Artificial intelligence is rapidly changing many aspects of how we work and live. (How many stories did you read last week about self-driving cars and job-stealing robots? Perhaps your holiday shopping involved some AI algorithms, as well.) But despite the constant flow of news, many misconceptions about AI remain.
AI doesn’t think in our sense of the word at all, Scriffignano explains. “In many ways, it’s not really intelligence. It’s regressive.”
IT leaders should make deliberate choices about what AI can and can’t do on its own. “You have to pay attention to giving AI autonomy intentionally and not by accident,”
5,209 Comments
Tomi Engdahl says:
Angela Chen / The Verge:
Voice-enabled tech has given rise to voice analysis research that provides insight into human behaviors, but raises concerns about privacy and accuracy
Why companies want to mine the secrets in your voice
https://www.theverge.com/2019/3/14/18264458/voice-technology-speech-analysis-mental-health-risk-privacy
Voices are highly personal, hard to fake, and contain surprising information about our mental health and behaviors.
Tomi Engdahl says:
Kamen Aims to Deliver AI to FedEx
Wheelchair for vets gives delivery robots a boost
https://www.eetimes.com/document.asp?doc_id=1334442
Tomi Engdahl says:
Why I joined MLPerf
https://www.eetimes.com/author.asp?section_id=36&doc_id=1334441
My involvement in MLPerf is the culmination of a long-standing interest in machine learning and my involvement in the world of computer architecture. My curiosity about machine learning dates back to my years studying mathematics and economics at the University of Chicago.
To pick two examples, Greg Diamos was one of the leaders on Baidu’s DeepSpeech and had kicked off DeepBench, a spiritual predecessor to MLPerf. I had spent a fair bit of time studying DeepBench and my discussions with Greg were always enjoyable and enlightening.
Jonah Alben has led the development for several generations of Nvidia GPUs.
Jonah Alben has led the development for several generations of Nvidia GPUs. The chips represent a tremendously successful example of hardware/software co-design given, for instance, their recent hardware acceleration for ray-tracing. Jonah has insightful perspectives that are often informed by customers and cutting-edge researchers and a keen eye for commercial relevance.
MLPerf also included key contributors from Google’s TPU team, several of Intel’s products groups including the former Nervana team, Microsoft Azure, and many others. Simply put, I saw that MLPerf was bringing together some of the people who were going to build the future of machine learning. I could tell that working with this group would accelerate my own learning process and build invaluable connections.
Tomi Engdahl says:
Android offline speech recognition natively on PC
https://hackaday.io/project/164399-android-offline-speech-recognition-natively-on-pc
Porting the Android on-device speech recognition found in GBoard to TensorFlow Lite or LWTNN
Tomi Engdahl says:
https://www.uusiteknologia.fi/2019/03/20/teollisuuskooste-hakee-nostetta-tekoalysta/
Tomi Engdahl says:
https://www.uusiteknologia.fi/2019/03/20/tekoaly-kiinnostaa-mutta-koetaan-yrityksissa-viela-etaiseksi/
Tomi Engdahl says:
Anand Murali / FactorDaily:
With the global market for AI data preparation expected to reach $1.2B by the end of 2023, India is emerging as a hub for data labelling and annotation work
Tomi Engdahl says:
Anand Murali / FactorDaily:
With the global market for AI data preparation expected to reach $1.2B by the end of 2023, India is emerging as a hub for data labelling and annotation work
How India’s data labellers are powering the global AI race
https://factordaily.com/indian-data-labellers-powering-the-global-ai-race/
Thousands of staffers at iMerit’s other offices in Kolkata, Ranchi, Bhubaneswar, Vizag and Shillong do similar work, labelling millions of data to help train and power AI algorithms developed by companies across the globe.
With global enterprise giants embracing AI, and the datasets that feed the AI algorithms increasingly becoming proprietary, companies need a higher degree of engagement with data labelling teams in terms of requirements, quality control, feedback and deliverables.
Because of the business process outsourcing boom around the turn of the century, Indians are no strangers to such jargon and demands. Data annotation and labelling, too, is process-driven, requiring precision work and skills that even people with a high-school education can be trained on.
