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,”
5,336 Comments
Tomi Engdahl says:
The Linux Foundation launches a deep learning foundation
https://techcrunch.com/2018/03/26/the-linux-foundation-launches-a-deep-learning-foundation/
Despite its name, the Linux Foundation has long been about more than just Linux. These days, it’s a foundation that provides support to other open source foundations and projects like Cloud Foundry, the Automotive Grade Linux initiative and the Cloud Native Computing Foundation. Today, the Linux Foundation is adding yet another foundation to its stable: the LF Deep Learning Foundation.
The idea behind the LF Deep Learning Foundation is to “support and sustain open source innovation in artificial intelligence, machine learning, and deep learning while striving to make these critical new technologies available to developers and data scientists everywhere.”
https://www.linuxfoundation.org/projects/deep-learning/
Tomi Engdahl says:
AI Medicine Software Science
New Deep-Learning Software Knows How To Make Desired Organic Molecules
https://science.slashdot.org/story/18/03/28/232206/new-deep-learning-software-knows-how-to-make-desired-organic-molecules?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+Slashdot%2Fslashdot%2Fto+%28%28Title%29Slashdot+%28rdf%29%29
dryriver shares a report from Nature about a neural network-based, deep-learning software that is as good as trained chemists in figuring out what reagents and reactions may lead to the successful creation of a desired organic molecule:
Chemists have a new lab assistant: artificial intelligence. Researchers have developed a “deep learning” computer program that produces blueprints for the sequences of reactions needed to create small organic molecules, such as drug compounds. The pathways that the tool suggests look just as good on paper as those devised by human chemists. The tool, described in Nature on March 28, is not the first software to wield artificial intelligence (AI) instead of human skill and intuition
Need to make a molecule? Ask this AI for instructions
https://www.nature.com/articles/d41586-018-03977-w?error=cookies_not_supported&code=75546f74-67b1-4d92-8301-5aebf96df3f3
Artificial-intelligence tool that has digested nearly every reaction ever performed could transform chemistry.
Tomi Engdahl says:
Algorithms Can’t Tell When They’re Broken–And Neither Can We
https://www.fastcompany.com/40549744/algorithms-cant-tell-when-theyre-broken-and-neither-can-we?utm_campaign=Technology+Review&utm_source=facebook.com&utm_medium=social
When algorithms go haywire (and they do) we often don’t know what caused the problem, or even that the problem exists.
Tomi Engdahl says:
Algorithms also make mistakes because they pick up on features of the environment that are correlated with outcomes, even when there is no causal relationship between them. In the algorithmic world, this is called overfitting. When this happens in a brain, we call it superstition.
https://www.fastcompany.com/40549744/algorithms-cant-tell-when-theyre-broken-and-neither-can-we?utm_campaign=Technology+Review&utm_source=facebook.com&utm_medium=social
Tomi Engdahl says:
Nvidia Taps Memory, Switch for AI
Intel’s Nervana aims to leapfrog Nvidia in 2019
https://www.eetimes.com/document.asp?doc_id=1333124
At its annual GTC event, Nvidia announced system-level enhancements to boost the performance of its GPUs in training neural networks and a partnership with ARM to spread its technology into inference jobs.
Nvidia offered no details of its roadmap, presumably for 7-nm graphics processors in 2019 or later. It has some breathing room, given that AMD is just getting started in this space, Intel is not expected to ship its Nervana accelerator until next year, and Graphcore — a leading startup — has gone quiet. A few months ago, both Intel and Graphcore were expected to release production silicon this year.
The high-end Tesla V100 GPU from Nvidia is now available with 32-GBytes memory, twice the HBM2 stacks of DRAM that it supported when launched last May. In addition, the company announced NVSwitch, a 100-W chip made in a TSMC 12nm FinFET process. It sports 18 NVLink 2.0 ports that can link 16 GPUs to shared memory.
Nvidia became the first company to make the muscular training systems expected to draw 10 kW of power and deliver up to 2 petaflops of performance. Its DGX-2 will pack 12 NVSwitch chips and 16 GPUs in a 10U chassis that can support two Intel Xeon hosts, Infiniband, or Ethernet networks and up to 60 solid-state drives.
Cray, Hewlett Packard Enterprise, IBM, Lenovo, Supermicro, and Tyan said that they will start shipping systems with the 32-GB chips by June. Oracle plans to use the chip in a cloud service later in the year.
Tomi Engdahl says:
Optane™ – The New Smart System Accelerator by Intel
Zip into a responsive computing experience with the power of Optane™.
https://www.arrow.com/en/campaigns/emea-intel-q1-2018-optane
Its speed is just the beginning. As a performance and productivity accelerator, Intel Optane™ uses 3D XPoint™ memory media, Intel Memory and Storage Controllers, Intel Interconnect IP, and Intel software to bridge the gap between storage and the DRAM for an incredibly responsive experience.
