3 AI misconceptions IT leaders must dispel

https://enterprisersproject.com/article/2017/12/3-ai-misconceptions-it-leaders-must-dispel?sc_cid=7016000000127ECAAY

 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,”

7,003 Comments

  1. Tomi Engdahl says:

    Eta Compute Debuts Spiking Neural Network Chip for Edge AI
    https://spectrum.ieee.org/tech-talk/semiconductors/processors/eta-compute-debuts-spiking-neural-network-chip-for-edge-ai

    At Arm Tech Con today, West Lake Village, Calif.-based startup Eta Compute showed off what it believes is the first commercial low-power AI chip capable of learning on its own using a type of machine learning called spiking neural networks. Most AI chips for use in low-power or battery-operated IoT devices have a neural network that has been trained by a more powerful computer to do a particular job. A neural network that can do what’s called unsupervised learning can essentially train itself: Show it a pack of cards and it will figure out how to sort the threes from the fours from the fives.

    Reply
  2. Tomi Engdahl says:

    https://www.tivi.fi/Kaikki_uutiset/tasmalaaketta-osaajapulaan-huippuyliopisto-satsaa-miljardi-dollaria-tekoalyyn-6745135

    MIT is investing $1 billion in an AI college
    One of the most sizable commitments yet to building out the talent-scarce AI field
    https://www.theverge.com/2018/10/15/17978056/mit-college-of-computing-ai-interdisciplinary-research

    Ever since the beginning of the AI boom in the early 2010s, there’s been a corresponding drought in talented AI developers and researchers. The way to fix this is to educate more of them, and today, MIT announced a $1 billion initiative to do exactly that: it will establish a new college of computing to train the next generation of machine learning mavens.

    Importantly, the college isn’t just about training AI skills. Instead, it will focus on what MIT president L. Rafael Reif calls “the bilinguals of the future.” By that, he means students in fields like biology, chemistry, physics, politics, history, and linguistics who also know how to apply machine learning to these disciplines. (Presumably, Reif feels safe borrowing the term “bilingual” because, in the future, AI will be doing all the translation anyway.)

    Reply
  3. Tomi Engdahl says:

    Cate Cadell / Reuters:
    Baidu joins the Partnership on AI, led by US companies like Apple and Google, becoming the first Chinese company to join an AI ethics group
    https://www.reuters.com/article/us-china-ai-baidu/search-engine-baidu-becomes-first-china-firm-to-join-u-s-ai-ethics-group-idUSKCN1MR0J3

    Reply
  4. Tomi Engdahl says:

    AI on a MEMS Device Brings Neuromorphic Computing to the Edge
    For the first time, artificial intelligence has been integrated into a MEMS device
    https://spectrum.ieee.org/tech-talk/robotics/artificial-intelligence/artificial-intelligence-on-a-mems-device-brings-neuromorphic-computing-to-the-edge

    Reply
  5. Tomi Engdahl says:

    The case for open source classifiers in AI algorithms
    https://opensource.com/article/18/10/open-source-classifiers-ai-algorithms?sc_cid=7016000000127ECAAY

    Machine bias is a widespread problem with potentially serious human consequences, but it’s not unmanageable.

    Reply
  6. Tomi Engdahl says:

    Hospitals Roll Out AI Systems to Keep Patients From Dying of Sepsis
    https://spectrum.ieee.org/biomedical/diagnostics/hospitals-roll-out-ai-systems-to-keep-patients-from-dying-of-sepsis

    In hospitals, doctors and nurses keep vigilant watch over patients’ vital signs and blood tests to catch the first symptoms of sepsis. In this life-threatening condition, the body responds to an infection with widespread inflammation that can lead to organ failure. Cases can progress rapidly to severe sepsis and then to septic shock, which has a mortality rate of almost 50­ percent in the United States.

    But even the most vigilant humans get tired, make mistakes, and miss subtle patterns. That’s why several hospitals are experimenting with artificially intelligent sepsis detectors. Researchers say these pilot projects are the first real examples of AI being integrated into hospital operations, with data flowing from electronic medical records and alerts being incorporated into physicians’ workflows.