As the first generation of such work that was mainly crowdsourced gave way to more advanced requirements, companies such as Infolks, iMerit and Playment have come up catering to global clients and making India an emerging hub for data labelling and annotation work.
“This is an emerging sector… in India and everybody has begun to realise the humongous opportunity it presents,”
What is data labelling?
Data labelling and annotation is a process by which datasets — from unstructured sources such as cameras, sensors, emails and social media among others, as well as from structured sources such as databases — are labelled, marked, coloured or highlighted to mark up differences, similarities or types. This is so that when the data are fed into an algorithm for training an AI system, the algorithm can rightfully identify the data and learn from it.
Say you want to train an algorithm to understand road signs using images captured by a camera onboard a vehicle. Data annotators or labellers will go through the dataset of images and mark or highlight road signs using annotation tools and feed this to an AI algorithm to learn from. The next time the algorithm encounters a road sign during a live drive through an area, it should be able to recognize the sign. The more images of road signs the algorithm is trained on, the better its accuracy.
Driving the surge in AI or machine-learning is the access to plentiful data made available from the internet, social media, sensors and other sources. Algorithms today have the ability to absorb more data and, hence, be more accurate. As long as the data is good and clean, feeding another million datasets to an algorithm will inch up its accuracy. This has caused an unending hunger for well-annotated and labelled data for AI algorithms and applications.
Today, data preparation and engineering tasks account for more than 80% of the time involved in most AI and machine-learning projects, according to the Cognilytica report.
Tomi Engdahl says:
China Sees U.S. Ahead in AI
Recent stories from EET-China portray an AI underdog
https://www.eetimes.com/document.asp?doc_id=1334424
While some engineers — and politicians — in the U.S. express concerns that China is getting ahead of them in AI, engineers in China appear to have similar concerns about their U.S. counterparts.
A look at the most popular AI stories in our sister publication EE Times-China in the period from Nov. 25, 2018, to Feb. 25, 2019, shows that six out of 10 articles had a similar thread — China may be behind the U.S. in AI.
Tomi Engdahl says:
10 Views of Nvidia’s GTC 2019
Intel’s Nervana on the radar, but not an immediate threat
https://www.eetimes.com/document.asp?doc_id=1334462
The show floor at Nvidia’s annual Graphics Technology Conference didn’t need a new chip to power lots of energy. It was running at a high frequency on existing chips in robots for cities, farms and factories, expanding virtual and augmented headsets, lots of training servers and several semi-autonomous cars and trucks.
A new chip from a big rival made an appearance at a nearby Intel cocktail reception. The 400W Nervana training accelerator came dressed in the new PCIe mezzanine carrier unveiled at last week’s Open Compute Summit.
Tomi Engdahl says:
Artificial Intelligence Tutorial | AI Tutorial for Beginners | Artificial Intelligence | Simplilearn
https://www.youtube.com/watch?v=FWOZmmIUqHg
This Artificial Intelligence tutorial video will help you understand what is Artificial Intelligence, types of Artificial Intelligence, ways of achieving Artificial Intelligence and applications of Artificial Intelligence. In the end, we will also implement a use case on TensorFlow in which we will predict whether a person has diabetes or not.
Tomi Engdahl says:
https://truththeory.com/2018/11/22/ai-generates-hilarious-inspirational-quotes/
Tomi Engdahl says:
AI Has Beaten Humans at Lip-reading
https://www.technologyreview.com/s/602949/ai-has-beaten-humans-at-lip-reading/?utm_campaign=owned_social&utm_source=facebook.com&utm_medium=social
A pair of new studies show that a machine can understand what you’re saying without hearing a sound.
Tomi Engdahl says:
Facebook’s AI couldn’t spot mass murder
https://techcrunch.com/2019/03/21/facebooks-ai-couldnt-spot-mass-murder/
Tomi Engdahl says:
Morphin instantly Deepfakes your face into GIFs
https://techcrunch.com/2019/03/20/morphin-avatars/
Want to star in your favorite memes and movie scenes? Upload a selfie to Morphin, choose your favorite GIF and your face is grafted in to create a personalized copy you can share anywhere.
Morphin started three years ago with the intention to build personalized avatars for games and VR so you could be a FIFA soccer player or Skyrim knight.
