Reject it or embrace it. Either way, artificial intelligence is here to stay.

Nobody can predict the future with absolute precision.

But when it comes to the impact of artificial intelligence (AI) on peoples careers, the recent past provides some intriguing clues.

Rhonda Scharfs bookAlexa Is Stealing Your Job: The Impact of Artificial Intelligence on Your Futureoffers some insights and predictions that are well worth our consideration.

In the first two parts of my conversation with Rhonda (see What Role Will [Doe]) Artificial Intelligence Play In Your Life? and Artificial Intelligence, Privacy, And The Choices You Must Make) we discussed the growth of AI in recent years and talked about the privacy concerns of many AI users.

In this final part, we look at how AI is affectingand will continue to affectpeoples career opportunities.

Spoiler alert: theres some good news here.

Rodger Dean Duncan:You quote one researcher who says robots are not here to take away our jobs, theyre here to give us a promotion. What does that mean?

Rhonda Scharf:Much like the computer revolution, we need jobs to maintain the systems that have been created. This creates new, desirable jobs where humans work alongside technology. These new jobs are called the trainers, explainers, and sustainers.

Trainers will teach a machine what it needs to do. For instance, we need to teach a machine that when I yell at it (loud voice), I may be frustrated. It needs to be taught that when I ask it to call Robert, who Robert is and what phone number should be used. Once the machine has a basic understanding, it continues to self-learn, but it needs the basics taught to it (like children do.)

Rhonda Scharf

Explainers are human experts who explain computer behavior to others. They would explain, for example, why a self-driving car performed in a certain way. Or why AI sold shares in a stock at a certain point of the day. The same way lawyers can explain why someone acted in self-defense, when initially his or her actions seemed inappropriate, we need explainers to tell us why a machine did what it did.

Sustainers ensure that our systems are functioning correctly, safely, and responsibly. In the future, theyll ensure that AI systems uphold ethical standards and that industrial robots dont harm humansbecause robots dont understand that were fragile, unlike machinery.

There are going to be many jobs that AI cant replace. We need to think, evolve, interpret, and relate. As smart as a chatbot can be, it will never have the same qualities as my best friend. We will need people for the intangible side of relationships.

Duncan:What should people look for to maximize their careers through the use of AI?

Scharf:According to the World Economic Forum, the top 10 in-demand skills for 2020 include complex problem-solving, critical thinking, creativity, emotional intelligence, judgment and decision-making, and cognitive flexibility. These are the skills that will provide value to your organization. By demonstrating all of these skills, you will be positioning yourself as a valuable resource. Well have AI to handle basic tasks and administrative work. People need complex thinking to propel organizations forward.

Duncan:Bonus: What question do you wish I had asked, and how would you respond?

If you don't want to be left behind, you'd better get educated on AI.

Scharf:I wished you had asked how I felt about artificial intelligence. If I was afraid for my future, for the future of my children, and my childrens children?

The answer is no. I dont think that AI is all the doom and gloom that has been publicized. I also dont believe were about to lead a life of leisure and have the world operate on its own either.

As history has shown us, these types of life-altering changes happen periodically. This is the next one. I believe the way we work is about to change, the same way it changed during the Industrial Revolution, the same way it evolved in response to automation. The way we live is about to change. (Think pasteurization and food storage.) Those who adapt will have a better life for it, and those who refuse to adapt will suffer.

Im confident that I will still be employed for as long as I want to be. My children have only known a life with computers and are open to change, and my future grandchildren will only know a life with AI.

Im excited about our future. Im excited about what AI can bring to my life. I embrace Alexa and all her friends and welcome them into my home.

Link:

How Will Your Career Be Impacted By Artificial Intelligence? - Forbes

Emulators speed up simulations, such as this NASA aerosol model that shows soot from fires in Australia.

By Matthew HutsonFeb. 12, 2020 , 2:35 PM

Modeling immensely complex natural phenomena such as how subatomic particles interact or how atmospheric haze affects climate can take many hours on even the fastest supercomputers. Emulators, algorithms that quickly approximate these detailed simulations, offer a shortcut. Now, work posted online shows how artificial intelligence (AI) can easily produce accurate emulators that can accelerate simulations across all of science by billions of times.