    Reply
  7. Tomi Engdahl says:

    SkyKnit: How an AI Took Over an Adult Knitting Community
    https://www.theatlantic.com/technology/archive/2018/03/the-making-of-skyknit-an-ai-yarn/554894/

    Ribald knitters teamed up with a neural-network creator to generate new types of tentacled, cozy shapes.

    Reply
  8. Tomi Engdahl says:

    Looking For The Next Big Innovation
    https://semiengineering.com/looking-for-the-next-big-innovation/

    AI is challenging the entire design ecosystem in new ways.

    Reply
  9. Tomi Engdahl says:

    AI-Enhanced Cameras Facilitate Service, Surveillance, and Security
    https://www.eeweb.com/profile/max-maxfield/articles/ai-enhanced-cameras-facilitate-service-surveillance-and-security

    The AI allows for all sorts of things, from recognizing VIP return guests at venues to determining that someone is in the process of shoplifting and alerting security

    A few days ago, I was given a “heads up” that a California-based technology company called Kogniz was poised to announce a $4M seed financing round led by The Entrepreneurs’ Fund, Tom Chavez, Auren Hoffman, and other industry investors who are uniquely focused on machine learning and artificial intelligence (at the time of this writing, this seed financing round has already taken place).

    What makes these cameras special is that, in addition to a high-resolution 4K sensor, they are equipped with AI that allows them to process video in real time to recognize people, objects, and activities using next-generation facial- and object-recognition technologies.

    Reply
  10. Tomi Engdahl says:

    Tech Heavyweights Back AI Chip Startup Syntiant
    https://www.eetimes.com/document.asp?doc_id=1333899

    The venture capital arms of six prominent tech companies — including Microsoft, Amazon and Intel — invested Syntiant, an AI chip startup focusing initially on voice recognition.

    The $25 million Series B funding round — which led by M12 (formerly known as Microsoft Ventures) — brings to just over $30 million the amount raised so far by Syntiant, an Irvine, Calif.-based chip startup founded last year.

    Syntiant is currently sampling its first product, a 20 tera-operations/watt machine-learning chip that the company calls a neural decision processor (NPD). It uses an array of hundreds of thousands of NOR cells to compute TensorFlow neural-network jobs in the analog domain.

    Reply
  11. Tomi Engdahl says:

    Artificial Intelligence Submits Its First Art for Auction
    https://robbreport.com/shelter/auctions/artificial-intelligence-submits-first-art-for-auction-2824380/

    Christie’s charts new territory by tendering a portrait generated entirely by computer.

    In a novel move, Christie’s is offering Portrait of Edmond de Belamy (2018), the first work of art to hit the auction block that was generated not by an artist of the likes of Leonardo but by an algorithm—a computer program—in its sale of Prints & Multiples on October 25.

    The portrait, one of a series of eleven unique images of members of the fictitious Belamy family, was published by Hugo Caselles-Dupré, Pierre Fautrel, and Gauthier Vernier of the Paris-based collective Obvious Art.

    Reply
  12. Tomi Engdahl says:

    Smartsheet:
    Automation and AI will change the pace of business – here’s how. — AI has tremendous potential to reduce costs and increase productivity for the enterprise. But its true impact comes to life when combined with human ingenuity.

    How Automation and AI Will Change the Pace of Business
    https://www.smartsheet.com/blog/how-automation-and-ai-will-change-pace-business?utm_source=TechMeme&utm_medium=sponcon&utm_campaign=automation

    It’s long been the stuff of science fiction, but increasingly, the practical side of artificial intelligence (AI) is very real. On the consumer side, personal assistants such as Siri, Alexa, and Cortana help automate the simplest of tasks. Meanwhile, AI adoption is also accelerating within your workspace.

    According to a recent Spiceworks survey of more than 500 IT professionals, nearly a quarter of large companies have already implemented digital assistants and another 40% expect to follow suit by 2019. For small and medium businesses, that number is closer to 25%.

    It’s virtually indisputable that AI (in the form of natural-language processing, machine learning, and deep learning) has tremendous potential to reduce costs and increase productivity for the enterprise. But its true impact comes to life when combined with human ingenuity.

    Reply
  13. Tomi Engdahl says:

    Building AI SoCs
    How to develop AI chips when algorithms are changing so quickly.
    https://semiengineering.com/building-ai-socs/

    Reply
  14. Tomi Engdahl says:

    What Makes A Good AI Accelerator
    https://semiengineering.com/what-makes-a-good-accelerator/

    Optimizing processor architectures requires a broader understanding data flow, latency, power and performance.