Morphin’s tech no longer requires 3D scanning hardware and it works with just a regular selfie. You just snap a headshot, select a GIF from its iOS or Android app’s library and a few seconds later you have a CGI version of yourself in the scene (with no watermark) that you can export and post. “We wanted it to be super straightforward because we wanted people to relate to the content,”
Tomi Engdahl says:
Fear of AI takeover: Why are we afraid of machines?
https://www.getjenny.com/blog/fear-of-ai-takeover-why-are-we-afraid-of-machines?utm_campaign=Blog%20posts&utm_source=facebook&utm_medium=paidsocial&utm_content=%20Blog%20post:%20Fear%20of%20AI%20takeover
The robots are taking over – dystopian science fiction or distant reality?
From taking our jobs to controlling the human race, there’s definitely a fear surrounding AI and its applications today. Some scientists think that the ‘singularity’ – the moment when AI becomes self-conscious – may happen in this generation. However, the majority of them are less convinced.
Tomi Engdahl says:
Make deep learning faster and simpler
https://phys.org/news/2019-03-deep-faster-simpler.amp
Researchers are using automatic differentiation and other techniques to make deep learning faster and simpler.
Artificial intelligence systems based on deep learning are changing the electronic devices that surround us.
Tomi Engdahl says:
Kuinka välttää koneoppimisen sudenkuopat
by Rami Juhela
https://bluugo.fi/blog/kuinka-valttaa-koneoppimisen-sudenkuopat/
Koneoppiminen tarjoaa yrityksille monia innovatiivisia keinoja tehostaa liiketoimintansa prosesseja, mutta projektien toteuttaminen osoittautuu usein odotettua vaikeammaksi. Lue mitkä ovat koneoppimisprojektien yleisimmät sudenkuopat ja miten ne voidaan välttää.
Tomi Engdahl says:
Nvidia AI turns sketches into photorealistic landscapes in seconds
https://techcrunch.com/2019/03/18/nvidia-ai-turns-sketches-into-photorealistic-landscapes-in-seconds/
Today at Nvidia GTC 2019, the company unveiled a stunning image creator. Using generative adversarial networks, users of the software are with just a few clicks able to sketch images that are nearly photorealistic. The software will instantly turn a couple of lines into a gorgeous mountaintop sunset. This is MS Paint for the AI age.
Called GauGAN, the software is just a demonstration of what’s possible with Nvidia’s neural network platforms.
Tomi Engdahl says:
AI as a Competitive Advantage in Telecom Industry
https://medium.com/mmt-business-publishing/ai-as-a-competitive-advantage-in-telecom-industry-b4b3577e8f1f?sk=937576792313318aa86278a1f88a5fbb
Will Artificial Intelligence be the catalyst to reinvent Telecom companies or will it merely improve operations? One is certain — companies in Telecom Industry are well positioned to use AI, but past technological developments lead to only minor dents in their business models. Why this time might be different?
Tomi Engdahl says:
Untold History of AI: Charles Babbage and the Turk
https://spectrum.ieee.org/tech-talk/tech-history/dawn-of-electronics/untold-history-of-ai-charles-babbage-and-the-turk
Tomi Engdahl says:
Your Dog Can Talk to You with the Help of This Machine Learning Device
https://blog.hackster.io/your-dog-can-talk-to-you-with-the-help-of-this-machine-learning-device-a321d717bd61
Inspired by the Disney film Up, the device relies on two relatively new technologies: EEG (electroencephalogram) brain wave detection, and machine learning.
Tomi Engdahl says:
Nvidia’s Jetson Nano Is an AI Computer for the Masses
BY TOM BRANT 19 MAR 2019, MIDNIGHT
https://uk.pcmag.com/news/120106/nvidias-jetson-nano-is-an-ai-computer-for-the-masses
A new $99 single-board computer from the GPU giant is aimed at amateur programmers interested in building basic artificial intelligence algorithms.