Tomi Engdahl says:
AI Has Learned To Make Nude Art But It Might Be The Least Sexy Thing You’ll Ever See
http://www.iflscience.com/technology/ai-has-learned-to-make-nude-art-but-it-might-be-the-least-sexy-thing-youll-ever-see/
Let’s just say that we’re quite glad robots haven’t taken over the world just yet. Because we’re not sure we’re ready for AI art.
That includes these rather odd artificial intelligence-generated nude portraits, which AI researcher Robbie Barrat posted to Twitter. He used what’s called a Generative Adversarial Network (GAN) and fed it thousands of nude portraits, before asking it to make its own. The results are, well, pretty odd.
Tomi Engdahl says:
An AI-powered design trick could help prevent accidents like Uber’s self-driving car crash
https://qz.com/1241119/accidents-like-ubers-self-driving-car-crash-could-be-prevented-with-this-ai-powered-design-trick/
Tomi Engdahl says:
Nvidia wants AI to Get Out of the Cloud and Into a Camera, Drone, or Other Gadget Near You
https://spectrum.ieee.org/view-from-the-valley/computing/embedded-systems/nvidia-wants-ai-to-get-out-of-the-cloud-into-a-camera-drone-or-other-gadget-near-you
People are just now getting comfortable with the idea that data from many electronic gadgets they use flies up to the cloud. But going forward, much of that data will stick closer to Earth, processed in hardware that lives at the so-called edge—for example, inside security cameras or drones.
That’s why Nvidia, the processor company whose graphics processing units (GPUs) are powering much of the boom in deep learning, is now focused on the edge.
Tomi Engdahl says:
Tom Simonite / Wired:
How tech companies are fighting bias in facial recognition software: detecting race, diversifying training data, and disclosing recognition accuracy by group
How Coders Are Fighting Bias in Facial Recognition Software
https://www.wired.com/story/how-coders-are-fighting-bias-in-facial-recognition-software
Tomi Engdahl says:
How deep learning is enhancing machine vision
https://www.vision-systems.com/articles/print/volume-23/issue-1/features/how-deep-learning-is-enhancing-machine-vision.html?cmpid=enl_vsd_vsd_newsletter_2018-04-02&pwhid=6b9badc08db25d04d04ee00b499089ffc280910702f8ef99951bdbdad3175f54dcae8b7ad9fa2c1f5697ffa19d05535df56b8dc1e6f75b7b6f6f8c7461ce0b24&eid=289644432&bid=2052184
Developers increasingly apply deep learning and artificial neural networks to improve object detection and classification.
Johannes Hiltner
Digitalization has a firm grip on industrial production, with processes increasingly automated as part of the Industrial Internet of Things (IIoT). In the IIoT, which is also known as Industry 4.0, various machines and robots take on more everyday production tasks. In assembly for example, new, compact and mobile robots, such as collaborative robots (cobots), often work hand in hand with their human colleagues.
The IIoT’s highly automated and universally networked production flows characterized by machine-to-machine interaction depend on machine vision to reliably identify a wide range of objects in the flow of goods within factories and the rest of the process chain. Machine vision increases the efficiency and safety of these workflows, and has become an indispensable tool for engineers seeking to automate and speed up production.
Analyze and evaluate large data sets
In an effort to make the identification process even more robust and adaptable to the requirements of flexible and networked IIoT processes, machine vision software developers increasingly rely on methods from the field of artificial intelligence (AI). Deep learning is an area of machine learning that enables computers to be trained and learn through architectures such as convolutional neural networks (CNNs).
The special attribute of AI, machine-learning and deep-learning technologies is that they comprehensively analyze and evaluate large amounts of data (big data) in order to train many different classes and thereby more effectively distinguish between objects. Increasingly, this data is generated within the IIoT. This can be digital image information as well as data from sensors, scanners, and other process components.
In order to use deep learning, CNNs must first be trained. This training process relates to certain external features that are typical of the object, such as color, shape, texture, and surface structure. The objects are divided into different classes based on these properties to allocate them more precisely later.
In conventional machine vision methods, a developer must laboriously define and verify the individual features manually. With deep learning, however, self-learning algorithms are used to automatically find and extract the unique patterns in order to differentiate between the particular classes.
Conclusion
Technologies based on artificial intelligence, such as deep learning and CNNs, are an important part of modern machine vision solutions today. Deep learning enables companies to train neural networks themselves without any in-depth expert knowledge and minimal effort, especially when programming defect classes during error inspections. The result is that companies can save money and benefit from much more robust recognition rates as well as better classification results.
Tomi Engdahl says:
The Week In Review: Design
Nvidia deep learning architecture for Arm
https://semiengineering.com/the-week-in-review-design-123/
Nvidia will integrate its open-source NVIDIA Deep Learning Accelerator (NVDLA) architecture into Arm’s Project Trillium platform for machine learning. The NVDLA hardware supports a wide range of IoT devices and is supported by Nvidia’s suite of developer tools.
