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,”
5,231 Comments
Tomi Engdahl says:
Colin Lecher / The Verge:
A report by AI Now Institute finds that about 80% of AI professors are men, while just 15% of AI research staff at Facebook and 10% at Google are women
The artificial intelligence field is too white and too male, researchers say
A new report explores AI’s ‘diversity crisis’
https://www.theverge.com/2019/4/16/18410501/artificial-intelligence-ai-diversity-report-facial-recognition
Tomi Engdahl says:
Detecting malaria with deep learning
https://opensource.com/article/19/4/detecting-malaria-deep-learning?sc_cid=7016000000127ECAAY
Artificial intelligence combined with open source tools can improve diagnosis of the fatal disease malaria.
Tomi Engdahl says:
AI and machine learning: Shaking up the space industry
https://www.edn.com/design/sensors/4461772/AI-and-machine-learning–Shaking-up-the-space-industry?utm_content=bufferfd271&utm_medium=social&utm_source=facebook.com&utm_campaign=buffer
Tomi Engdahl says:
The most overlooked path to commercialize AI is for companies to do it themselves
https://techcrunch.com/2019/04/15/the-most-overlooked-path-to-commercialize-ai-is-for-companies-to-do-it-themselves/?tpcc=ECFB2019
Tomi Engdahl says:
Facebook’s AI team maps the whole population of Africa
https://techcrunch.com/2019/04/09/facebooks-ai-team-maps-out-where-everyone-in-africa-lives/?tpcc=ECFB2019
Tomi Engdahl says:
KeyMe raises $50 million for AI-powered kiosks that duplicate keys in seconds
https://venturebeat.com/2019/04/10/keyme-raises-50-million-for-ai-powered-kiosks-that-duplicate-keys-in-seconds/
For the uninitiated, KeyMe offers key duplication and locksmith services at over 2,300 locations in retailers across 46 states
Tomi Engdahl says:
Facebook Has Released a Machine Learning Tool to Help Engineers Code
https://www.designnews.com/electronics-test/facebook-has-released-machine-learning-tool-help-engineers-code/185360139560620?ADTRK=UBM&elq_mid=8208&elq_cid=876648
Facebook AI has released Aroma, a machine learning tool capable of searching large libraries and making code recommendations to engineers.
Tomi Engdahl says:
The Role Of EDA In AI
https://semiengineering.com/the-role-of-eda-in-ai/
Experts at the Table, part 3: Which aspects of AI implementation should EDA create tools for?
Tomi Engdahl says:
Don’t Worry, Be H(AI)ppy
https://www.eetimes.com/document.asp?doc_id=1334572
Tomi Engdahl says:
Qualcomm’s Latest Chip Wants to be the Most Powerful for Cloud-Based AI
https://www.designnews.com/electronics-test/qualcomms-latest-chip-wants-be-most-powerful-cloud-based-ai/129866657760599?ADTRK=UBM&elq_mid=8200&elq_cid=876648
Qualcomm’s Cloud AI 100 is a purpose-built chip for AI inferencing in the cloud that the company says outperforms similar offerings from Nvidia and other competitors.
Tomi Engdahl says:
InferX X1 Coprocessor Takes on Inference at the Edge
https://www.electronicdesign.com/iot/inferx-x1-coprocessor-takes-inference-edge?sfvc4enews=42&cl=article_2_b&utm_rid=CPG05000002750211&utm_campaign=24783&utm_medium=email&elq2=95a7ec9ea8624b4bbf55d3899aff98ff
The InferX X1 edge inference coprocessor developed by Flex Logix delivers artificial-intelligence inference to IoT applications.
Flex Logix’s nnMax machine-learning (ML) inference engine technology, originally developed for embedded FPGAs (eFPGAs), will now be available in the InferX X1 coprocessor
Tomi Engdahl says:
Multi-Layer Processing Boosts Inference Throughput/Watt
How to boost performance in AI chips.
https://semiengineering.com/multi-layer-processing-boosts-inference-throughput-watt/
The focus in discussion of inference throughput is often on the computations required.
