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,”
5,208 Comments
Tomi Engdahl says:
AI and Machine Learning for Kids © GPL3+
https://create.arduino.cc/projecthub/Fryden-Learning/ai-and-machine-learning-for-kids-2baa1f
Teach ALL kids the fundamentals of AI and machine learning, through games and hardware and software design.
Our mission when creating this project was :
• Teach ALL kids the fundamentals of AI and machine learning
• Bring AI to life with games and hardware and software creation activities
• Send kids home with a computer they assembled and a machine learning application they trained
Tomi Engdahl says:
Moni pelkää kontrollin katoavan, kun tekoäly yleistyy – siksi suomalainen tekoälyä selittävä kurssi on huippusuosittu ulkomailla asti
https://yle.fi/uutiset/3-10721388
Tomi Engdahl says:
Complexity’s Impact On Security
https://semiengineering.com/complexitys-impact-on-security/
How interactions between components can compromise AI inferencing models.
Tomi Engdahl says:
Redefining Expectations for Test
https://semiengineering.com/redefining-expectations-for-test-2/
How to tackle testing automotive and AI systems
Tomi Engdahl says:
Japan and Israel Tech Experts Join Forces to Create AI Products
https://www.designnews.com/automation-motion-control/japan-and-israel-tech-experts-join-forces-create-ai-products/53903235160667?ADTRK=UBM&elq_mid=8275&elq_cid=876648
Musashi Seimitsu Corp. and Poliakine Innovation have created the consortium, Musashi AI, in order to produce an automatic inspection system and a self-driving forklift.
Tomi Engdahl says:
Benefits of AI and machine learning for automation safety systems
https://www.controleng.com/articles/benefits-of-ai-and-machine-learning-for-automation-safety-systems/
Artificial intelligence (AI) and machine learning can use the information from robots and machines to help improve safety standards by learning what was done in the past and applying it to future situations.
Tomi Engdahl says:
Machine-learning algorithms used to make agriculture taste better
https://www.controleng.com/articles/machine-learning-algorithms-used-to-make-agriculture-taste-better/
MIT researchers have developed a method using machine learning to reveal optimal growing conditions to maximize taste and other features.
Tomi Engdahl says:
AI Researchers Answer the 5 Big Questions About Fairness
https://www.eetimes.com/document.asp?doc_id=1334639
The experts in academia and industry we interviewed for this special report provided the following answers to five basic questions about AI fairness and ethics.
Tomi Engdahl says:
Kyle Wiggers / VentureBeat:
Facebook CTO Mike Schroepfer and others on AI techniques and progress in identifying image, video, and multi-lingual text content violating Facebook’s rules
Believe it or not, Facebook is using AI to reduce abuse and bias
https://venturebeat.com/2019/05/01/believe-it-or-not-facebook-is-using-ai-to-reduce-abuse-and-bias/
Frederic Lardinois / TechCrunch:
Facebook launches two new open source AI tools: BoTorch, based on PyTorch, for Bayesian library optimization and Ax, a platform for managing AI experiments
Facebook open-sources Ax and BoTorch to simplify AI model optimization
https://techcrunch.com/2019/05/01/facebook-open-sources-ax-and-botorch-to-simplify-ai-model-optimization/
At its F8 developer conference, Facebook today launched Ax and BoTorch, two new open-source AI tools.
BoTorch, which, as the name implies, is based on PyTorch, is a library for Bayesian optimization. That’s a pretty specialized tool. Ax, on the other hand, is the more interesting launch, as it’s a general-purpose platform for managing, deploying and automating AI experiments. Both tools, though, are part of the same overall work at Facebook, which focuses on what the company calls “adaptive experimentation.” Indeed, Ax interfaces with BoTorch and, internally, Facebook has used the two tools for tasks that vary from optimizing Instagram’s back-end infrastructure to improving the response rates of user surveys.
Tomi Engdahl says:
Machine Learning Challenges at Facebook Scale
https://www.eetimes.com/document.asp?doc_id=1334641
In a crowded presentation at the AI Expo event here, Facebook Engineering Manager Aditya Kalro described the challenges of rolling out AI at the scale required by the social media behemoth.
Kalro described a major push for AI across all the different Facebook products, but it’s used primarily for personalization of the news feed for all 2.7 billion users. A single user simply logging into their account immediately generates 70 or 80 predictions, he said. Around 200 trillion predictions are carried out every day, in the form of face recognition in photos, what content should be seen, advertisement placement, etc.