Tomi Engdahl says:
Robots Will Help Spectators at Tokyo 2020 Olympics
https://spectrum.ieee.org/automaton/robotics/medical-robots/tokyo-2020-olympics-helper-robots
Tomi Engdahl says:
https://spectrum.ieee.org/tech-talk/computing/hardware/algorithms-and-hardware-for-deep-learning
Tomi Engdahl says:
DNA Computer Shows Programmable Chemical Machines Are Possible
https://spectrum.ieee.org/tech-talk/computing/hardware/dna-computer-shows-programmable-chemical-machines-are-possible
Tomi Engdahl says:
Muscle-Controlling Mini Brains Grown In The Lab For The First Time
https://www.iflscience.com/brain/musclecontrolling-mini-brains-grown-in-the-lab-for-the-first-time/
Tomi Engdahl says:
Your Dog Can Talk to You with the Help of This Machine Learning Device
https://blog.hackster.io/your-dog-can-talk-to-you-with-the-help-of-this-machine-learning-device-a321d717bd61
Unfortunately, dogs lack the vocal cord development necessary to speak like a parrot, and the paw dexterity to use sign language like a chimpanzee. But, that doesn’t mean they aren’t thinking; they just can’t express those thoughts to us clearly. That may change soon, thanks to a device created by University of Illinois students that uses machine learning to give dogs a voice.
https://youtu.be/UpLDRoqXFys
Tomi Engdahl says:
Scientists Can Read Your Mind Before You Know You’ve Made A Decision
https://www.iflscience.com/brain/scientists-can-read-your-mind-before-you-know-youve-made-a-decision/
It’s not quite Minority Report, but scientists can sometimes see your decisions before you know them yourself. When participants in a study were asked to choose between two patterns, the scientists running the test used images on an fMRI machine to foresee which they would choose. The researchers consider this evidence that our decisions are primed by an unconscious “stand-by” mode.
Tomi Engdahl says:
AI Code Wags Hardware — Vigorously
Pioneer says self-supervising systems will be the next big thing
https://www.eetimes.com/document.asp?doc_id=1334443
In AI, hardware is the tail and software is the dog — and this is a very active dog. One need only browse the popular arXiv.org site to find one- or two-dozen new research papers posted daily.
Wei Li, who leads a software group at Intel devoted to machine learning, rattles off a list of a dozen popular convolutional, recurrent, and other neural-network models. Adding another layer, most big cloud and chip vendors have created their own frameworks to build and run the models optimally on their platforms.
“There’s a variety of topologies and frameworks to test,” he said.
Don’t let the complexity overwhelm you, said Chris Rowen, chief executive of BabbleLabs, a startup creating DNN engines for audio tasks. “The structure of a neural net can be important to efficiency, but any of them can get the job done,” he said. “In many cases, it’s almost a question of style.”
Automated learning is perhaps the most powerful megatrend that will drive change in the software. It could take decades to evolve into what is still considered a kind of science fiction — machines that can learn independently of humans. Meanwhile, researchers are helping today’s neural nets take baby steps in that direction.
Tomi Engdahl says:
AI Trolls for Data Center Woes
HPE monitors hard drives with neural networks
https://www.eetimes.com/document.asp?doc_id=1334423
Hewlett-Packard Enterprise is using neural networks to predict failures on some of the 4 million hard disk drives that its InfoSight service monitors. The project taught HPE that using neural networks takes time, specialized expertise, and some big iron.
Tomi Engdahl says:
AI Silicon Sprouts in the Dark
Engineers are deep in a jungle of architecture options
https://www.eetimes.com/document.asp?doc_id=1334445
The potential for new architectures to accelerate deep learning is enormous. So far, only one novel chip has been fully described and benchmarked — Google’s TPU — but the pipeline is full and a few of the techniques are becoming clear.
The jungle is dense with possibilities. They include analog computing, a variety of emerging memory and packaging types, and a basket of techniques specific to handling neural networks such as pruning and quantization.
“It’s wide open with people working at every level,” said Marian Verhelst, a professor at KU Leuven in Belgium who worked on research chips exploring binary precision formats. Analog computing looks useful, especially for 3- to 8-bit formats, she said.
