 

Tesla’s use of batteries for its electric vehicles are crucial to their function. The company’s battery systems are the industry standard as they offer more range and density than their competitors. But despite this lead, a recently submitted patent for an aggregated battery system could put Tesla’s batteries head-and-shoulders above the rest of the pack, bar none.

Tesla Pickup Truck still on track for November unveiling

AN ARTIST’S RENDER OF THE TESLA PICKUP TRUCK. (CREDIT: EMRE HUSMAN)

Simon Alvarez
By
Posted on 

It appears that Tesla’s highly-anticipated Pickup Truck is still poised to be unveiled sometime this coming November. The update was shared on Twitter by CEO Elon Musk while responding to an inquiry about the upcoming vehicle’s official reveal date.

While Musk did not specify a date in his recent tweet, a previous announcement from the CEO last month estimated a November unveiling event for the Tesla Pickup Truck. Prior to this, Musk noted in late July that while the vehicle was “close,” the truck’s reveal was “maybe 2 to 3 months” away. This coming November is just a bit over this estimate.

Steve Jobs Ghost 👻@tesla_truth
 · 
Replying to @elonmusk

Have you decided on a date for the pickup reveal? Still targeting November?

Elon Musk

✔@elonmusk

No change

1,075

Twitter Ads info and privacy
128 people are talking about this
Learn more about the benefits of certified pre-owned Acura models.

Interestingly, a November reveal for the Tesla Pickup Truck would mark around two years since the unveiling event of the company’s Semi, which could only be described as the company’s most exciting reveal event to date. Tesla surprised both its enthusiasts and the auto industry as a whole during the Semi event by unveiling its next-generation Roadster, a successor to the car that started it all for the company that boasts an insane 0-60 mph time of 1.9 seconds and range of 620 miles per charge.

Little details are known about the upcoming Tesla Pickup Truck unveiling apart from its expected date, though considering recent developments in the electric car maker’s lineup of vehicles, there seems to be a chance that Tesla could do a “One More Thing” portion on its pickup’s reveal event. With this in mind, a potential vehicle that might make a surprise appearance could be the Model S Plaid Powertrain variant.

Thanks to the Model S’ track capabilities as exhibited in the Plaid Powertrain variant’s performance in Laguna Seca and the Nurburgring, interest in Tesla’s flagship sedan is fairly high once more. Thus, it would be a good idea for Tesla to showcase some of its recent projects involving its flagship sedan during the pickup truck’s unveiling. Such a gesture will likely reaffirm the Model S’ place in the premium EV sedan market, especially considering the arrival of vehicles like the Porsche Taycan Turbo S, a car that was bred on the track.

Elon Musk has teased several notable aspects of Tesla’s upcoming pickup truck, with the CEO stating during the 2019 Annual Shareholder Meeting that the vehicle will feature performance that’s comparable to a base Porsche 911 while boasting a towing capacity that can match industry leaders like the Ford F-150. “If the (Ford) F-150 can tow it, the Tesla truck can do it,” Musk said.

Perhaps the most interesting aspect of the Tesla Pickup Truck would be its starting price. Musk has stated that the vehicle will be priced at “well under” $50,000. The CEO also added that at most, the vehicle should have a starting price of around $49,000. “You should be able to buy a really great truck for $49k or less,” Musk said.

https://www.teslarati.com/tesla-patent-hood-hinge-pedestrian-safety/

~ brucelandonblog ~ Leave a comment

Tesla patents novel hood hinge that optimizes pedestrian safety during collisions

THE MODEL X IS TESLA’S LARGEST VEHICLE IN ITS CURRENT LINEUP. (CREDIT: NICK.LAUER VIA MY TESLA ADVENTURE/INSTAGRAM)

Simon Alvarez
By
Posted on 

Tesla’s electric cars are known for being extremely quick, and they are also known for being extremely safe. The Model 3, the company’s most affordable car to date, for example, has aced safety ratings across the globe, earning a 5-Star rating from the NHTSA in the US, the Euro NCAP in Europe, and the ANCAP in Australia. Even the IIHS gave the Tesla Model 3 its highest rating, Top Safety Pick+.

But this is Tesla, and the electric car maker is known for being a company that refuses to stay still. Its cars are already quick enough to give passengers serious Gs while launching, yet the company remains hard at work on making them even quicker and more visceral in terms of speed (e.g. the Model S Plaid Powertrain). In the same light, while Teslas are already safe at their current state, it is no surprise that the company remains dedicated to finding ways to make its vehicles even safer, both for passengers in the cabin and for pedestrians on the road.