For example, YOLOv3, a power real time object detection and recognition model, requires 227 BILLION MACs (multiply-accumulates) to process a single 2 Mega Pixel image! This is with the Winograd Transformation; it’s more than 300 Billion without it.
And there is a lot of discussion of the large size of the weights required for the model: 62 Million in the case of YOLOv3.
What is often overlooked is the size of the intermediate activations.
YOLOv3 has slightly over 100 layers.
The input to the first layer is the 2 Megapixel image: 2M x 3 RGB bytes = 6MegaBytes (MB).
The output of the first layer is an activation of 64MB!
Tomi Engdahl says:
Wave Computing uncorked its AI platform, TritonAI 64, that provides 8-to-32-bit integer-based support for high-performance AI inferencing at the edge. It also includes bfloat16 and 32-bit floating point-based support for edge training in the future. Features include a MIPS 64-bit SIMD engine that is integrated with Wave’s dataflow and tensor-based configurable technology. Additional features include access to MIPS IDE as well as a Linux-based TensorFlow programming environment.
Source: https://semiengineering.com/week-in-review-design-low-power-38/
More:
Wave Computing Unveils New Licensable 64-Bit AI IP Platform to Enable High-Speed Inferencing and Training in Edge Applications
https://wavecomp.ai/wave-computing-unveils-new-licensable-64-bit-ai-ip-platform-to-enable-high-speed-inferencing-and-training-in-edge-applications/
Tomi Engdahl says:
Qualcomm’s Latest Chip Wants to be the Most Powerful for Cloud-Based AI
https://www.designnews.com/electronics-test/qualcomms-latest-chip-wants-be-most-powerful-cloud-based-ai/129866657760599?ADTRK=UBM&elq_mid=8131&elq_cid=876648
Qualcomm’s Cloud AI 100 is a purpose-built chip for AI inferencing in the cloud that the company says outperforms similar offerings from Nvidia and other competitors.
Tomi Engdahl says:
How to Attract AI Talent for Your Medical Device
https://www.mddionline.com/how-attract-ai-talent-your-medical-device?ADTRK=UBM&elq_mid=8135&elq_cid=876648
Competition for AI talent is white hot. So how can medtech companies differentiate themselves to snag the best AI engineers?
Artificial intelligence may still be in its infancy, but the demand for AI talent is already growing at warp speed, according to a recent report from jobs site Indeed Inc. Posts for AI-related roles on Indeed nearly doubled between June 2015 and June 2018. Meanwhile, during the same time period, the percent of searches on Indeed using “AI” or “machine learning” increased by 182%.
Tomi Engdahl says:
From AI Algorithm To Implementation
https://semiengineering.com/from-ai-algorithm-to-implementation/
Experts at the Table, part 2: The transformation from algorithm to implementation has some significant problems that are not being properly addressed today.
Tomi Engdahl says:
”Tekoäly on mullistanut haittaohjelmien aikakauden”
https://www.tivi.fi/uutiset/tv/41849d38-c9f7-4a21-b9d2-717c589f7cdb
HP Sure Sense torjuu haittaohjelmia tekoälyn avulla
https://www.epressi.com/tiedotteet/tietotekniikka/hp-sure-sense-torjuu-haittaohjelmia-tekoalyn-avulla.html
HP Sure Sense hyödyntää syväoppivaa tekoälyä reaaliaikaiseen haittaohjelmien torjumiseen.