Tomi Engdahl says:
https://www.marmai.fi/uutiset/tekoalyn-merkitys-kasvaa-markkinoinnissa-suomessa-on-teknologiset-mahdollisuudet-ottaa-ai-sta-kaikki-irti-6710299
Tomi Engdahl says:
Google’s AI Becomes “Aggressive” In Certain
Situations Thanks To A Very Human Problem
http://www.iflscience.com/technology/googles-ai-becomes-aggressive-in-certain-situations-thanks-to-a-very-human-problem/
It’s been reported that an artificial intelligence (AI) has learned to adopt “highly aggressive” strategies when it feels it’s about to lose a simulated game. While that certainly sounds a little scary, the study this is linked to was investigating a larger social problem far more fascinating and enlightening that mere aggression.
Google’s in-house AI, machine learning, and neural networking development teams are working on some truly remarkable projects at present: From AlphaGo Zero, the AI that learned 3,000 years’ worth of Go tactics in a matter of days, to AutoML, a system that makes self-correcting AI “children” that are designed to perform specific tasks, their output is nothing but impressive.
Tomi Engdahl says:
New York Times:
Apple hires John Giannandrea, formerly Google’s chief of search and AI, to run Apple’s machine learning and AI strategy, reporting to Tim Cook — SAN FRANCISCO — Apple has hired Google’s chief of search and artificial intelligence, John Giannandrea, a major coup in its bid to catch …
Apple Hires Google’s A.I. Chief
https://www.nytimes.com/2018/04/03/business/apple-hires-googles-ai-chief.html
Tomi Engdahl says:
How machine-learning code turns a mirror on its sexist, racist masters
Word-analyzing AI study reveals ‘historical social changes’
https://www.theregister.co.uk/2018/04/04/ai_algorithms_discrimination_bias/
Be careful which words you feed into that machine-learning software you’re building, and how.
A study of news articles and books written during the 20th and 21st century has shown that not only are gender and ethnic stereotypes woven into our language, but that algorithms commonly used to train code can end up unexpectedly baking these biases into AI models.
Basically, no one wants to see tomorrow’s software picking up yesterday’s racism and sexism.
A paper published in the Proceedings of the US National Academy of Sciences on Tuesday describes how word embeddings, a common set of techniques used by machine-leaning applications to develop associations between words, can pick up social attitudes towards men and women, and people of different ethnicities, from old articles and novels.
Tomi Engdahl says:
Why This Doctor Is Building an ‘AI Sherlock’ for Medical Pros
https://www.pcmag.com/news/358639/why-this-doctor-is-building-an-ai-sherlock-for-medical-pro
Dr. Anthony C. Chang from the Children’s Hospital of Orange County wants his fellow doctors to think outside the box when it comes to knowledge and expertise. Can an ‘AI Sherlock’ help?
In the medical community, a clinician’s opinion holds sway, so wariness among professionals about artificial intelligence is not surprising.
But Dr. Anthony C. Chang, practicing Pediatric Cardiologist and Chief Intelligence and Innovation Officer at the Children’s Hospital of Orange County (CHOC) saw the writing on the wall several years ago and has embraced the emerging technology. He even went back to school after several decades as a doctor to study biomedical data science and AI at Stanford School of Medicine.
Today, he’s often referred to as “Dr. AI,” speaks regularly at Singularity University Exponential Medicine conferences, and presides over the Medical Intelligence and Innovation Institute (MI3), the first of its kind inside a hospital, funded by the Sharon Disney Lund Foundation.
Tomi Engdahl says:
Nvidia: One Analyst Thinks It’s Decimating Rivals in A.I. Chips
https://www.barrons.com/articles/nvidia-one-analyst-thinks-its-decimating-rivals-in-a-i-chips-1522703007
The fastest-growing part of chip maker Nvidia’s (NVDA) business is its “data center” chips product line, driven in part by sales of graphics chips — “GPUs” — that are widely used for artificial intelligence tasks such as machine learning.
That division looks to have a very bright future, according to one analyst who attended Nvidia’s annual “GTC” conference last week.
“What Nvidia did with their announcements last week was to cause everyone, including Intel (INTC), but also startups, to re-examine their roadmaps,” says Hans Mosesmann of Rosenblatt Securities.
I chatted with Mosesmann by phone on Friday. Mosesmann, who has a Buy rating on Nvidia stock, and a $300 price target, foresees the company having something of a lock on the A.I. chip market.