This is a big deal, says Donald Lucas, who runs climate simulations at Lawrence Livermore National Laboratory and was not involved in the work. He says the new system automatically creates emulators that work better and faster than those his team designs and trains, usually by hand. The new emulators could be used to improve the models they mimic and help scientists make the best of their time at experimental facilities. If the work stands up to peer review, Lucas says, It would change things in a big way.

A typical computer simulation might calculate, at each time step, how physical forces affect atoms, clouds, galaxieswhatever is being modeled. Emulators, based on a form of AI called machine learning, skip the laborious reproduction of nature. Fed with the inputs and outputs of the full simulation, emulators look for patterns and learn to guess what the simulation would do with new inputs. But creating training data for them requires running the full simulation many timesthe very thing the emulator is meant to avoid.

The new emulators are based on neural networksmachine learning systems inspired by the brains wiringand need far less training. Neural networks consist of simple computing elements that link into circuitries particular for different tasks. Normally the connection strengths evolve through training. But with a technique called neural architecture search, the most data-efficient wiring pattern for a given task can be identified.

The technique, called Deep Emulator Network Search (DENSE), relies on a general neural architecture search co-developed by Melody Guan, a computer scientist at Stanford University. It randomly inserts layers of computation between the networks input and output, and tests and trains the resulting wiring with the limited data. If an added layer enhances performance, its more likely to be included in future variations. Repeating the process improves the emulator. Guan says its exciting to see her work used toward scientific discovery. Muhammad Kasim, a physicist at the University of Oxford who led the study, which was posted on the preprint server arXiv in January, says his team built on Guans work because it balanced accuracy and efficiency.

The researchers used DENSE to develop emulators for 10 simulationsin physics, astronomy, geology, and climate science. One simulation, for example, models the way soot and other atmospheric aerosols reflect and absorb sunlight, affecting the global climate. It can take a thousand of computer-hours to run, so Duncan Watson-Parris, an atmospheric physicist at Oxford and study co-author, sometimes uses a machine learning emulator. But, he says, its tricky to set up, and it cant produce high-resolution outputs, no matter how many data you give it.

The emulators that DENSE created, in contrast, excelled despite the lack of data. When they were turbocharged with specialized graphical processing chips, they were between about 100,000 and 2 billion times faster than their simulations. That speedup isnt unusual for an emulator, but these were highly accurate: In one comparison, an astronomy emulators results were more than 99.9% identical to the results of the full simulation, and across the 10 simulations the neural network emulators were far better than conventional ones. Kasim says he thought DENSE would need tens of thousands of training examples per simulation to achieve these levels of accuracy. In most cases, it used a few thousand, and in the aerosol case only a few dozen.

Its a really cool result, said Laurence Perreault-Levasseur, an astrophysicist at the University of Montreal who simulates galaxies whose light has been lensed by the gravity of other galaxies. Its very impressive that this same methodology can be applied for these different problems, and that they can manage to train it with so few examples.

Lucas says the DENSE emulators, on top of being fast and accurate, have another powerful application. They can solve inverse problemsusing the emulator to identify the best model parameters for correctly predicting outputs. These parameters could then be used to improve full simulations.

Kasim says DENSE could even enable researchers to interpret data on the fly. His team studies the behavior of plasma pushed to extreme conditions by a giant x-ray laser at Stanford, where time is precious. Analyzing their data in real timemodeling, for instance, a plasmas temperature and densityis impossible, because the needed simulations can take days to run, longer than the time the researchers have on the laser. But a DENSE emulator could interpret the data fast enough to modify the experiment, he says. Hopefully in the future we can do on-the-spot analysis.

Read the rest here:

From models of galaxies to atoms, simple AI shortcuts speed up simulations by billions of times - Science Magazine

In a world first, a medicine developed by artificial intelligence may be used to treat patients with obsessive-compulsive disorder. The news is remarkable and hints that in the future, AI may help drug development become faster and more efficiently than ever before.