Tomi Engdahl says:
AI Researchers Answer the 5 Big Questions About Fairness
https://www.eetimes.com/document.asp?doc_id=1334639
Tomi Engdahl says:
The State of Machine Learning in 2019
https://blogs.cisco.com/security/the-state-of-machine-learning-in-2019
An ensemble cast featuring machine learning
Anyone who has built an effective security analytics pipeline knows that job one is to ensure that it is resilient to active evasion. Threat actors know as much or more than you do about the detection methods within the environments they wish to penetrate and persist. The job of security analytics is to find the most stealthy and evasive threat actor activity in the network and to do this, you cannot just rely on a single technique. In order for that detection to happen, you need a diverse set of techniques all of which complement one another. While a threat actor will be able to evade one or two of them simultaneously, they don’t stand a chance against hundreds of them! Detection in diversity!
Tomi Engdahl says:
Improving Edge Inferencing
https://semiengineering.com/improving-edge-inferencing/
Where are the bottlenecks in AI chips and how to boost the efficiency.
Tomi Engdahl says:
Hackathon Winners Present Diverse Solutions
https://www.eeweb.com/profile/eeweb/news/hackathon-winners-present-diverse-solutions
DutchChain has named the winners of its Odyssey open innovation program, billed as the largest blockchain and AI hackathon in the world. The four-day event connects innovative ideas with government, corporate, and nonprofit partners to solve complex challenges using blockchain, artificial intelligence, and other emerging technologies.
Tomi Engdahl says:
Should Computers Run the World? – with Hannah Fry
https://www.youtube.com/watch?v=Rzhpf1Ai7Z4
Algorithms are increasingly used to make decisions in healthcare, transport, finance and security. How can they best be used and what happens when things go wrong?
Tomi Engdahl says:
Deepfake Videos Are Getting Terrifyingly Real I NOVA I PBS
https://www.youtube.com/watch?v=T76bK2t2r8g
Artificially intelligent face swap videos, known as deepfakes, are more sophisticated and accessible than ever.
Tomi Engdahl says:
Life-like Gaming is Now Possible (Thanks to A.I.)
https://www.youtube.com/watch?v=fQlQQSsC47g
A.I. is making its way into gaming, with real-time ray tracing now possible, what other things does A.I. have in store?
Tomi Engdahl says:
Caffe in Jetson Nano
https://hackaday.io/project/165293-caffe-in-jetson-nano
Caffe is a commonly used deep learning framework, we can install it to Jetson Nano.
Tomi Engdahl says:
Tarina Lidlin “hifi”kuulokkeiden takana.
https://jazmanaut.blogspot.com/2018/11/tarina-lidlin-hifikuulokkeiden-takana.html?m=1
Tomi Engdahl says:
U.S. Tech Needs Hard Lines on China
https://foreignpolicy.com/2019/05/03/u-s-tech-needs-hard-lines-on-china-artificial-intelligence-technology-microsoft-google-defense/
Researchers must stay aware of how easily AI work can be turned to repressive ends
Tomi Engdahl says:
Microsoft launches a drag-and-drop machine learning tool
https://techcrunch.com/2019/05/02/microsoft-launches-a-drag-and-drop-machine-learning-tool-and-hosted-jupyter-notebooks/?tpcc=ECTW2019
The new interface for Azure’s automated machine learning tool makes creating a model as easy as importing a data set and then telling the service which value to predict. Users don’t need to write a single line of code, while in the backend, this updated version now supports a number of new algorithms and optimizations that should result in more accurate models. While most of this is automated, Microsoft stresses that the service provides “complete transparency into algorithms, so developers and data scientists can manually override and control the process.”
