3 AI misconceptions IT leaders must dispel


 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,”


  1. Tomi Engdahl says:

    Cutting-Edge Face Recognition is Complicated. These Spreadsheets Make it Easier.

    9 Steps to Building a Deep Convolutional Neural Net in Excel for Normal Humans.

  2. Tomi Engdahl says:

    Space station robot goes rogue: International Space Station’s artificial intelligence has turned belligerent

    CIMON isn’t much to look at. It’s just a floating ball with a cartoonish face on its touch screen. It’s built to be a personal assistant for astronauts working on the International Space Station (ISS).

    It’s also supposed to be something more.

    CIMON stands for Crew Interactive MObile compinioN.

    Yes, it’s a personality prototype.

    You can tell, can’t you?

    But, as numerous books and movies have clearly warned us — shortly after being switched on for the first time, CIMON has developed a mind of its own.

    And it appears CIMON wants to be the boss.

    This has CIMON’s ‘personality architects’ scratching their heads.

  3. Tomi Engdahl says:

    A new brain-inspired architecture could improve how computers handle data and advance AI

  4. Tomi Engdahl says:

    AI desperately needs regulation and public accountability, experts say

    Artificial intelligence systems and creators are in dire need of direct intervention by governments and human rights watchdogs, according to a new report from researchers at Google, Microsoft and others at AI Now. Surprisingly, it looks like the tech industry just isn’t that good at regulating itself.

    In the 40-page report (PDF) published this week, the New York University-based organization


  5. Tomi Engdahl says:

    Neuromorphic computing gives AI a real-time boost

    now there are real-time applications emerging at the network edge that require instantaneous learning from small data sets, using minimal computational resources, and these applications can’t afford the network latency inherent in relying on data centers. A type of machine learning called neuromorphic computing fits that bill. One type of neuromorphic computing, known as spiking neural networks (SNN), is particularly well-suited to those demands.

  6. Tomi Engdahl says:

    DeepMind’s AlphaZero now showing human-like intuition in historical ‘turning point’ for AI

    DeepMind’s artificial intelligence programme AlphaZero is now showing signs of human-like intuition and creativity, in what developers have hailed as ‘turning point’ in history.

  7. Tomi Engdahl says:

    AI in 2019: 8 trends to watch

    Forget the job-stealing robot predictions. Let’s focus on artificial intelligence trends – around talent, security, data analytics, and more – that will matter to IT leaders

  8. Tomi Engdahl says:


    Piilaaksolainen Achronix tunnetaan sekä erittäin suorituskykyisistä Speedster-piireistä, että järjestelmäpiirille sulautettavista ohjelmoitavista eFPGA-piireistä. Nyt yhtiö on esitellyt neljännen polven sulautettavat Speedcore-ytimet, joilla suunnitteluihin saadaan tuotua erillinen koneoppimisprosessori.

  9. Tomi Engdahl says:

    Inferencing In Hardware
    How to improve the efficiency of neural networks.

    Cheng Wang, senior vice president of engineering at Flex Logix, examines shifting neural network models, how many multiply-accumulates are needed for different applications, and why programmable neural inferencing will be required for years to come.


  10. Tomi Engdahl says:

    China’s Big AI Plan: Do Toys Count, Too?

    The most convincing indication of China’s rise in the AI market that I recently came across was the plethora of AI-driven toys displayed, demonstrated, and pitched at shop after shop in the Shenzhen airport.

    China’s big AI plan — to dominate the global market with its AI technology by 2030 — is widely known and deeply feared in the rest of the world. The lingering mystery, though, is how much of the best-laid plans of China will come to pass, if ever.

    China’s State Council proposed in July 2017 “A Next-Generation Artificial Intelligence Development Plan.” It says that China will ultimately become the world leader in artificial intelligence, with a domestic AI industry worth almost $150 billion by 2030. The first step is to catch up with the United States on AI technology and applications by 2020.

    It’s easy to call this mere bravado, but it’s probably not wise. As I see it, the Chinese genie is already out of the bottle.

    Beijing’s “national champion” technology leaders — Baidu, Alibaba, Tencent, and iFlyTek — are established, successful firms. They are leading the development of innovation platforms such as self-driving cars, smart cities, computer vision for medical diagnosis, and voice intelligence.

  11. Tomi Engdahl says:

    Inferencing In Hardware
    How to improve the efficiency of neural networks.

  12. Tomi Engdahl says:

    Neuromorphic computing gives AI a real-time boost

    There are a number of different approaches to machine learning, like decision tree learning, inductive logic programming, and association rule learning, but perhaps the most successful and widespread technique is the use of artificial neural networks, or ANNs.