The first non-man made drug molecule, DSP-1181, has now entered Phase 1 clinical trials, European Pharmaceutical Review reported. The molecule is a long-acting potent serotonin 5-HT1A receptor agonist and was developed using AI that was the product of a partnership between Japans Sumitomo Dainippon Pharma and Exscientia in the UK. The compound was developed in a remarkable time, with AI able to complete in 12 months what typically takes five years.

We are very excited with the results of the joint research that resulted in the development of candidate compounds in a very short time, said Toru Kimura, Senior Executive Officer and Senior Executive Research Director of Sumitomo Dainippon Pharma. We will continue to work hard to make this clinical study success so that it may deliver new benefits to patients as soon as possible.

The drug was created by using algorithms, which AI was able to sift through faster than any human could.

Binary code symbols are seen on a laptop screen in this photo illustration on October 15, 2018 in ... [+] Warsaw, Poland. (Photo by Jaap Arriens/NurPhoto via Getty Images)

"There are billions of decisions needed to find the right molecules and it is a huge decision to precisely engineer a drug," Exscienta chief executive Prof Andrew Hopkins told the BBC. "But the beauty of the algorithm is that they are agnostic, so can be applied to any disease," he added.

This is not the first time that AI has played a hand in medicine, as increasing research shows AIs ability to accurately diagnose disease, sometimes even better than doctors. For example, in the case of breast cancer diagnoses, a publication in the journal Nature showed that a computer model using an algorithm was more successful than radiologists reading mammograms, the BBC reported. The study concluded that the AI was as good as the current system, which uses two doctors to reach a single mammogram, but better than a single doctor. The AI also had fewer false negatives than the doctors, which is important as undiagnosed cancer is serious.

In the case of developing drugs, DSP-1181 may have been the first but scientists are convinced it is far from the last.

"This year was the first to have an AI-designed drug but by the end of the decade all new drugs could potentially be created by AI," said Hopkins, BBC reported.

See more here:

For The First Time Ever, A Drug Developed By AI Will Be Tested In Human Trials - Forbes

Gamers could soon be guiding military robots afterresearchers won a grant to investigate what they can teach an AI about controlling swarms of drones.

Scientists from the University of Buffalowillstudy the decisions, brainwaves and eye movements of people playing video games. They will then use this data to build an AI that can controlautonomous air and ground robots

The participants will play a real-time strategy game developed by the research team thats comparable to the likes of Starcraft and Company of Heroes,UBNow revealed.

[Read: How DeepMinds AI defeated top players at StarCraft II]

While they play, the researchers will record their decisions, track their eye movements with high-speed cameras, and monitor their brain wave patterns through electroencephalography (EEG)headsets.

The data they extract will be used to create algorithms that guideswarms of up to 250 military drones.

We dont want the AI system just to mimic human behavior; we want it to form a deeper understanding of what motivates human actions. Thats what will lead to more advanced AI, principal investigatorChowdhury told UBNow.

The study will be funded by the Defense Advanced Research Projects Agency (DARPA), which recentlypublished a video showing how itsdrone swarms could conduct anurban raid.

StarCraft experts left unimpressed by DARPAs efforts now have a chance to demonstrate that their own tactics are superior.

Youre here because you want to learn more about artificial intelligence. So do we. So this summer, were bringing Neural to TNW Conference 2020, where we will host a vibrant program dedicated exclusively to AI. With keynotes by experts from companies like Spotify, RSA, and Medium, our Neural track will take a deep dive into new innovations, ethical problems, and how AI can transform businesses. Get your early bird ticket and check out the full Neural track.

Published February 12, 2020 18:24 UTC

Read more:

Gamers will teach AI how to control military drone swarms - The Next Web

In a building across from its main office in Seattle, the Allen Institute for Artificial Intelligence (AI2) has enough Ikea furniture to configure 14 different apartments. The lab isnt going into interior designnot exactly. The resources are meant to train smarter algorithms for controlling robots.