Move Over Moore’s Law, Make Way for Huang’s Law
https://spectrum.ieee.org/view-from-the-valley/computing/hardware/move-over-moores-law-make-way-for-huangs-law
Graphics processors are on a supercharged development path that eclipses Moore’s Law, says Nvidia’s Jensen Huang
Tomi Engdahl says:
Three Concepts for Managing AI
https://www.eetimes.com/author.asp?section_id=36&doc_id=1333151
Three key ideas should drive how AI is rolled out in the electronics community to reap the full benefits of the new technology.
In the manufacturing and high-tech industry, 77 percent of leaders reported using AI to automate business processes, which is where most organizations initially focus their AI initiatives.
According to the research, 87 percent of organizations in late or final stages of their AI deployments saw significant and measurable benefits from AI technologies. Of those in the later stages of AI deployments, 80 percent of IT decision makers said that they are using AI to augment existing solutions or build new business-critical solutions and services to optimize insights and consumer experience.
Looking at all the organizations that saw ROI ranging from improved process times to new or increased market share, one key component was consistent: Developing a clear AI strategy that tethers to a business strategy is of the utmost importance.
Forty-nine percent of leaders reported that they were focusing on hiring new employees that have different skill sets related to AI. However, in the manufacturing and high-tech sector, 56 percent of executives reported challenges in finding qualified staff to lead AI integration.
Bringing in new talent with AI expertise–such as a chief AI officer or strategist–can be critical to success. But this should be done in parallel with reskilling in-house employees.
More than half of the CEOs surveyed (52 percent) said they are fearful that leadership will have less visibility into their business due to AI and automation.
Tomi Engdahl says:
Are We Master or Machine?
https://www.eetimes.com/author.asp?section_id=36&doc_id=1333144
The two AI leaders are the US and China. In the US, it’s entirely driven by the private sector… Chinese players collect a lot of data driven by a government. Neither reflects our principles and values, says French President.
Most U.S. consumers, swept into the era of Big Data, in a prosperous nation where Google, Apple, Facebook and Amazon have become the definition of Big Business, don’t readily dwell on the constitutional complexities of personal privacy.
One of my friends on Facebook (yes, I use Facebook) posted a message last Thursday: “Why is anyone surprised that Facebook sells data about its users? That’s their entire business!”
My colleague Rick Merritt also penned an op-ed pointing out: “Traders may pound Facebook, stock and regulators may draw up rules for online privacy, but the Pandora’s box of big data is forever open.”
Touché. I reckon that the popular sentiment is that privacy train left the station a long time ago. We already feel powerless. It’s easier to give up than fight it, because after all, no one has the individual power to cram Pandora back into her box.
Tomi Engdahl says:
Are We Short of Deep Learning Experts?
DeepScale gets $15 million investment
https://www.eetimes.com/document.asp?doc_id=1333147
DeepScale, a Mountain View, Calif.-based startup founded in 2015 to develop deep-learning perception software for ADAS and self-driving vehicles, has just closed its Series A funding with $15 million from two venture firms, Point72 and next47.
Most revealing, however, in a phone interview with DeepScale CEO Forrest Iandola, was his acknowledgement that the world doesn’t have enough deep-learning experts.
Not enough deep learning experts
Both car OEMs and tier ones’ appetite for software expertise — deep learning in particular — has only grown over the last 18 months. The industry, in general, suffers a chronic knowledge gap in deep learning and how to leverage it to develop software.
Even DeepScale, co-founded by Iandola, a PhD from U.C. Berkeley working on deep neural networks and computer vision systems, feels pressed to internally scale up its expertise more quickly to meet the external demands.
In short, deep learning experts are people whose knowledge matches up in unprecedented ways. It requires specialists in deep learning models, plus people who understand deep learning infrastructure, and others experienced with deep learning data sets. Iandola said, “We plan to offer internal classes and assigning a mentor to every two to three more junior engineers.”
Tomi Engdahl says:
Chipmakers Expand R&D Amid France’s New AI Push
https://www.eetimes.com/document.asp?doc_id=1333155
French President Emmanuel Macron last week launched a bold new AI strategy, backed by up to €1.5 billion (about $1.84 billion) government funding over the next five years. Key players like Fujitsu, Samsung and DeepMind also announced they are establishing increased AI research in the country.
The strategy states data is a key competitive advantage in the global AI race, and it is therefore essential to have a data and AI policy if France and the European Union wish to attain the goals of sovereignty and strategic autonomy. It says this is especially significant to counter the dominance of digital giants in China, Russia and the United States, which have built up their positions by focusing on data collection and use and have a considerable head start.
The policy has ambitious goals, which it adds are necessary steps in the creation of a French and European AI industry, supporting research, encouraging startups, and collecting data that can be used, and shared, by engineers.
France’s AI strategy will focus on four sectors — healthcare, transport-mobility, the environment and defense/security.