    The rapid growth and dynamic nature of AI and machine learning algorithms is sparking a rush to develop accelerators that can be optimized for different types of data. Where one general-purpose processor was considered sufficient in the past, there are now dozens vying for a slice of the market.

    As with any optimized system, architecting an accelerator — which is now the main processing engine in many SoCs today — requires a deep level of understanding about what data is going to be processed and where. So data flow and latency are just as important as the balance between power and performance, and system architectures are becoming much more complex than a series of components connected by a wire.

    Reply
  15. Tomi Engdahl says:

    http://www.etn.fi/index.php/13-news/8624-uusi-tekoalypiiri-jattaa-nvidian-varjoonsa

    Nvidian kehittämää Xavier-tekoälypiiriä pidetään yleisesti tämän hetken mittatikkuna. Nyt startup Novumind on kertonut lisätietoja uudesta NovuTensor-piiristään, joka iskee pöytään kovemmat suorituskykylukemat.

    Reply
  16. Tomi Engdahl says:

    Using AI To Pull Call Signs From SDR-Processed Signals
    https://hackaday.com/2018/10/09/using-ai-to-pull-call-signs-from-sdr-processed-signals/

    AI is currently popular, so [Chirs Lam] figured he’d stimulate some interest in amateur radio by using it to pull call signs from radio signals processed using SDR. As you’ll see, the AI did just okay so [Chris] augmented it with an algorithm invented for gene sequencing.

    His experiment was simple enough. He picked up a Baofeng handheld radio transceiver to transmit messages containing a call sign and some speech. He then used a 0.5 meter antenna to receive it and a little connecting hardware and a NooElec SDR dongle to get it into his laptop. There he used SDRSharp to process the messages and output a WAV file. He then passed that on to the AI, Google’s Cloud Speech-to-Text service, to convert it to text.

    “Make amateur radio cool again”, said Mr Artificial Intelligence.
    https://towardsdatascience.com/make-amateur-radio-cool-again-said-mr-artificial-intelligence-36cb32978fb2?gi=2c767f95e8b5

    A project on building a speech recognition system for amateur radio communication.

    Reply
  17. Tomi Engdahl says:

    Playing Mortal Kombat with TensorFlow.js. Transfer learning and data augmentation
    https://blog.mgechev.com/2018/10/20/transfer-learning-tensorflow-js-data-augmentation-mobile-net/

    While experimenting with enhancements of the prediction model of Guess.js, I started looking at deep learning. I’ve focused mainly on recurrent neural networks (RNNs), specifically LSTM because of their “unreasonable effectiveness” in the domain of Guess.js. In the same time, I started playing with convolutional neural networks (CNNs), which although less traditionally, are also often used for time series. CNNs are usually used for image classification, recognition, and detection.

    https://github.com/mgechev

    Reply
  18. Tomi Engdahl says:

    Artificial intelligence—parking a car with only 12 neurons
    https://m.techxplore.com/news/2018-10-artificial-intelligenceparking-car-neurons.html

    A vehicle is manoevred into a parking space by a tiny neural network.

    Reply
  19. Tomi Engdahl says:

    New York Times:
    A look at AI-related research for predicting earthquakes by using seismic data, which scientists say is similar to the audio data used to train voice assistants — SAN FRANCISCO — Countless dollars and entire scientific careers have been dedicated to predicting where and when the next big earthquake will strike.
    http://www.nytimes.com/2018/10/26/technology/earthquake-predictions-artificial-intelligence.html

    Reply
  20. Tomi Engdahl says:

    4 human-caused biases we need to fix for machine learning
    https://thenextweb.com/contributors/2018/10/27/4-human-caused-biases-machine-learning/

    Bias is an overloaded word. It has multiple meanings, from mathematics to sewing to machine learning, and as a result it’s easily misinterpreted.

    When people say an AI model is biased, they usually mean that the model is performing badly. But ironically, poor model performance is often caused by various kinds of actual bias in the data or algorithm.

    Machine learning algorithms do precisely what they are taught to do and are only as good as their mathematical construction and the data they are trained on. Algorithms that are biased will end up doing things that reflect that bias.