Tomi Engdahl says:
Autonomous Mars Rover using Raspberry Pi, Arduino and Pi Camera
https://medium.com/@rishavrajendra/autonomous-mars-rover-using-raspberry-pi-arduino-and-pi-camera-5b285be452c1
Tomi Engdahl says:
THE DEFINITIVE GUIDE FOR BUILDING AN ANOMALY DETECTION SYSTEM
https://go.anodot.com/building-large-scale-wp-part-1?utm_source=facebook&utm_medium=paid&utm_campaign=wp_post_us_website_conversions&utm_term=6095012501885
Tomi Engdahl says:
Taylor drift: Finally, a use for AI emerges? Cyber-smut star films scene in Tesla with Autopilot but warns: ‘I wouldn’t recommend it’
Model X? More like Model XXX
https://www.theregister.co.uk/2019/05/04/tesla_nooky_video/
Tomi Engdahl says:
Not a Lot of Debiasing, Auditing Tools Yet
https://www.eetimes.com/document.asp?doc_id=1334650
There are lots and lots of guidelines and best practices for defining AI fairness and what to do about bias, as we describe in a companion article in this special report, Can AI Fairness Be Regulated? There are not a lot of tools that design engineers could use to detect and correct bias in algorithms or datasets, however. Such tools are lacking both for engineers developing their own products and for customers who want to optimize third-party AI systems.
Tomi Engdahl says:
Google Duplex A.I. – How Does it Work?
https://www.youtube.com/watch?v=IuIpgArEZig
In this video we take a look at Google’s Duplex Assistant extension and how it works.
Tomi Engdahl says:
Frederic Lardinois / TechCrunch:
IntelliCode, Microsoft’s tool for AI-assisted coding that was trained on GitHub projects with 100+ stars, is now generally available
Microsoft’s IntelliCode for AI-assisted coding comes out of preview
https://techcrunch.com/2019/05/06/microsofts-intellicode-for-ai-assisted-coding-is-now-generally-available/
IntelliCode, Microsoft’s tool for AI-assisted coding, is now generally available. It supports C# and XAML in Visual Studio and Java, JavaScript, TypeScript and Python in Visual Studio Code. By default, it is now also included in Visual Studio 2019, starting with the second preview of version 16.1, which the company also announced that.
Tomi Engdahl says:
Frederic Lardinois / TechCrunch:
Microsoft announces the limited preview of a new Azure-based platform for building autonomous robots, based on its acquisition of AI startup Bonsai
Microsoft launches a new platform for building autonomous robots
https://techcrunch.com/2019/05/06/microsoft-launches-a-new-platform-for-building-autonomous-robots/
Tomi Engdahl says:
Frederic Lardinois / TechCrunch:
Microsoft Word Online is getting a new feature called Ideas, an AI-powered editor that suggests ways to make your writing more concise and readable
Word’s new AI editor will improve your writing
https://techcrunch.com/2019/05/06/words-new-ai-features-will-help-you-write-better/
If you write in Microsoft Word Online, you’ll soon have an AI-powered editor at your side. As the company announced today, Word will soon get a new feature called “Ideas” that will offer writers all kinds of help with their documents.
Tomi Engdahl says:
Miljardisijoitus tekoälyyn – ilmainen huippukoulutus avataan kaikille
https://www.tivi.fi/uutiset/tv/96f7a1a1-e815-4dff-8242-4741e0f542cf
Koulutusyhteistyöstä vastaava johtaja Sean O’Brien kertoo yhtiön avaavan aiemmin professoreille ja yliopisto-opiskelijoille tarjolla ollutta SAS-koulutusta myös kaikille muille halukkaille.
Toukokuun loppuun mennessä koulutus ja yhtiön ohjelmistot kuitenkin tuodaan kaikkien saataville Coursera-palvelun kautta. Neliosaisen tekoälykurssin voi käydä ilmaiseksi. Mikäli lopuksi haluaa osaamisestaan myös sertifikaatin, siitä pitää maksaa erikseen.
Tomi Engdahl says:
AI Software Learns to Make AI Software
https://www.technologyreview.com/s/603381/ai-software-learns-to-make-ai-software/?fbclid=IwAR2PIa85aqXVqvbYP8cY99BXUodD-LzDWDJ91z0nvYj-gFCotjDh4bPR4Ts
Google and others think software that learns to learn could take over some work done by AI experts.
Tomi Engdahl says:
Vaak AI Spots Shoplifters Before They Steal
BY MATTHEW HUMPHRIES 5 MAR 2019, 12:19 P.M.
https://uk.pcmag.com/news/119918/vaak-ai-spots-shoplifters-before-they-steal
Japanese company Vaak developed software that uses artificial intelligence to detect body language suggesting someone intends to shoplift and alerts staff so they can intervene
Tomi Engdahl says:
Will Machines Ever Learn to Be Fair?
https://www.eetimes.com/document.asp?doc_id=1334649
Will machines ever learn to be fair? Should they? What will it take? Who will decide what that means? And what will the consequences be if they’re not made to be fair?