Tomi Engdahl says:
The truth about AI, NLP and ML — human involvement is mandatory
https://medium.com/@mikeslone/the-truth-about-ai-nlp-and-ml-human-involvement-is-mandatory-51dab138bccb
The technology revolution today — led by artificial intelligence (AI), natural language processing (NLP), machine learning (ML) etc, is like an awkward adolescent trying to make its way in the world. Many of today’s leading edge technologies will mature and become part of the business mainstream, others will struggle to cope.
The analogy with adolescence works, to an extent, but one related area of concern is the historical tendency to apply generic, traditional human terms to very specific machine advances. This can confuse matters and skew expectations.
For example, a chatbot is not really able to “chat”, except in a very limited sense, while”deep learning” only means that an artificial neural network has several hidden layers which are not nearly as deep as one would assume. It is a trend that is coming back to bite us.
Tomi Engdahl says:
AI Programmed To Solve Zodiac Killer Mystery Is Doing Something Even Creepier On The Side
http://www.iflscience.com/technology/ai-programmed-to-solve-zodiac-killer-mystery-creates-creepy-poetry-on-the-side/
CARMEL, the supercomputer detective, moonlights as a poet, as revealed in a new documentary on HISTORY.
Rather excitingly, there is an online tool where you can pick a topic and CARMEL will come up with a short poem for you about said topic.
http://52.24.230.241/poem//advance/
Tomi Engdahl says:
Blair Hanley Frank / VentureBeat:
AWS’ Amazon SageMaker service now allows developers to experiment with machine learning models locally on personal computers before moving them to the cloud — Amazon Web Services announced today a new way for machine learning developers to build and deploy models through its cloud.
Amazon expands services for AI training, translation, and transcription
https://venturebeat.com/2018/04/04/amazon-expands-services-for-ai-training-translation-transcription/
Amazon Web Services announced today a new way for machine learning developers to build and deploy models through its cloud. The company’s SageMaker AI service gained support for a local mode that lets developers start testing intelligent systems on their personal computers before moving to the cloud.
Using local mode, a developer can first test out different approaches, then send them out to SageMaker for more extensive training on Amazon’s cloud. Customers also have a broader set of compute instances to choose from for powering that training.
Tomi Engdahl says:
New York Times:
Letter circulating within Google, signed by 3,100+ employees, protests the company’s involvement in a Pentagon program that uses AI to interpret video imagery — WASHINGTON — Thousands of Google employees, including dozens of senior engineers, have signed a letter protesting the company’s involvement …
‘The Business of War’: Google Employees Protest Work for the Pentagon
https://www.nytimes.com/2018/04/04/technology/google-letter-ceo-pentagon-project.html
Tomi Engdahl says:
Amazon on AI and Accelerators
From folding proteins to growing clouds
https://www.eetimes.com/document.asp?doc_id=1333158
Not quite 18 months into his job as Mr. AI for Amazon Web Services, Matt Wood is convinced the fledgling business will someday be bigger than the $20 billion/year AWS itself. At a corporate event in San Francisco, Wood talked with EE Times about the status and outlook of deep learning, what Amazon wants in semiconductors for it and his not-so-strange career path from genomics to cloud computing.
After earning a PhD in bioinformatics in 2004, Wood went to work for a U.K. institute that handled a third of the initial work decoding the human genome.
“It was just a sample to get a blueprint. We did 40 other species including zebra fish and the duck-billed platypus — an odd creature with 10 sex chromosomes,” Wood quipped.
Technology caught up with what was a billion-dollar effort that took a decade. A nearby U.K. startup developed a $100,000 system that could sequence a genome in a week.
“They were just around the corner, so they sent their first instrument over in the back of a taxi. Within six months we had 200 more, working on thousands of genomes and cell lines,” he recalled.
Tomi Engdahl says:
Three Concepts for Managing AI
https://www.eetimes.com/author.asp?section_id=36&doc_id=1333151
Three key ideas should drive how AI is rolled out in the electronics community to reap the full benefits of the new technology.
AI impacts everything, from leadership to business outcomes, across industries and countries, according to a recent Infosys report. The study, Leadership in the Age of AI, surveyed more than 1,000 business and IT leaders at enterprises in seven countries.
In the manufacturing and high-tech industry, 77 percent of leaders reported using AI to automate business processes, which is where most organizations initially focus their AI initiatives. While automation provides clear benefits around efficiency and accuracy, it’s really the tip of the iceberg of the potential ROI organizations could be reaping.
According to the research, 87 percent of organizations in late or final stages of their AI deployments saw significant and measurable benefits from AI technologies. Of those in the later stages of AI deployments, 80 percent of IT decision makers said that they are using AI to augment existing solutions or build new business-critical solutions and services to optimize insights and consumer experience.
Tomi Engdahl says:
Chipmakers Expand R&D Amid France’s New AI Push
https://www.eetimes.com/document.asp?doc_id=1333155
French President Emmanuel Macron last week launched a bold new AI strategy, backed by up to €1.5 billion (about $1.84 billion) government funding over the next five years. Key players like Fujitsu, Samsung and DeepMind also announced they are establishing increased AI research in the country.