    There are four distinct types of machine learning bias that we need to be aware of and guard against.

    1. Sample bias
    Sample bias is a problem with training data.

    2. Prejudice bias
    Prejudice bias is a result of training data that is influenced by cultural or other stereotypes.

    3. Measurement bias
    Systematic value distortion happens when there’s an issue with the device used to observe or measure.

    4. Algorithm bias
    This final type of bias has nothing to do with data. In fact, this type of bias is a reminder that “bias” is overloaded.

    Data scientists who understand all four types of AI bias will produce better models and better training data.

    Reply
  21. Tomi Engdahl says:

    Davey Alba / BuzzFeed News:
    After Orlando let its pilot with Amazon’s facial “Rekognition” tech expire amid public outcry in June, FOIA docs show a new pilot has begun under a mutual NDA

    With No Laws To Guide It, Here’s How Orlando Is Using Amazon’s Facial Recognition Technology
    https://www.buzzfeednews.com/article/daveyalba/amazon-facial-recognition-orlando-police-department

    New documents obtained by BuzzFeed News reveal the most detailed picture yet of how the Orlando Police Department is using Amazon Rekognition, the tech giant’s facial recognition technology.

    Reply
  22. Tomi Engdahl says:

    A Painting Created by Open-Source AI Sells for $432K
    https://www.linuxjournal.com/content/painting-created-open-source-ai-sells-432k-selks5-beta-released-mirantis-launches-mirantis

    A painting created by an open-source neural network sold this week for $432K at a London auction house. Obvious is the group behind the work that “used 19-year-old Robbie Barrat’s GAN package, available here on Github, and sourced paintings from Wiki Commons” to create the painting. See the post on TNW for details on the “first portrait ever sold at auction that was made with the assistance of an AI”.

    Reply
  23. Tomi Engdahl says:

    AI-Generated Portrait Sells for $432,500 in an Auction First
    https://www.bloomberg.com/news/articles/2018-10-25/ai-generated-portrait-is-sold-for-432-500-in-an-auction-first

    Print on canvas is the brainchild of a Paris-based collective
    Christie’s had estimated it would fetch as little as $7,000

    It’s signed by the artist: min G max D Ex[log(D(x))] + Ez[log(1-D(G(z)))].

    A portrait created by artificial intelligence fetched $432,500 at Christie’s in New York on Thursday, the first time a computer-generated artwork was offered by a major auction house.

    The print on canvas, titled “Edmond de Belamy, from La Famille de Belamy,” depicts a blurry and unfinished image of a man.

    The work was the brainchild of Obvious Art, a Paris-based collective, with help from an algorithm known as GAN (Generative Adversarial Network).

    “We fed the system with a data set of 15,000 portraits painted between the 14th century to the 20th,” collective member Hugo Caselles-Dupre told Christie’s.

    Reply
  24. Tomi Engdahl says:

    Amazon’s neural net offer to border cops, Waymo charges fares, the first AI portrait sold at auction, and more
    Including: Bonus IBM Watson Health drama
    https://www.theregister.co.uk/2018/10/29/ai_roundup/

    Reply
  25. Tomi Engdahl says:

    Microsoft to staff: We remain locked and loaded with US military – and will keep adding voice to AI ethics debate
    https://www.theregister.co.uk/2018/10/29/microsoft_defends_jedi_bid/

    Servicemen and women ‘deserve’ access to our tech… even if that includes Windows 10 October 2018 Update

    Reply
  26. Tomi Engdahl says:

    Top AI conference NIPS won’t change its name amid growing protest over ‘bad taste’ acronym
    Not a great look for an industry tackling data bias issues
    https://www.theregister.co.uk/2018/10/29/nips_ai_conference/

    Like something out of HBO’s TV satire Silicon Valley, Neural Information Processing Systems is one of the world’s top AI conferences.

    Yes, N. I. P. S. NIPS. And a decision to keep calling it that has somewhat split the machine-learning community.

    NIPS is one of the must-attend AI events of the year for those working or studying in the industry, and has been running since the late 1980s. Tickets for this year’s conference, taking place in Montreal, Canada, in December, sold out in 11 minutes and 38 seconds – faster than Burning Man, the anarchic gathering held in the Nevada desert, USA, and beloved by techies.