    All neural networks might be considered “artificial” in that they all seek to imitate the neural activity in the brain.

    ANNs have proven to be very effective at a number of tasks, especially those involving pattern recognition. That includes such applications as computer vision, speech recognition, or medical diagnosis from symptoms or scans.

    Data centers versus the edge
    For the past few decades, neural networks have largely been implemented in software, operating as a model, executed on general-purpose processors. The software emulates the way that each individual neuron functions, as well as the interconnections between them that govern their collective behavior.

  13. Tomi Engdahl says:

    Miten rakennat sillan tekoälyhypen ja todellisuuden välille?

    Tekoälyhypeltä on mahdotonta välttyä, sillä tekoälystä kirjoitetaan nyt paljon. Ennen kuin otsikon kysymykseen voi järkevästi vastata, täytyy selventää pari perusasiaa.

  14. Tomi Engdahl says:

    How to build deep learning inference through Knative serverless framework

    Using deep learning to classify images when they arrive in object storage

  15. Tomi Engdahl says:

    How to get started in AI

    Before you can begin working in artificial intelligence, you need to acquire some human intelligence.

  16. Tomi Engdahl says:

    Your Next SoC Will Probably Include AI Acceleration
    It may be possible to get an SoC without AI acceleration, but the trend is to provide mainstream machine-learning support.

    System-on-chip (SoC) solutions continue to get more complex as more specialized hardware is added to optimize the SoC for new applications. Qualcomm’s latest Snapdragon 855 (Fig. 1) highlights this change. The 855 includes a number of blocks including the Snapdragon X24 cellular modem and wireless Wi-Fi, Bluetooth, and GPS support from the Adreno 640 GPU, the Hexagon 690 DSP, the Kryo 485 processor cluster, and the Spectra 380 image signal processor (ISP).

    More Machine Learning

    ML support is showing up at all levels. MediaTek’s Helio P70 is built around an octal core big.LITTLE configuration with four 2.1-GHz Arm Cortex-A73s and four power-efficient, 2-GHz Cortex-A53s. There’s also a 900-MHz Arm Mali-G72 GPU and a dual-core AI processing unit (APU). The APU is designed to handle chores like human pose recognition in real-time as well as augmentation to still images and video. It can deliver 280 GMACs.

    Moving further down the scale is Renesas’ RZ/A2M with DRP. It’s designed to support human machine interfaces (HMIs) including systems with cameras. It has an Arm Cortex-A9 along with Renesas’ Dynamically Reconfigurable Processor (DRP) that provides ML support. The DRP is programmable in C and has optimized DMA support to minimize data movement

  17. Tomi Engdahl says:

    Searching for the Perfect Artificial Synapse for AI

    What’s the best type of device from which to build a neural network? Of course, it should be fast, small, consume little power, have the ability to reliably store many bits-worth of information. And if it’s going to be involved in learning new tricks as well as performing those tricks, it has to behave predictably during the learning process.

  18. Tomi Engdahl says:

    A radical new neural network design could overcome big challenges in AI

    Researchers borrowed equations from calculus to redesign the core machinery of deep learning so it can model continuous processes like changes in health

    An AI researcher at the University of Toronto, he wanted to build a deep-learning model that would predict a patient’s health over time. But data from medical records is kind of messy: throughout your life, you might visit the doctor at different times for different reasons, generating a smattering of measurements at arbitrary intervals. A traditional neural network struggles to handle this. Its design requires it to learn from data with clear stages of observation. Thus it is a poor tool for modeling continuous processes, especially ones that are measured irregularly over time.

  19. Tomi Engdahl says:

    Flex Logix’s Geoff Tate examines how to beat latency in neural networks while retaining high hardware utilization.

    High Neural Inferencing Throughput At Batch=1

    Beating latency in neural networks while retaining high hardware utilization.

  20. Tomi Engdahl says:

    The Autonomous Car’s Big Challenge: Using the Hyperscale Server Fleet to Train AI Neural Networks
    Gloria Lau, head of hardware engineering for Uber Technologies, Inc, explains how hyperscale hardware technology accelerates the training of the neural networks behind self-driving vehicles.

  21. Tomi Engdahl says:

    Deep Learning Cars

    A small 2D simulation in which cars learn to maneuver through a course by themselves, using a neural network and evolutionary algorithms.

    Explained In A Minute: Neural Networks

    Artificial Neural Networks explained in a minute.

    As you might have already guessed, there are a lot of things that didn’t fit into this one-minute explanation. Y

  22. Tomi Engdahl says:

    But what *is* a Neural Network? | Deep learning, chapter 1

  23. Tomi Engdahl says:

    MarI/O – Machine Learning for Video Games

    MarI/O is a program made of neural networks and genetic algorithms that kicks butt at Super Mario World.
    Source Code: http://pastebin.com/ZZmSNaHX

  24. Tomi Engdahl says:

    What are Deep Neural Networks Learning About Malware?