The strategy states data is a key competitive advantage in the global AI race, and it is therefore essential to have a data and AI policy if France and the European Union wish to attain the goals of sovereignty and strategic autonomy. It says this is especially significant to counter the dominance of digital giants in China, Russia and the United States, which have built up their positions by focusing on data collection and use and have a considerable head start.
Tomi Engdahl says:
Packing Neural Networks Into End-User Client Devices
How number representation shrinks the footprint.
https://semiengineering.com/packing-neural-networks-into-end-user-client-devices/
https://pages.arm.com/machine-learning-whitepaper-simplifying-neural-networks.html
Tomi Engdahl says:
The Ideal Solution For AI Applications — Speedcore eFPGAs
https://semiengineering.com/the-ideal-solution-for-ai-applications-speedcore-efpgas/
AI requires a careful balance of datapath performance, memory latency, and throughput that requires an approach based on pulling as much of the functionality as possible into an ASIC or SoC. But that single-chip device needs plasticity to be able to handle the changes in structure that are inevitable in machine-learning projects. Adding eFPGA technology provides the mixture of flexibility and support for custom logic that the market requires.
Tomi Engdahl says:
CometML wants to do for machine learning what GitHub did for code
https://techcrunch.com/2018/04/05/cometml-wants-to-do-for-machine-learning-what-github-did-for-code/?utm_source=tcfbpage&utm_medium=feed&utm_campaign=Feed%3A+Techcrunch+%28TechCrunch%29&sr_share=facebook
Comet.ml allows data scientists and developers to easily monitor, compare and optimize their machine learning models. The New York-based company is launching its product today
https://www.comet.ml/
Tomi Engdahl says:
A Step-by-Step Guide to Synthesizing Adversarial Examples
https://www.anishathalye.com/2017/07/25/synthesizing-adversarial-examples/
Synthesizing adversarial examples for neural networks is surprisingly easy: small, carefully-crafted perturbations to inputs can cause neural networks to misclassify inputs in arbitrarily chosen ways. Given that adversarial examples transfer to the physical world and can be made extremely robust, this is a real security concern.
In this post, we give a brief introduction to algorithms for synthesizing adversarial examples, and we walk through the process of implementing attacks in TensorFlow, building up to synthesizing a robust adversarial example following this technique.
Intriguing properties of neural networks
https://arxiv.org/abs/1312.6199
Tomi Engdahl says:
Benjamin Haas / The Guardian:
50+ AI academics call for boycott of South Korean university, KAIST, following a since-deleted post announcing a partnership with weapons-maker Hanwha Systems
‘Killer robots’: AI experts call for boycott over lab at South Korea university
https://www.theguardian.com/technology/2018/apr/05/killer-robots-south-korea-university-boycott-artifical-intelligence-hanwha
Academics around the world voice ‘huge concern’ over KAIST’s collaboration with defence company on autonomous weapons
Tomi Engdahl says:
Ex-Google Executive Opens a School for AI, With China’s Help
https://www.wired.com/story/ex-google-executive-opens-a-school-for-ai-with-chinas-help
When China’s government said last summer it intends to surpass the US and lead the world in artificial intelligence by 2030, skeptics pointed to a major problem. Despite gobs of data from the world’s largest online population, lightweight privacy rules, and 8 million fresh college graduates in 2017, the country doesn’t have enough people skilled in AI to overtake America.
This week Kai-Fu Lee, onetime head of Google’s operations in China, launched a new project to help close the country’s AI talent gap. His helpers include the Chinese government and some of North America’s leading computer scientists. The project is an example of how US and Chinese efforts to progress in AI are entangled, despite recent rhetoric about superpower technology rivalry.
Tomi Engdahl says:
Google Turns to Users to Improve Its AI Chops Outside the US
https://www.wired.com/story/google-turns-to-users-to-improve-its-ai-chops-outside-the-us
Smart algorithms have taken Google a long way. They helped the company dominate search and create the first software to conquer the complex board game Go. Now the company is betting that algorithms that understand images and text will draw business to its cloud services, make augmented reality popular, and prompt us to search using our smartphone cameras. But some of the algorithms Google is staking its future on aren’t equally smart everywhere.
Tomi Engdahl says:
From AI to Russia, Here’s How Estonia’s President Is Planning for the Future
https://www.wired.com/story/from-ai-to-russia-heres-how-estonias-president-is-planning-for-the-future
At 48 years old, Kersti Kaljulaid is Estonia’s youngest president ever, and its first female president.
Known for its digital government, tax, and medical systems, Estonia is planning for the future. The country’s “e-resident” program—which allows global citizens to obtain a government-issued ID card and set up remotely-operated businesses in Estonia—has attracted 35,000 people since 2014.