    The NIPS acronym is distasteful, unprofessional, or inappropriate, according to its critics, who say some people feel uncomfortable attending a male-dominated conference called NIPS.

    Reply
  27. Tomi Engdahl says:

    AI can predict the structure of chemical compounds thousands of times faster than quantum chemistry
    Traditional math heavy calculations are just too slow
    https://www.theregister.co.uk/2018/10/30/ai_can_predict_the_structure_of_chemical_compounds_thousands_of_times_faster_than_quantum_chemistry/

    AI can help chemists crack the molecular structure of crystals much faster than traditional modelling methods, according to research published in Nature Communications on Monday.

    Reply
  28. Tomi Engdahl says:

    Suomalainen startup kehitti “puolivahingossa” 95 prosenttia työtä säästävän tekoälyn – valtiokin innostui
    https://www.tivi.fi/Kaikki_uutiset/suomalainen-startup-kehitti-puolivahingossa-95-prosenttia-tyota-saastavan-tekoalyn-valtiokin-innostui-6746869

    Reply
  29. Tomi Engdahl says:

    Brain Cell Electronics Explains Wetware Computing Power
    https://hackaday.com/2018/10/28/brain-cell-electronics-explains-wetware-computing-power/

    Neural networks use electronic analogs of the neurons in our brains. But it doesn’t seem likely that just making enough electronic neurons would create a human-brain-like thinking machine. Consider that animal brains are sometimes larger than ours — a sperm whale’s brain weighs 17 pounds — yet we don’t think they are as smart as humans or even dogs who have a much smaller brain. MIT researchers have discovered differences between human brain cells and animal ones that might help clear up some of that mystery.

    Electrical properties of dendrites help explain our brain’s unique computing power
    https://news.mit.edu/2018/dendrites-explain-brains-computing-power-1018

    Neurons in human and rat brains carry electrical signals in different ways, scientists find.

    Reply
  30. Tomi Engdahl says:

    Applying AI to insurance documents, omni:us raises its first Series A funding
    https://techcrunch.com/2018/10/30/appyling-ai-to-insurance-documents-omnius-raises-its-first-series-a-funding/?sr_share=facebook&utm_source=tcfbpage

    Earlier this year omni:us, an AI-driven service which is able to process digital documents (some of which contain handwriting) by classifying them and extracting the valuable data, revealed it was working with over half of the top 10 insurance providers in the German-speaking DACH region.

    Reply
  31. Tomi Engdahl says:

    Privacy group calls on U.S. government to adopt universal AI guidelines to protect safety, security and civil liberties
    https://techcrunch.com/2018/10/29/us-government-universal-artificial-intelligence-guidelines/?sr_share=facebook&utm_source=tcfbpage

    After months of work, a set of guidelines designed to protect humanity from a range of threats posed by artificial intelligence have been proposed.

    Now, a privacy group wants the U.S. government to adopt them too.

    The set of 12 universal guidelines revealed at a meeting in Brussels last week are designed to “inform and improve the design and use of AI” by maximizing the benefits while reducing the risks. AI has been for years a blanket term for machine-based decision making, but as the technology gets better and are more widely adopted, the results of AI-based outcomes are having a greater effect on human lives — from gaining credit, employment, and even to criminal sentencing.

    Reply
  32. Tomi Engdahl says:

    What Makes A Good AI Accelerator
    https://semiengineering.com/what-makes-a-good-accelerator/

    Optimizing processor architectures requires a broader understanding data flow, latency, power and performance.

    Reply
  33. Tomi Engdahl says:

    Building AI SoCs
    https://semiengineering.com/building-ai-socs/

    How to develop AI chips when algorithms are changing so quickly.

    Reply
  34. Tomi Engdahl says:

    Schools Enlist AI to Detect Vaping and Bullies in Bathrooms
    https://spectrum.ieee.org/tech-talk/robotics/artificial-intelligence/schools-enlist-ai-to-detect-vaping-and-bullies-in-bathrooms

    Schools have been removing bathroom doors, posting bathroom monitors, and even closing bathrooms in their struggles to handle the surging popularity of vaping among middle school and high school students. That has translated into steady business for a U.S. company offering AI-assisted school surveillance capable of alerting teachers and administrators to suspected vaping or bullying in bathrooms.

    Reply

Leave a Comment

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

*

*