Tomi Engdahl says:
Artificial intelligence speeds efforts to develop clean, virtually limitless fusion energy
https://m.phys.org/news/2019-04-artificial-intelligence-efforts-virtually-limitless.html
Artificial intelligence speeds efforts to develop clean, virtually limitless fusion energy
Home
Share
Feedback
Add to favorites
Comments
Newsletter
Full version
Science X profile
About
Artificial intelligence speeds efforts to develop clean, virtually limitless fusion energy
April 17, 2019 by John Greenwald, Princeton Plasma Physics Laboratory
Artificial intelligence speeds efforts to develop clean, virtually limitless fusion energy
Depiction of fusion research on a doughnut-shaped tokamak enhanced by artificial intelligence. Credit: Eliot Feibush/PPPL and Julian Kates-Harbeck/Harvard University
Artificial intelligence (AI), a branch of computer science that is transforming scientific inquiry and industry, could now speed the development of safe, clean and virtually limitless fusion energy for generating electricity. A major step in this direction is under way at the U.S. Department of Energy’s (DOE) Princeton Plasma Physics Laboratory (PPPL) and Princeton University, where a team of scientists working with a Harvard graduate student is for the first time applying deep learning—a powerful new version of the machine learning form of AI—to forecast sudden disruptions that can halt fusion reactions and damage the doughnut-shaped tokamaks that house the reactions.
Tomi Engdahl says:
Untold History of AI: Algorithmic Bias Was Born in the 1980s
https://spectrum.ieee.org/tech-talk/tech-history/dawn-of-electronics/untold-history-of-ai-the-birth-of-machine-bias
Tomi Engdahl says:
Can artificial intelligence help end fake news?
https://horizon-magazine.eu/article/can-artificial-intelligence-help-end-fake-news.html?utm_source=fb&utm_medium=share
Fake news has already fanned the flames of distrust towards media, politics and established institutions around the world. And while new technologies like artificial intelligence (AI) might make things even worse, it can also be used to combat misinformation.
Tomi Engdahl says:
Inmates in Finland are training AI as part of prison labor
Empowerment or exploitation?
https://www.theverge.com/2019/3/28/18285572/prison-labor-finland-artificial-intelligence-data-tagging-vainu
Tomi Engdahl says:
Industrial robotics giant Fanuc is using AI to make automation even more automated
https://techcrunch.com/2019/04/18/industrial-robotics-giant-fanuc-is-using-ai-to-make-automation-even-more-automated/?tpcc=ECFB2019
Then the conventional process of training bin-picking robots means teaching it many rules so it knows what parts to pick up.
“Making these rules in the past meant having to through a lot of iterations and trial and error. It took time and was very cumbersome,”
Tomi Engdahl says:
Nvidia launches its Isaac SDK to help democratize AI-powered robot development
https://techcrunch.com/2019/04/18/nvidia-launches-its-isaac-sdk-to-help-democratize-ai-powered-robot-development/?tpcc=ECFB2019
Tomi Engdahl says:
GPUs Holding Back AI Innovation
https://www.eetimes.com/document.asp?doc_id=1334579&utm_content=buffer9bd79&utm_medium=social&utm_source=facebook.com&utm_campaign=buffer
A computational graph of a deep learning model, visualized by Graphcore’s IPU compiler, Poplar. Millions of vertices and edges represent computation processes and communication between processes, respectively. Clusters represent intensive communication between processes in each layer of the network, with lighter communication between layers.
Tomi Engdahl says:
The right way to do AI in security
https://techcrunch.com/2019/04/10/translating-ai-in-security/?tpcc=ECFB2019
AI often falls short when it comes to cybersecurity, but the benefits can be significant when done correctly
Tomi Engdahl says:
The AI Race Expands: Qualcomm Reveals “Cloud AI 100” Family of Datacenter AI Inference Accelerators for 2020
https://www.anandtech.com/show/14187/qualcomm-reveals-cloud-ai-100-family-of-datacenter-ai-inference-accelerators-for-2020
Tomi Engdahl says:
‘You Can’t Take My Door’, A Country Song Created by a Neural Network That Studied a Catalog of Country Hits
https://laughingsquid.com/country-song-created-by-neural-network/
Elle O’Brien and a team from Botnik Studios created a predictive AI country song entitled “You Can’t Take My Door”. The song was created by training a neural network to learn country music hits and then produce one of its own. The song was then arranged and performed by humans.