Tomi Engdahl says:
OpenAI challenges you to beat 1990s classic Sonic the Hedgehog using machine learning
Sega golden oldie repackaged as a research testbed
https://www.theregister.co.uk/2018/04/05/openai_challenge_developers_to_play_sonic_the_hedgehog_with_ai/
OpenAI has launched a new competition using classic Sonic the Hedgehog games as a testbed for transfer learning in AI.
Reinforcement learning is an area of machine learning that tries to teach an agent specific behaviours in a fixed environment. The agent is programmed to explore its environment and experiments with different actions, whenever it makes a good move it is rewarded. Since it’s been programmed to try and maximise its score, the idea is that it should improve and learns how to complete a specific task over time.
It’s often explored in old school video games like Doom, Pacman, or Q*bert. Now, it’s time to bring Sonic the Hedgehog back.
OpenAI has released Gym Retro, a new platform made up of 58 specific scenarios or “save states” from the games: Sonic the Hedgehog, Sonic the Hedgehog 2, and Sonic 3 & Knuckles.
Tomi Engdahl says:
https://www.itewiki.fi/blog/2018/03/suomalainen-tekoalysovellus-gdpr-ai-miner-sai-merkittavan-tunnustuksen-new-yorkissa/
Tomi Engdahl says:
Elon Musk warns that artificial intelligence could turn into an “immortal dictator”
https://www.indy100.com/article/elon-musk-artificial-intelligence-immortal-dictator-spacex-tesla-technology-do-you-trust-this-8291351
Elon Musk seems determined to improve the world with technology, no matter what.
He’s intent on exploring the solar system and travelling to Mars, and even built a huge battery which is powering a region of Australia.
However, one thing that he doesn’t want the world to invest too heavily in is artificial intelligence – which he really, really doesn’t like.
In the past, the South African entrepreneur has said that “AI is far more dangerous than nukes” and now he has delivered another chilling prophecy on the subject.
Tomi Engdahl says:
The Linux Foundation and Open Source Community Members Launch LF Deep Learning to Drive Open Source Growth In AI
https://www.linuxfoundation.org/press-release/the-linux-foundation-and-open-source-community-members-launch-lf-deep-learning-to-drive-open-source-growth-in-ai/
Tomi Engdahl says:
Three Concepts for Managing AI
https://www.eetimes.com/author.asp?section_id=36&doc_id=1333151
Three key ideas should drive how AI is rolled out in the electronics community to reap the full benefits of the new technology.
AI impacts everything, from leadership to business outcomes, across industries and countries, according to a recent Infosys report. The study, Leadership in the Age of AI, surveyed more than 1,000 business and IT leaders at enterprises in seven countries.
Tomi Engdahl says:
IoT Startup Wants Lower Cost AI
Sub-$5 accelerator, cheaper DRAM needed
https://www.eetimes.com/document.asp?doc_id=1333165
A startup that designed a lighting system for smart homes called for cheaper SoCs and DRAMs to bring machine learning to the Internet of Things. Noon Home is not using neural networks yet, but it wants to and it needs lower cost chips to do it.
To serve consumer markets, “we look to single-digit component costs…[and] a lot of AI apps require a lot of memory bandwidth,” said Saket Vora, a former engineering manager on the first Apple Watch and now head of hardware engineering at Noon Home.
Tomi Engdahl says:
Amazon on AI and Accelerators
From folding proteins to growing clouds
https://www.eetimes.com/document.asp?doc_id=1333158
Not quite 18 months into his job as Mr. AI for Amazon Web Services, Matt Wood is convinced the fledgling business will someday be bigger than the $20 billion/year AWS itself. At a corporate event in San Francisco, Wood talked with EE Times about the status and outlook of deep learning, what Amazon wants in semiconductors for it and his not-so-strange career path from genomics to cloud computing.
Tomi Engdahl says:
AI doc ‘Do You Trust This Computer?’ to stream for free this weekend, thanks to Elon Musk: Watch the trailer
http://www.syfy.com/syfywire/ai-documentary-do-you-trust-this-computer-trailer
There’s a new documentary warning about the perils of artificial intelligence out there, and Elon Musk wants you to see it. So much so that he’s making it available to stream for free this weekend.
The documentary — Do You Trust This Computer? — explores the rise of machine intelligence and its possible consequences. There’s a new generation of self-learning computers that has begun to reshape every aspect of our lives. Incomprehensible amounts of data are being created, interpreted, and fed back to us via a flood of apps, personal assistants, smart devices, and targeted advertisements.
Virtually every industry on earth is experiencing this transformation, from job automation, to medical diagnostics, even military operations.