One such example of this was highlighted in a recently published patent that was simply titled “Hinge Assembly for a Vehicle Hood.” Based on the electric car maker’s discussion, the novel hinge assembly has the potential to protect pedestrians who happen to hit the vehicle’s hood during a collision. Similar systems are in place in vehicles today, though Tesla maintained that conventional designs have lots of areas for improvement.

A side view of Tesla’s hinge assembly. (Credit: US Patent Office)
If you’re shopping for a new vehicle, it’s good to know the key differences between these two models.

“Modern vehicles are mandated by safety standards to protect pedestrians from head-impact injuries, including a scenario in which a pedestrian would contact the vehicle’s hood. To meet these requirements. Current state of the art safety systems are active systems that typically include a sensor system to detect a collision with pedestrian and fire (using a pyrotechnic) an actuator to lift the front hood into a protective position before pedestrian impact. However, such systems may be falsely triggered and can only be used once because the pyrotechnic is not reversible. The pyrotechnic is also expensive, adding to overall cost of the vehicle. Therefore, there is a need for a safety system that overcomes the aforementioned drawbacks.”

Tesla noted in its patent’s description that its hinge assembly includes a body member and a hood member, with the latter being “pivotally coupled with a body member through a pivot pin.” In the event of a collision, a portion of the vehicle’s hood member or body member “deforms such that the hood member or body member disengages from the pivot pin.” This allows Tesla to use the hinge as a passive pedestrian safety system that does not require any additional components such as sensors or controllers. The design outlined in Tesla’s patent is also more practical than the pyrotechnic system used in conventional pedestrian impact safety systems.

Tesla describes how its hood hinge works in a collision in the following section.

A side view of Tesla’s shows the hinge assembly being impacted by a pedestrian head. (Credit: US Patent Office)

“FIG. 6 illustrates impact of a headform 602 on hinge assembly 116. Headform 602 represents the head (or portion thereof) of a pedestrian or other living being. As illustrated, when a collision occurs such that headform 602 hits a portion of hood member 108 of vehicle 100 along direction of an axis X-X′, a force is generated. When the force is great enough, the impact force causes hood member 108 to disengage from hinge assembly 116. The impact force typically causes deformation of portion 314 of hood member 108 adjacent to notch 312 such that pivot pin 202 disengages with second opening 304 of hood member 108. In embodiments, the width W of notch 312 is altered to change the impact force at which the hood member 108 disengages from hood member 108. In embodiments the impact force causes deformation of the pivot pin 202 to allow disengagement of hood member 108 from body member 110.

“In an event of collision, hood member 108 may disengage with hinge assembly 116 such that safety standards can be met. Hood member 108 may move down due to impact force and disengagement with hinge assembly 116. To allow movement of hood member 108, sufficient space may be provided by trimming away portions of hood member 108 and body member 110. Advantageously, this would lower weight of components while maintaining the safety standards for vehicle 100.”

Tesla is a carmaker that will likely never stay still. Despite its significant lead in the electric car segment thanks to its vehicles’ batteries and powertrain, Tesla is in a continuous process of improvement. The hood hinge outlined in this patent might be quite simple, but it contributes to the overall safety of Tesla’s lineup of vehicles nonetheless. Such initiatives, if any, further prove that when it comes to safety, no part is too small for innovation — and in the event of a collision, it’s these factors that can make all the difference.

Tesla’s patent for its hinge assembly could be accessed below.

Tesla Hood Patent by Simon Alvarez on Scribd

https://www.sciencealert.com/babies-who-are-cuddled-more-seem-to-have-their-genetics-altered-for-years-afterwards

~ brucelandonblog ~ Leave a comment

Babies Who Are Cuddled More Seem to Have Their Genetics Altered For Years Afterwards

DAVID NIELD
12 OCT 2019

The amount of close and comforting contact that young infants get doesn’t just keep them warm, snug, and loved.

A 2017 study says it can actually affect babies at the molecular level, and the effects can last for years.

Based on the study, babies who get less physical contact and are more distressed at a young age, end up with changes in molecular processes that affect gene expression.

The team from the University of British Columbia in Canada emphasises that it’s still very early days for this research, and it’s not clear exactly what’s causing the change.

But it could give scientists some useful insights into how touching affects the epigenome – the biochemical changes that influence gene expression in the body.