Household robots like the Roomba function well only because their tasks are relatively simple. Meandering around, doubling back, and returning to the same spots over and over dont really matter when the objective is to relentlessly clean the same floor.

But anything that requires more efficient or complex navigation still trips up many state-of-the-art robots. The research needed to improve this status quo is also expensivelimiting most cutting-edge progress to well-funded commercial labs.

Sign up for The Algorithm artificial intelligence, demystified

Now AI2 wants to kill two birds with one stone. On Tuesday, it announced a new challenge called RoboTHOR (THOR for The House Of inteRactionsyes, really). It will double as a way to crowdsource better navigation algorithms and lower the financial barriers for researchers who may not have robotics resources of their own.

The ultimate goal is to more rapidly advance AI by getting more research groups involved. Different communities should bring different perspectives and use cases that will expand the repertoire of robot capabilities, driving the field closer to more generalizable intelligence.

AI2

The lab has designed an easily reconfigurable room, the size of a cramped studio, to be the staging ground for all 14 apartment variations. It has also re-created identical virtual replicas in Unity, a popular video-game engineas well as 75 other configurationsthat have all been open-sourced online. Together, these 89 total configurations will offer realistic simulation environments for teams around the world to train and test their navigation algorithms. The environments also come pre-loaded with models of AI2s robots and mirror real-world physics like gravity and light reflections as closely as possible.

The challenge specifically asks teams to develop algorithms that can get a robot from a random starting location within a room to an object in that room just by telling it the objects name. This will be more difficult than simple navigation because it will require the robot to understand the command and recognize the object in its visual field as well.

AI2

Teams will compete in three phases. In phase one, they will be given the 75 digital-only simulation environments to train and validate their algorithms. In phase two, the highest performers will then be given four new simulation environments with corresponding physical doppelgangers. The teams will be able to remotely refine their algorithms by loading them into AI2s real robots.

In the final phase, the highest performers will need to demonstrate the generalizability of their algorithms in the last 10 digital and corresponding physical apartments. Whichever teams perform the best in this final phase will win bragging rights and an invitation to demo their models at the Conference on Computer Vision and Pattern Recognition, a leading AI research conference for vision-based systems.

AI2

After the challenge is over, AI2 plans to keep the setup available, giving anyone access to the environment to continue conducting robotics research. Researchers who clear a certain threshold of accuracy in the simulated environmentsproving they wont crash the robotswill be allowed to remotely deploy their algorithms in the physical ones. The room will rotate between the different furniture configurations.

We are going to maintain this environment, and we are going to maintain these robots, says Ani Kembhavi, a research scientist at AI2 who is leading the project. His team plans to develop a time-sharing system to allow different researchers to take turns remotely testing their algorithms in the real world.

AI2 hopes the strategy will make robotics research more accessible by eliminating as much of the associated hardware costs as possible. It also hopes that the scheme will inspire other well-funded organizations to open up their resources in similar ways. Additionally, it purposely designed its reconfigurable room with low materials costs and globally available Ikea furniture; the setup cost roughly $10,000. Should other researchers want to build their own physical training spaces, they can easily replicate it locally and still match the virtual simulation environments.

Kembhavi, whose dad is an astronomer, likens the idea to the global sharing of telescopes. Communities like astronomy have figured out how to take expensive resources and make it available to researchers all around the world, he says.

That's our vision for this environment, he adds. Embodied AI for all.

Link:

An ever-changing room of Ikea furniture could help AI navigate the world - MIT Technology Review

With countless predictions about whats in store for artificial intelligence in 2020, Im eager to see what will come true and what will fall by the wayside. I think that one of the more paradigm-changing predictions will be moving AIs learning ability to the edge.

Under the hood of AIs generic name, a variety of approaches are hidden, spanning from huge models that crunch data on a distributed cloud infrastructure to tiny, edge-friendly AI that analyze and mine data on small processors.

From my academic research at Boston University to cofounding Neurala, I have always been keenly aware of the difference between these two types of AIlets call them "heavy" and "light" AI. Heavy AI requires hefty compute substrates to run, while light AI can do what heavy AI is capable of but on smaller compute power.