    FireEye’s deep learning classifier can successfully identify malware using only the unstructured bytes of the Windows PE file.
    Import-based features, like names and function call fingerprints, play a significant role in the features learned across all levels of the classifier.
    Unlike other deep learning application areas, where low-level features tend to generally capture properties across all classes, many of our low-level features focused on very specific sequences primarily found in malware.
    End-to-end analysis of the classifier identified important features that closely mirror those created through manual feature engineering, which demonstrates the importance of classifier depth in capturing meaningful features.

  25. Tomi Engdahl says:

    Kent Walker / Google:
    Google says it won’t sell general-purpose facial recognition APIs before working through “important technology and policy questions” — More than 400 million people in the world have diabetes. A third of them have diabetic retinopathy, a complication that can cause permanent blindness.

    AI for Social Good in Asia Pacific

    More than 400 million people in the world have diabetes. A third of them have diabetic retinopathy, a complication that can cause permanent blindness. The good news is that this blindness can be prevented if diabetic retinopathy is detected early. The not-so-good news—the illness is often going undetected because people don’t always get screenings. In major part, this is due to limited access to eye care specialists and staff capable of screening for the disease. In Thailand, for example, there are only about 1,400 eye doctors for approximately five million diabetics.

    This is a problem that AI can help us solve. A few years ago, we worked with eye specialists in India and the U.S. on an AI system to help doctors analyze images of the back of the eye for signs of diabetic retinopathy. The results were promising. Our AI model now detects diabetic retinopathy with a level of accuracy on par with human retinal specialists.

    We should work to make the benefits of AI available to everyone. Besides rolling out this diabetic retinopathy initiative in clinics in India with our partner Verily, we’ve also been conducting research in Thailand over the past few months.

    To gather more of these ideas, we recently launched the Google AI Impact Challenge. Selected organizations who apply to the challenge will receive support from Google’s AI experts and Google.org grant funding from a $25 million pool.


  26. Tomi Engdahl says:

    AI Still Has Trust Issues

    A lot has been accomplished in the last year to improve comprehension, accuracy and scalability of artificial intelligence, but 2019 will see efforts focused on eliminating bias and making decision making more transparent.

    Jeff Welser, vice president at IBM Research, says the organization has hit several AI milestones in the past year and is predicting three key areas of focus for 2019. Bringing cognitive solutions powered by AI to a platform businesses can easily adopt is a strategic business imperative for the company, he said, while also increasing understanding of AI and addressing issues such as bias and trust.

  27. Tomi Engdahl says:

    Searching for the Perfect Artificial Synapse for AI
    Researchers tried out several new devices to get closer to the ideal needed for deep learning and neuromorphic computing

  28. Tomi Engdahl says:

    This Canadian Genius Created Modern AI

    For nearly 40 years, Geoff Hinton has been trying to get computers to learn like people do, a quest almost everyone thought was crazy or at least hopeless – right up until the moment it revolutionized the field. In this Hello World video, Bloomberg Businessweek’s Ashlee Vance meets the Godfather of AI.

  29. Tomi Engdahl says:

    What is backpropagation really doing? | Deep learning, chapter 3

    What’s actually happening to a neural network as it learns?

    Backpropagation calculus | Deep learning, chapter 4

  30. Tomi Engdahl says:

    This AI Can Select Materials for Design Engineers

    Researchers at the National Institute for Materials Science have developed an AI to assist engineers in materials design and selection.

  31. Tomi Engdahl says:


    What if you could generate visual materials such as ads, illustrations and app layouts with a touch of a button? You already can. Kind of.

  32. Tomi Engdahl says:


    Onlookers immediately notice a series of things wrong with the robot

    The android – known as Boris – took to the stage at a Russian technology conference to delight the world with its entirely lifelike moving and dancing.

    It was clear that if the robot was real it would be one of the most advanced

    Local reports straight away noted a variety of things wrong with the robot.

    The organisers of the event did not claim that the android was anything other than a man in a suit, according to local reports. But when footage of it was rebroadcast on state TV, the confusion about whether the robot was real seemed to arise

  33. Tomi Engdahl says:

    Artificial Intelligence Composes New Christmas Songs

    One of the most common uses of neural networks is the generation of new content, given certain constraints. A neural network is created, then trained on source content – ideally with as much reference material as possible. Then, the model is asked to generate original content in the same vein. This generally has mixed, but occasionally amusing, results. The team at [Made by AI] had a go at generating Christmas songs using this very technique.