During the study, parents of 94 babies were asked to keep diaries of their touching and cuddling habits from five weeks after birth, as well as logging the behaviour of the infants – sleeping, crying, and so on.

Four-and-a-half years later, DNA swabs were taken of the kids to analyse a biochemical modification called DNA methylation.

It’s an epigenetic mechanism in which some parts of the chromosome are tagged with small carbon and hydrogen molecules, often changing how genes function and affecting their expression.

The researchers found DNA methylation differences between “high-contact” children and “low-contact” children at five specific DNA sites, two of which were within genes: one related to the immune system, and one to the metabolic system.

DNA methylation also acts as a marker for normal biological development and the processes that go along with it, and it can be influenced by external, environmental factors as well.

Then there was the epigenetic age, the biological ageing of blood and tissue. This marker was lower than expected in the kids who hadn’t had much contact as babies, and had experienced more distress in their early years, compared with their actual age.

“In children, we think slower epigenetic ageing could reflect less favourable developmental progress,” said one of the team, Michael Kobor.

In fact, similar findings were spotted in a study from 2013 looking at how much care and attention young rats were given from a very early age.

Gaps between epigenetic age and chronological age have been linked to health problems in the past, but again it’s too soon to draw those kind of conclusions: the scientists readily admit they don’t yet know how this will affect the kids later in life.

We are also talking about less than 100 babies in the study, but it does seem that close contact and cuddles do somehow change the body at a genetic level.

Of course it’s well accepted that human touch is good for us and our development in all kinds of ways, but this is the first study to look at how it might be changing the epigenetics of human babies.

It will be the job of further studies to work out why, and to investigate whether any long-term changes in health might appear as a consequence.

“We plan to follow up on whether the ‘biological immaturity’ we saw in these children carries broad implications for their health, especially their psychological development,” said one of the researchers, Sarah Moore.

“If further research confirms this initial finding, it will underscore the importance of providing physical contact, especially for distressed infants.”

The research was published in Development and Psychopathology.

A version of this article was first published in November 2017.

Learn More

  1. Milestones in transcription and chromatin published in the Journal of Biological Chemistry
    Joel M. Gottesfeld, Journal of Biological Chemistry, 2019
  2. Disorders of infant feeding
    Helen McElroy, BMJ Best Practice, 2018
  1. Genetics as a Modernization Program: Biological Research at the Kaiser Wilhelm Institutes and the Political Economy of the Nazi State
    Bernd Gausemeier, Historical Studies in the Natural Sciences, 2010
  2. Biliary atresia
    Jessi Erlichman et al., BMJ Best Practice, 2018

https://phys.org/news/2019-10-quantum-faster.html

~ brucelandonblog ~ Leave a comment

New compiler makes quantum computers two times faster

by Epiqc, University of Chicago

UChicago-Developed Compiler Makes Quantum Computers 2x Faster
A flow chart describing the compiling of variational algorithms to speed up quantum computations. Credit: EPiQC/University of Chicago

A new paper from researchers at the University of Chicago introduces a technique for compiling highly optimized quantum instructions that can be executed on near-term hardware. This technique is particularly well suited to a new class of variational quantum algorithms, which are promising candidates for demonstrating useful quantum speedups. The new work was enabled by uniting ideas across the stack, spanning quantum algorithms, machine learning, compilers, and device physics. The interdisciplinary research was carried out by members of the EPiQC (Enabling Practical-scale Quantum Computation) collaboration, an NSF Expedition in Computing.

Adapting to a New Paradigm for Quantum Algorithms

The original vision for  dates to the early 1980s, when physicist Richard Feynman proposed performing molecular simulations using just thousands of noise-less qubits (quantum bits), a practically impossible task for traditional computers. Other algorithms developed in the 1990s and 2000s demonstrated that thousands of noise-less qubits would also offer dramatic speedups for problems such as database search, integer factoring, and matrix algebra. However, despite recent advances in quantum hardware, these algorithms are still decades away from scalable realizations, because current hardware features noisy qubits.

To match the constraints of current and near-term quantum computers, a new paradigm for variational quantum algorithms has recently emerged. These algorithms tackle similar computational challenges as the originally envisioned quantum algorithms, but build resilience to noise by leaving certain internal program parameters unspecified. Instead, these internal parameters are learned by variation over repeated trials, guided by an optimizer. With a robust optimizer, a variational  can tolerate moderate levels of noise.