Tomi Engdahl says:
Tiny Neural Network Library in 200 Lines of Code
https://hackaday.com/2018/04/08/tiny-neural-network-library-in-200-lines-of-code/
Neural networks have gone mainstream with a lot of heavy-duty — and heavy-weight — tools and libraries. What if you want to fit a network into a little computer? There’s tinn — the tiny neural network. If you can compile 200 lines of standard C code with a C or C++ compiler, you are in business. There are no dependencies on other code.
On the other hand, there’s not much documentation, either. However, between the header file and two examples, you should be able to figure it out. After all, it isn’t much code.
The tiny neural network library
https://github.com/glouw/tinn
Tomi Engdahl says:
ECTC 2018 To Focus on Heterogeneous Integration for Artificial Intelligence, Wearables, and More
https://www.3dincites.com/2018/04/ectc-2018-to-focus-on-heterogeneous-integration-for-artificial-intelligence-wearables-and-more/
Did you get your ECTC Advanced Program and Registration in the mail yet? Mine arrived yesterday, and I was pleased to see that this year’s Electronics Components and Technologies Conference (ECTC 2018), which takes place May 29-June 1, will be all about the new technology darlings driving development in heterogeneous integration: artificial intelligence, the human-machine interface, wearables, Big Data, with only small dose of autonomous vehicles thrown in for good measure. I don’t know about you, but I’ve grown weary of automotive electronics keynotes and panel discussions. It’s time to move on. I’m also happy that this is the year ECTC is at the Sheraton San Diego Hotel and Marina. Personally, it’s my favorite of the three rotating locations, the other two being Orlando and Las Vegas.
artificial intelligence, the human-machine interface, wearables, Big Data, with only small dose of autonomous vehicles thrown in for good measure.
Tomi Engdahl says:
Artificial, With Questionable Intelligence
Debate begins to stir up about choices machines make.
https://semiengineering.com/artificial-with-questionable-intelligence/
A common theme is emerging in the race to develop big machines that can navigate through a world filled with people, animals, and other assorted objects—if an accident is inevitable, what options are available to machines and how should they decide?
This question was raised at a number of semiconductor industry conferences over the past few weeks, which is interesting because this idea has been kicked around for at least a couple years. But in the past, it was merely theory. Now, as autonomous vehicles begin to roll onto city streets and highways and real accidents occur, this issue has seen a sustained resurgence.
On a business level, the rollout of autonomous vehicles has been generating an increasing number of questions about who is financially responsible for these accidents. AI has fascinating possibilities, and machine learning/deep learning are valuable tools, but these technologies are still deep in the research phase even though they are being applied to real products today. Changes to algorithms are being made almost daily, which is why most of the chips being used for these devices are either inexpensive arrays of GPUs for the training, or some type of programmable logic for the inferencing.
Tomi Engdahl says:
Neural Networks: All YOU Need to Know
https://towardsdatascience.com/nns-aynk-c34efe37f15a
The backbone of any large scale ML project starts with a Network… A Neural Network and Here’s all you need to know about them.
As stated in the sub-title, Neural Nets(NNs) are being used almost everywhere, where there is need of a heuristic to solve a problem. This article will teach you all you need to know about a NN. After reading this article, you should have a general knowledge of NNs, how they work, and how to make one yourself.
Tomi Engdahl says:
Artificial intelligence in the industrial enterprise
https://www.controleng.com/single-article/artificial-intelligence-in-the-industrial-enterprise/2e0569050f5a89c15ab4c1b898fa313e.html
Analytics can deliver insight as to how things are going, but artificial intelligence (AI) doesn’t become a thing until you start using machine learning and semantics for insight.
Automation can improve a process. Productivity can gain from examination of workflows and leading indicators. And analytics deliver insight as to how things are going. But it isn’t till you step over into the cognitive, with things like machine learning and semantics, that the realm of artificial intelligence (AI) is entered.
For the Industrial Internet of Things (IIoT), predictive maintenance of machinery and equipment is the first application demonstrating wide commercial acceptance. “This can be done with classic regression and predictive analytics. With artificial intelligence, however, you go beyond the structured deterministic to the fuzzier stochastic,” said Jeff Kavanaugh, vice president, senior partner, Infosys. “With machine learning based on input such as audio signatures, the computer learns as a human would, by first paying attention to how a machine sounds when it’s healthy and then understanding anomalies.”
Sample set asymmetry
A question often asked is whether companies have the data needed to enable machine learning, and whether the data is in a form suitable for such use. “People have more data than they think, but less than they hope,” said Kavanaugh. “While there are a lot of data stores that don’t lend themselves to machine learning, there are instances where great amounts of data simply aren’t needed. At other times, companies can build on the power of accumulated data. Industrial manufacturers do have deep troves of simple data which can be converted to use cases, where they can go deep.”
“We’re talking about things that are inherently cognitive, in other words fuzzy. While the earlier transformation was from full analog to computerized operations, the current one is more pervasive, more connected, more intelligent—and ultimately—more profound.”