The introduction of commodity processors such as GPUsand later their portabilityhas made it technically and economically viable to bring AI/deep learning/DNN/neural network algorithms to the edge in a multitude of industries.

Bandwidth, latency, cost and just plain logic dictate the era of edge AI and will help make our next technology jump a reality. But before we can do so, it is important to understand the technologys nuances, because making AI algorithms run on the small compute edge has a few. In fact, there are at least two processes at play: inference, or "predictions" generated by an edge (e.g., I see a normal frame vs. one with a possible defect), and edge learningnamely, using the acquired information to change, improve, correct and refine the edge AI. This is a small, often overlooked difference with huge implications.

Living At The Edge

I first realized this difference between inference/predictions and edge learning while working with NASA back in 2010. My colleagues and I implemented a small brain emulation to control a Mars Rover-like device with AI that needed to be capable of running and learning at the edge.

For NASA, it was important that a robot be capable of learning "new things" completely independently of any compute power available on Earth. A data bottleneck, latency and a plethora of other issues meant they needed to explore different breeds of AI than what had been developed at that time. They needed algorithms that had the ability to digest and learnnamely, adapt the behavior of AI and the available datawithout requiring huge amounts of compute power, data and time.

Unfortunately, traditional deep neural network (DNN) models were just not up to par, so we went on to build our own AI that would meet these requirements. Dubbed "lifelong deep neural network" (Lifelong-DNN), this new approach to DNNs had the ability to learn throughout its lifetime (versus traditional DNNs that can only learn once, before deployment).

Learn At The Edge Or Die

One of the biggest challenges when it comes to the implementation of AI today is its inflexibility and lack of adaptability. AI algorithms can be trained on huge amounts of data, when available, and can be fairly robust if all data is captured for their training beforehand. But unfortunately, this is not how the world works.

We humans are so adaptable because our brains have figured out that lifelong learning (learning every day) is key, and we cant rely solely on the data we are born with. Thats why we do not stop learning after our first birthday: We continuously adapt to changing environments and scenarios we encounter throughout our lives and learn from them. As humans, we do not discard data, we use it constantly to fine-tune our own AI.

Humans are a primary example of edge learning-enabled machines. In fact, if human brains acted in the same way as a DNN, our knowledge would be restricted to our college years. We would go about our 9-to-5s and daily routines only to wake up the next morning without having learned anything new.

The Learning-Enabled, AI-Powered Edge

Traditional DNNs are the dominant paradigm in todays AI, with fixed models that need to be trained before deployment. But novel approaches such as Lifelong-DNN would enable AI-powered compute edges not only to understand the data coming to them but also adapt and learn.

So, if you too would like to harness the power of the edge, here is my advice. First off, you need to abandon the mindset (and restriction) that AI can only be trained before deployment. From there, a new need arises: a way for users to interact with the edge and add knowledge. This implies the need to visualize newly collected data and for the user to be able to select which ones to add. This can be done either manually by a user or automatically.

For instance, in a manufacturing scenario, a quality control specialist may reject a product coming out of a machine and, by doing so, provide AI a new clue that the product or the part of it that was just built has to be considered faulty. So, updating your AI training protocols to allow for the integration of continual training workflows, where AI is updated based on new clues, is a must for organizations and individuals looking to leverage this new breed of AI.

AI that learns at the edge is a paradigm-shifting technology that will finally empower AI to truly serve its purpose: shifting intelligence to the compute edge where it is needed, at speeds, latency and costs that make it affordable for every device.

Going forward, learning-enabled edges will survive natural selection in an increasingly competitive AI ecosystem. May the fittest AI survive!

Here is the original post:

2020 And The Dawn Of AI Learning At The Edge - Forbes

As a psychologist, Tom Gruber is in awe of Facebook. As a computer scientist and citizen of the earth, it scares the crap out of him.