Why are we talking about AI fairness? Because AIs are increasingly making decisions about people:
will you get hired or promoted for a particular job?
will you get a loan?
are you likely to get cancer?
what should your insurance rates be?
does your self-driving car recognize all people as human?
is the facial recognition app on your smartphone broken?
AIs are not always accurate. There are several reasons for that, but one of the most insidious is that the datasets used for training AIs can start out flawed. If the humans who generated the data have inherent biases, those biases will end up institutionalized in the data. The decisions that AIs programmed with innate bias will sometimes be unfair. Black-box algorithms based on that data will reflect and sometimes emphasize that bias, all unbeknownst to the developer.
Tomi Engdahl says:
Google Boosts Handset AI, Privacy
Pixel smartphones poised for better voice interface
https://www.eetimes.com/document.asp?doc_id=1334670
Tomi Engdahl says:
Google Duplex is less creepy assistant and more autofill on steroids
There was little mention of the human-sounding assistant.
https://www.cnet.com/news/google-duplex-is-less-creepy-assistant-and-more-autofill-on-steroids/
Last year, Google stoked intense controversy when it announced Duplex, a technology that uses eerily human-sounding artificial intelligence software to book restaurant reservations and hair appointments. Immediately, the project spurred a debate over ethics in artificial intelligence and generated fear over a robot’s ability to deceive people.
Now Google is unveiling its follow up — and it’s a lot less disturbing.
Tomi Engdahl says:
P.J. Bednarski / MediaPost:
Survey of 4,500 online shoppers: 17% buy regularly from voice assistants, up 6% YoY; 62% have clicked on an ad on social media with 31% making a purchase
Study: Voice Assistants Far From Hot Marketplace For Buying
https://www.mediapost.com/publications/article/335481/study-voice-assistants-far-from-hot-marketplace-f.html
Some shoppers like to get information about consumer goods from Alexa and the Google Home voice assistant. But when they get down to buying, those shoppers tend to go elsewhere.
Tomi Engdahl says:
Frederic Lardinois / TechCrunch:
Google says its 2nd- and 3rd-gen Cloud TPU Pods, with up to 1,000 Tensor Processing Units, are now in public beta, with the latter capable of 100 petaflops
Google’s newest Cloud TPU Pods feature over 1,000 TPUs
https://techcrunch.com/2019/05/07/googles-newest-cloud-tpu-pods-feature-over-1000-tpus/
Google today announced that its second- and third-generation Cloud TPU Pods — its scalable cloud-based supercomputers with up to 1,000 of its custom Tensor Processing Units — are now publicly available in beta.
The latest-generation v3 models are especially powerful and are liquid-cooled. Each pod can deliver up to 100 petaFLOPS. As Google notes, that raw computing power puts it within the top 5 supercomputers worldwide, but you need to take that number with a grain of salt given that the TPU pods operate at a far lower numerical precision.
Tomi Engdahl says:
We’re expecting AI to eliminate our human errors, fallacies and biases in decision making. However, technology cannot fix these problems. We need human interventions and conversations about AI ethics. Read why Sonal Makhija, anthropologist, lawyer and our latest AI ethics expert, thinks considering ethics can help.
AI – to bias or not to bias?
Tomi Engdahl says:
AI – to bias or not to bias?
https://www.solita.fi/en/blogs/ai-to-bias-or-not-to-bias/?utm_campaign=The%20Impact%20of%20AI&utm_content=90372319&utm_medium=social&utm_source=facebook&hss_channel=fbp-106572039365864
Tomi Engdahl says:
Benchmarking TensorFlow and TensorFlow Lite on the Raspberry Pi
https://blog.hackster.io/benchmarking-tensorflow-and-tensorflow-lite-on-the-raspberry-pi-43f51b796796
Using TensorFlow Lite we see a considerable speed increase when compared with the original results from our previous benchmarks using full TensorFlow.