    Christmas AI generated tunes

    What if AI was trained on Christmas tunes? How would it sound?

    We got the answer. We have trained a neural network to generate tunes with a Christmas feeling for you to enjoy

  34. Tomi Engdahl says:

    Putting Their Foot Down: GOAT Uses AI to Stomp

    Out Fake Air Jordan and Adidas Yeezy Sneakers
    The sneaker startup, which boasts more than 750,000 shoe listings, authenticates high-end kicks with image recognition AI.

  35. Tomi Engdahl says:

    Face Detection and Recognition on the ESP32

    Additionally, the newer ESP32 chips supports security features like secure boot and flash encryption needed to implement real ‘production ready’ Internet of Things smart devices, and interestingly, it now also has support for machine learning at the edge with the ESP-WHO framework


  36. Tomi Engdahl says:

    Russia’s High-Tech AI Robot Turns Out To Be Human In Robot Costume

    Russian state television covered a robotics forum this week. Whilst the whole forum – aimed at younger people getting into robotics – looks pretty cool, the part that caught the eye of Russia 24 was a freakishly human-like robot called Boris.

    If this robot is what Russia 24 reported, it would be an astonishing advance in robotics

    turns out that Boris isn’t a robot at all.

  37. Tomi Engdahl says:

    CDRP – architecture for multi-task deep learning to improve prognostic profiling

    Researchers from Fondazione Bruno Kessler introduce the CDRP (Concatenated Diagnostic-Relapse Prognostic) architecture for multi-task deep learning that incorporates a clinical algorithm, e.g., a risk stratification schema to improve prognostic profiling.

  38. Tomi Engdahl says:

    Parmy Olson / Forbes:
    Babylon Health has claimed its AI bot is as good at diagnosing as human doctors, but sources say the AI’s development was rushed and effectiveness exaggerated

    This Health Startup Won Big Government Deals—But Inside, Doctors Flagged Problems

    To prove their point, the doctors had spent about a day carrying out an audit on their own initiative, according to one current and one former staff member, who asked not to be named for fear of legal repercussions. They found that around 10% to 15% of the chatbot’s 100 most frequently suggested outcomes, such as a chest infection, either missed warning signs of a more serious condition like cancer or sepsis or were just flat-out wrong, according to one insider. The clinicians had gone directly to Parsa that Friday in the hope of stalling the new release. They made their case, and after some negotiation he agreed to delay the rollout.

  39. Tomi Engdahl says:

    This free online tool uses AI to quickly remove the background from images
    A handy utility if you don’t have access to Photoshop

    If you’ve ever needed to quickly remove the background of an image you know it can be tedious, even with access to software like Photoshop. Well, Remove.bg is a single-purpose website that uses AI to do the hard work for you. Just upload any image and the site will automatically identify any people in it, cut around the foreground, and let you download a PNG of your subject with a transparent background. Easy.


  40. Tomi Engdahl says:

    Artificial Intelligence in Cybersecurity is Not Delivering on its Promise

    The Cybersecurity Industry Doesn’t Have Artificial Intelligence Right Yet, But it is Promising Technology

    The application of artificial intelligence (AI) via the implementation of machine learning (ML) is the fastest growing area of

    cybersecurity. We are told that that ML-enhanced products produce results faster and more accurately than can be achieved by human

    operators; and this can result in cost savings through the need for fewer analyst employees.

    What has been largely missing from this assertion is independent verification that the theoretical benefits promoted by ML vendors

    translate to actual benefits in use.

  41. Tomi Engdahl says:

    Renesas Processor Puts Artificial Intelligence at the End Point

    Processor combines high performance with low power, reducing the need for AI computations to be done at the cloud.

    A new processor from Renesas Electronics promises to deliver on-board artificial intelligence (AI) to devices such as body cameras and service robots, reducing the need for intensive computations to be sent to the cloud.

    The RZ/A2M is said to offer ten times as much image processing performance as its predecessor, the RZ/A1, and reportedly does so at very low power consumption. “We can be in the tens or hundreds of tera-operations per second, and still maintain under 3 W of power consumption,” Mark Rootz, senior marketing director for Renesas’ Industrial Business Unit, told Design News. “And when you get under 3 W, you can get smaller size, you can get rid of fans, and you can embed intelligence in a portable application like a body cam.”

    The key to the processor’s capabilities is that it combines a 528-MHz ARM Cortex A9 CPU with a dynamically reconfigurable processor (DRP) module. The DRP is capable of dynamically changing the configuration of its processing circuit from one clock cycle to the next, enabling it to carry out real-time image processing at very low power


Leave a Comment

Your email address will not be published. Required fields are marked *