Facebook runs experiments on human behavior that psychologists can only dream about, Gruber says. The trials are done on millions of people, a sample size that's impossible in academia. Dozens of times a day, Mark Zuckerberg tweaks his artificial intelligence to see what will keep his 2.5 billion subscribers scrolling through Facebook, and to make them confuse advertising with news so they click on the ads, Gruber says.

"They have the world's largest psychology experiment at their disposal every single day," Gruber says. "They can do experiments that science can't do, at scale."

Gruber, who speaks at TechfestNW this April, is hardly a bomb-thrower. He is a pioneer in artificial intelligence and the co-inventor of Siri, the digital assistant on the iPhone that uses AI and speech recognition to answer billions of questions each year.

Since selling Siri to Apple in 2010, though, Gruber has become one of a small group of technologists who have grown wary of the AI they helped create. He plans to talk about the dangerand promiseof artificial intelligence at TechfestNW.

Facebook and YouTube have more than 2 billion users each, making them as big as the world's two biggest religions, Christianity and Islam, Gruber says.

"And I would add that even the people who pray to Mecca five times a day, only do it five times a day," Gruber says. "Our millennials check their phones 150 times a day."

Gruber has deep roots in techdom. He earned a bachelor's degree in computer science and psychology from Loyola University in New Orleans, got his Ph.D. in computer and information science from the University of Massachusetts, then did research at Stanford University for five years.

Siri grew out of a Stanford spinoff called SRI International. Gruber consulted at SRI in 2007, and, soon after, he and two others, Dag Kittlaus and Adam Cheyer, spun off newer digital-assistant technology that went beyond the DARPA work. They named the new company Siri, which means "beautiful woman who leads you to victory."

Siri is actually a collection of powerful neural networks: mathematical formulas running on computers that analyze huge amounts of data and learn the patterns within them. Turn a neural net loose on a million samples of spoken language, and it will start to recognize words and their meaning. No longer do programmers have to tell computers what to do, logic step by logic step.

Steve Jobs persuaded Gruber and his partners to sell to Apple in 2010 for some $200 million, according to Wired magazine.

Gruber retired from Apple in 2018 and founded Humanistic AI, a firm that helps companies use machine intelligence to collaborate with humans, not replaceor terrorizethem.

Unlike some AI doomsayers, including Tesla inventor Elon Musk and podcasting neuroscientist Sam Harris, Gruber thinks AI can be tamed. Right now, it's a science experiment gone wrong. Frankenstein never meant for his monster to become a killer, and Zuckerberg, he says, never intended Facebook to set us at each other's throats, over politics or anything else.

"My argument is that this is an unintended consequence," Gruber says. "We'll give them a pass on being evil geniuses. Maybe some of them are. But let's assume good intentions."

When it comes to Zuckerberg, assuming good intentions is controversial. In July, Facebook agreed to pay a record $5 billion fine to settle charges by the Federal Trade Commission that it abused users' personal information.

So call Gruber an optimist. He thinks the same algorithms that prey on our bad habits can be used to encourage good ones.

Tech companies make excuses for why they can't police their networks, and most involve money. So far, humans are better at sorting lies from truth, and hate from news. That means you have to hire a lot of humans, which is anathema to the tech monopolies. Gruber says they need to suck it up.

"It's like when the auto industry said, 'Air bags are going to put us out of business, so don't impose this onerous thing on us,'" Gruber says. "It's all bullshit."

And there's more. Why not run all these vast experiments on human behavior to improve human life, instead of wrecking it? Why not use AI to change the habits that lead to type 2 diabetes, heart disease, hypertension, and suicide?

"We have weak theories about what makes people tick, and what to do to help them do better things," Gruber says. "But AI has shown that if you want to get 2 billion people addicted to something that's not good for them, you can do it."

AI doesn't know if it's operating for good or evil, Gruber says. Someday it may, but for now, it's up to humans to direct it.

So far, we've been crappy shepherds.

GO: TechfestNW is at Portland State University's Viking Pavilion, 930 SW Hall St., techfestnw.com. Thursday-Friday, April 2-3. Visit the website for tickets.

View original post here:

The Father of Siri Has Grown Wary of the Artificial Intelligence He Helped Create - Willamette Week