Tomi Engdahl says:
http://www.etn.fi/index.php/13-news/9264-nvidialta-maailman-pienin-tekoalykortti
Tomi Engdahl says:
http://www.etn.fi/index.php/13-news/9270-analoginen-tietokone-ratkoo-yhtaloita-valolla
Tomi Engdahl says:
Anthony Ha / TechCrunch:
McDonald’s is acquiring Dynamic Yield, which uses machine learning to personalize online shopping such as recommending products; source says deal worth $300M+
McDonald’s is acquiring Dynamic Yield to create a more customized drive-thru
https://techcrunch.com/2019/03/25/mcdonalds-acquires-dynamic-yield/
McDonald’s is announcing an agreement to acquire personalization company Dynamic Yield.
The announcement does not include a price, but a source with knowledge of the deal said that it’s more than $300 million. This is the fast food chain’s largest acquisition in 20 years.
Dynamic Yield works with brands across e-commerce, travel, finance and media to create what’s been described as an Amazon-style personalized online experience.
Tomi Engdahl says:
New York Times:
A look at Robotics at Google, the company’s latest robotics program, which brings machine learning to robots simpler than Boston Dynamics’ machines — In 2013, the company started an ambitious, flashy effort to create robots. Now, its goals are more modest, but the technology is subtly more advanced.
Inside Google’s Rebooted Robotics Program
https://www.nytimes.com/2019/03/26/technology/google-robotics-lab.html
Tomi Engdahl says:
Cade Metz / New York Times:
Three pioneers of AI, Geoffrey Hinton, Yann LeCun, and Yoshua Bengio win the $1M Turing Award for their work on neural networks
Three Pioneers in Artificial Intelligence Win Turing Award
https://www.nytimes.com/2019/03/27/technology/turing-award-hinton-lecun-bengio.html
On Wednesday, the Association for Computing Machinery, the world’s largest society of computing professionals, announced that Drs. Hinton, LeCun and Bengio had won this year’s Turing Award for their work on neural networks. The Turing Award, which was introduced in 1966, is often called the Nobel Prize of computing, and it includes a $1 million prize, which the three scientists will share.
Tomi Engdahl says:
The Keyword:
Google announces an Advanced Technology External Advisory Council to consider ethical issues around AI like facial recognition and fairness in machine learning
An external advisory council to help advance the responsible development of AI
https://www.blog.google/technology/ai/external-advisory-council-help-advance-responsible-development-ai/
Last June we announced Google’s AI Principles, an ethical charter to guide the responsible development and use of AI in our research and products. To complement the internal governance structure and processes that help us implement the principles, we’ve established an Advanced Technology External Advisory Council (ATEAC). This group will consider some of Google’s most complex challenges that arise under our AI Principles, like facial recognition and fairness in machine learning, providing diverse perspectives to inform our work.
AI at Google: our principles
https://www.blog.google/technology/ai/ai-principles/
Tomi Engdahl says:
Eric Niiler / Wired:
A look at Estonia’s efforts to deploy AI in government services, from farm inspections via satellite image analysis to robot judges for small claims disputes
https://www.wired.com/story/can-ai-be-fair-judge-court-estonia-thinks-so/
Tomi Engdahl says:
Two Major Concerns about the Ethics of Facial Recognition in Public Safety
https://www.sealevel.com/2019/03/05/two-major-concerns-about-the-ethics-of-facial-recognition-in-public-safety/
From catching the right criminal to offering hope in overturning convictions, facial recognition offers public safety and the judicial system an effective way out of bureaucratic processes.
Tomi Engdahl says:
Algorithm developed to improve machine-learning models
https://www.controleng.com/articles/algorithm-developed-to-improve-machine-learning-models/
MIT researchers have developed an algorithm that designs optimized machine-learning models up to 200 times faster than traditional methods.
A new area in artificial intelligence involves using algorithms to automatically design machine-learning systems known as neural networks, which are more accurate and efficient than those developed by human engineers. But this so-called neural architecture search (NAS) technique is computationally expensive.
One of the state-of-the-art NAS algorithms recently developed by Google took 48,000 hours of work by a squad of graphical processing units (GPUs) to produce a single convolutional neural network, used for image classification and identification tasks. Google has the wherewithal to run hundreds of GPUs and other specialized circuits in parallel, but that’s out of reach for many others.
Tomi Engdahl says:
The Automation Of AI
https://semiengineering.com/the-automation-of-ai/
Experts at the Table, part 1: Will the separation of hardware and software for AI cause problems and how will hardware platforms for AI influence algorithm development?