While the noise resilience of variational algorithms is appealing, it poses a challenge for compilation, the process of translating a mathematical algorithm into the physical instructions ultimately executed by hardware.

“The trade-off between variational and traditional quantum algorithms is that while variational approaches are cheap in the number of gates, they are expensive in the number of repetitions needed,” said Fred Chong, the Seymour Goodman Professor of Computer Science at UChicago and lead PI for EPiQC. “Whereas traditional quantum algorithms are fully specified at execution time and thereby fully optimizable pre-execution, variational programs are only partially specified at execution time.”

Partial Compilation

The researchers address the issue of partially specified programs with a parallel technique called partial compilation. Pranav Gokhale, a UChicago PhD student explains, “Although we can’t fully compile a variational algorithm before execution, we can at least pre-compile the parts that are specified.” For typical variational algorithms, this simple heuristic alone is sufficient, delivering 2x speedups in quantum runtime relative to standard gate-based compilation techniques. Since qubits decay exponentially with time, this runtime speedup also leads to reductions in error rates.

For more complicated algorithms, the researchers apply a second layer of optimizations that numerically characterize variations due to the unspecified parameters, through a process called hyperparameter optimization. “Spending a few minutes on hyperparameter tuning and partial compilation leads to hours of savings in execution time”, summarizes Gokhale. Professor Chong notes that this theme of realizing cost savings by shifting resources—whether between traditional and quantum computing or between compilation and execution—echoes in several other EPiQC projects.

The researchers next aim to demonstrate their work experimentally. Such experimental validation has become possible only recently, with the release of cloud-accessible quantum computers that can be controlled at the level of analog pulses. This level of control is much closer to hardware than standard gate-based control, and the researchers expect to realize greater efficiency gains from this pulse interface.

The researchers’ paper, “Partial Compilation of Variational Algorithms for Noisy Intermediate-Scale Quantum Machines” (arXiv link) will be presented at the MICRO computer architecture conference in Columbus, Ohio on October 14. Gokhale and Chong’s co-authors include Yongshan Ding, Thomas Propson, Christopher Winkler, Nelson Leung, Yunong Shi, David I. Schuster, and Henry Hoffmann, all also from the University of Chicago.

Explore further

Research provides speed boost to quantum computers

More information: Partial Compilation of Variational Algorithms for Noisy Intermediate-Scale Quantum Machines, arXiv:1909.07522 [quant-ph] https://arxiv.org/abs/1909.07522 . DOI: 10.1145/3352460.3358313
Provided by University of Chicago

https://www.technologyreview.com/f/614551/ai-computer-vision-algorithms-on-your-phone-mit-ibm/

~ brucelandonblog ~ Leave a comment

Powerful computer vision algorithms are now small enough to run on your phone

An image of hand gestures being recognized on a mobile phone

Researchers have shrunk state-of-the-art computer vision models to run on low-power devices.

Growing pains: Visual recognition is deep learning’s strongest skill. Computer vision algorithms are analyzing medical images, enabling self-driving cars, and powering face recognition. But training models to recognize actions in videos has grown increasingly expensive. This has fueled concerns about the technology’s carbon footprint and its increasing inaccessibility in low-resource environments.

The research: Researchers at the MIT-IBM Watson AI Lab have now developed a new technique for training video recognition models on a phone or other device with very limited processing capacity. Typically, an algorithm will process video by splitting it up into image frames and running recognition algorithms on each of them. It then pieces together the actions shown in the video by seeing how the objects change over subsequent frames. The method requires the algorithm to “remember” what it has seen in each frame and the order in which it has seen it. This is unnecessarily inefficient.

In the new approach, the algorithm instead extracts basic sketches of the objects in each frame, and overlays them on top of one another. Rather than remember what happened when, the algorithm can get an impression of the passing of time by looking at how the objects shift through space in the sketches. In testing, the researchers found that the new approach trained video recognition models three times faster than the state of the art. It was also able to quickly classify hand gestures with a small computer and camera running only on enough energy to power a bike light.

Why it matters: The new technique could help reduce lag and computation costs in existing commercial applications of computer vision. It could, for example, make self-driving cars safer by speeding up their reaction to incoming visual information. The technique could also unlock new applications that previously weren’t possible, such as by enabling phones to help diagnose patients or analyze medical images.

Distributed AI: As more and more AI research gets translated into applications, the need for tinier models will increase. The MIT-IBM paper is part of a growing trend to shrink state-of-the-art models to a more manageable size.