Tomi Engdahl says:
Tacotron 2: Generating Human-like Speech from Text
https://ai.googleblog.com/2017/12/tacotron-2-generating-human-like-speech.html?m=1
Tomi Engdahl says:
Tekoäly käyttöön yhdellä koodirivillä tai parilla klikkauksella
http://www.etn.fi/index.php/13-news/9448-tekoaly-kayttoon-yhdella-koodirivilla-tai-parilla-klikkauksella
Jatkossa Salesforcen tekoälyn hyödyntäminen onnistuu hyvin ilman koodauksen tai tietokantojen syvällistä tuntemusta. Einstein-alustaan sisältyy koko joukko koneoppimiskykyjä – kuvantunnistusta, aikomuksien tunnista ja vastaavia. Aiemmin niiden hyödyntäminen on vaatinut mallin asettamista, mallin opettamista tai datatieteilijän taitoja.
Tomi Engdahl says:
Google Is Bringing AI to the Edge for Everyone in 2019
https://www.designnews.com/electronics-test/google-bringing-ai-edge-everyone-2019/1500467160761?ADTRK=UBM&elq_mid=8598&elq_cid=876648
The 2019 Google I/O developer conference brought a wave of artificial intelligence announcement as the company touted a shift toward edge-based AI, and a new focus on improved privacy.
Tomi Engdahl says:
Why Machine Learning Is Important to Embedded
https://www.designnews.com/electronics-test/why-machine-learning-important-embedded/3392145060759?ADTRK=UBM&elq_mid=8615&elq_cid=876648
Machine learning is opening up new features and applications that will forever change how users expect their systems to behave.
Machine learning for embedded systems has been gaining a lot of momentum over the past several years. For embedded developers, machine learning was something that data scientists were concerned with and something that lived up on the cloud, far from the resource-constrained microcontrollers that embedded developers work with on a daily basis.
What seems like almost overnight, however, machine learning is suddenly finding its way to microcontroller and edge devices. To some developers, this may seem baffling or at least intriguing. But why is machine learning so important to embedded developers now? Let’s explore a few possibilities.
Tomi Engdahl says:
Self-aware Robotic Arm
https://hackaday.com/2019/04/24/self-aware-robotic-arm/
Tomi Engdahl says:
But Can Your AI Recognize Slugs?
https://hackaday.com/2019/04/05/but-can-your-ai-recognize-slugs/
Tomi Engdahl says:
STMicro’s STM32 Summit Drives a Smarter World
https://www.eeweb.com/profile/eeweb/news/stmicros-stm32-summit-drives-a-smarter-world
ST’s AI Developers Lab demonstrated how the company is making neural networks run simple, fast, and optimized on its STM32 microcontrollers. ST also presented the AI functions of the STM32MP1 microprocessor through connected-home and motor-control demonstrations and a bone-conductive microphone demo, along with a range of STM32 MCUs with embedded AI.
Tomi Engdahl says:
Innovations Pushing AI Toward the Edge
https://www.eeweb.com/profile/richquinnell/articles/innovations-pushing-ai-toward-the-edge
Tomi Engdahl says:
TinyML Sees Big Hopes for Small AI
Engineers inspired, exasperated at inaugural summit
https://www.eetimes.com/document.asp?doc_id=1334484
A group of nearly 200 engineers and researchers gathered here to discuss forming a community to cultivate deep learning in ultra-low power systems, a field they call TinyML. In presentations and dialogs, they openly struggled to get a handle on a still immature branch of tech’s fastest-moving area in hopes of enabling a new class of systems.
“There’s no shortage of awesome ideas,” said Ian Bratt, a fellow in machine learning at Arm, kicking off a discussion.
“Four years ago, things were getting boring, and then machine learning came along with new floating-point formats and compression techniques—it’s like being young again. But there’s a big shortage of ways to use these ideas in a real system to make money,” Bratt said.
“The software ecosystem is a total wild West. It is so fragmented, and a bit of a land grab with Amazon, Google, Facebook and others all pushing their frameworks… So how can a hardware engineer get something out that many people can use,” he asked.
http://tinymlsummit.org/#home
Tomi Engdahl says:
The distributed intelligence triple-whammy: 5G, AI and tinyML
https://www.audioanalytic.com/distributed-intelligence-triple-whammy-5g-ai-tinyml/
Back in September 2018, Forbes contributor Prakash Sangam heralded two major tech trends, AI and 5G, and made the case, in autonomous driving, for distributed intelligence: AI running at the intelligent edge AND in the intelligent cloud. His point is to put critical ‘sensing’ that is required to act immediately in the car, while processing-intensive functions in the cloud. 5G ends up being the connective glue between both intelligent systems, offering low-latency, high bandwidth and high data transfer speeds.