Tomi Engdahl says:
The Value Of A Model
https://semiengineering.com/the-value-of-a-model/
How much would you pay for a model? Recently, the answer to that has been $$$.
Increased talk about the Digital Twin has brought models to the forefront of the discussion. What are the right models for particular applications? What is the correct level of abstraction? Where do the models come from and how are they maintained? How does one value a model?
Tomi Engdahl says:
The Dark Secret at the Heart of AI
https://www.technologyreview.com/s/604087/the-dark-secret-at-the-heart-of-ai/?utm_source=facebook.com&utm_medium=social&utm_campaign=Technology+Review
No one really knows how the most advanced algorithms do what they do. That could be a problem.
Tomi Engdahl says:
Amazon’s AWS Deep Learning Containers simplify AI app development
https://venturebeat.com/2019/03/27/amazons-aws-deep-learning-containers-simplify-ai-app-development/
Tomi Engdahl says:
Untold History of AI: Invisible Women Programmed America’s First Electronic Computer
https://spectrum.ieee.org/tech-talk/tech-history/dawn-of-electronics/untold-history-of-ai-invisible-woman-programmed-americas-first-electronic-computer
Tomi Engdahl says:
NVIDIA Research project uses AI to instantly turn drawings into photorealistic images
https://m.dpreview.com/news/7387722427/nvidia-research-project-uses-ai-to-instantly-turn-drawings-into-photorealistic-images
Tomi Engdahl says:
Dan Bilefsky / New York Times:
A profile of AI pioneer Yoshua Bengio, a co-recipient of the 2018 A.M. Turing Award, who says he supports regulating AI, including a ban on “killer robots” — MONTREAL — Yoshua Bengio is worried that innovations in artificial intelligence that he helped pioneer could lead to a dark future …
He Helped Create A.I. Now, He Worries About ‘Killer Robots.’
https://www.nytimes.com/2019/03/29/world/canada/bengio-artificial-intelligence-ai-turing.html
Tomi Engdahl says:
It’s Still Early Days for AI
Neural networks expand far beyond feline photos
https://www.eetimes.com/document.asp?doc_id=1334425
“We need to get to real AI because most of today’s systems don’t have the common sense of a house cat!” The keynoter’s words drew chuckles from an audience of 3,000 engineers who have seen the demos of systems recognizing photos of felines.
There’s plenty of room for skepticism about AI.
It’s true, deep neural networks (DNNs) are a statistical method — by their very nature inexact. They require large, labeled data sets, something many users lack.
It’s also true that DNNs can be fragile. The pattern-matching technique can return dumb results when the data sets are incomplete and misleading results when they have been corrupted. Even when results are impressive, they are typically inexplicable.
The emerging technique has had its share of publicity, sometimes bordering on hype. The fact remains that DNNs work. Though only a few years old, they already are being applied widely. Facebook alone uses sometimes simple neural nets to perform 3×1014 predictions per day, some of which are run on mobile devices, according to LeCun.
Cadence and Synopsys both have reported projects using them to help engineers design better chips. Intel helped Novartis use them to accelerate drug discovery. Siemens is using them to accelerate processing of medical images, and scientists are using them to speed up reading genomes of cancer patients.
Deep learning is with us to stay as a new form of computing. Its applications space is still being explored. Its underlying models and algorithms are still evolving, and hardware is trying to catch up with it all.
It’s “a fundamental transformation of the computing landscape,” said Chris Rowen, a serial entrepreneur who is “100% all in on deep learning” as CEO of startup BabbleLabs developing models for audio applications.
“I used to write an algorithm and tell a system what to do,” he said. “Now, we have a class of methods not described by an explicit algorithm but a set of examples, and the system figures out the pattern.”
“Just about everything software touches can be influenced by this new method, especially where the data is most ambiguous and noisy — where a conventional programmer could not distinguish relevant from irrelevant bits,” he added.
Tomi Engdahl says:
http://www.etn.fi/index.php/13-news/9293-matlab-yha-syvemmalle-tekoalyyn
https://se.mathworks.com/products/new_products/latest_features.html