Tomi Engdahl says:
Doing Machine Learning From the Arduino Development Environment?
https://blog.hackster.io/doing-machine-learning-from-the-arduino-development-environment-b6c63838a596
The Maixduino, a New MAix-Based Board From Sipeed
Tomi Engdahl says:
This YouTube Channel Streams AI-Generated Death Metal 24/7
https://motherboard.vice.com/en_us/article/xwnzm7/this-youtube-channel-streams-ai-generated-black-metal-247
Dadabots was developed by two music technologists who wanted to prove that a neural network was capable of capturing the subtle stylistic differences between Death Metal, Math Rock, and other lesser-known genres.
Tomi Engdahl says:
Kirsten Korosec / TechCrunch:
After ditching Nvidia, Tesla unveils its own custom chip for enabling “full self-driving” features, says all newly made Model S, 3, and X vehicles include it
https://techcrunch.com/2019/04/22/teslas-computer-is-now-in-all-new-cars-and-a-next-gen-chip-is-already-halfway-done/
Devin Coldewey / TechCrunch:
A look at Tesla’s first self-driving chip, manufactured by Samsung, which can be installed into existing Teslas, but the cost to do so is unknown
https://techcrunch.com/2019/04/22/tesla-vaunts-creation-of-the-best-chip-in-the-world-for-self-driving/
Tomi Engdahl says:
Deep fakes pose threat in 2020 race -US officials
https://mobile.reuters.com/video/2019/04/09/deep-fakes-pose-threat-in-2020-race-us-o?videoId=535655433&videoChannel=118208
Deep fakes are making a lot of people nervous, especially ahead of the 2020 election. Deep-learning computer applications can now generate fake video and audio that’s so slick, you can’t tell what’s real and what’s not. But now Congress and some states are considering legislation to criminalize the malicious creation and distribution of deep fakes.
Tomi Engdahl says:
The biggest danger of AI? Digitalised discrimination
https://www.solita.fi/en/blogs/the-biggest-danger-of-ai-digitalised-discrimination/?utm_campaign=The%20Impact%20of%20AI&utm_content=90087656&utm_medium=social&utm_source=facebook&hss_channel=fbp-106572039365864
There are no second chances with automated decision-making. You can’t plead your case or crack a joke with an AI system, in the same way you could with a human decision-maker. Add the threat that AI systems will mirror existing human biases and the future starts to sound quite scary. But this is not a foregone conclusion. Choices taken today will shape the future. This is why the debate around AI ethics is one of the most crucial issues facing humanity.
Tomi Engdahl says:
Scientists Unveil a ‘Brain Decoder’ That Turns Neural Activity Into Speech
https://www.sciencealert.com/a-new-kind-of-brain-decoder-turns-neural-activity-for-speaking-directly-into-speech
Tomi Engdahl says:
This Robot’s AI Can Develop Its Own Sense of Self-Awareness
https://blog.hackster.io/this-robots-ai-can-develop-its-own-sense-of-self-awareness-8245aa21f26b
Tomi Engdahl says:
Scientists pull speech directly from the brain
https://techcrunch.com/2019/04/24/scientists-pull-speech-directly-from-the-brain/?tpcc=ECFB2019
Tomi Engdahl says:
Google’s SpecAugment achieves state-of-the-art speech recognition without a language model
https://venturebeat.com/2019/04/22/googles-specaugment-achieves-state-of-the-art-speech-recognition-without-a-language-model/
Tomi Engdahl says:
This colorful printed patch makes you pretty much invisible to AI
https://www.theverge.com/2019/4/23/18512472/fool-ai-surveillance-adversarial-example-yolov2-person-detection
The patch only fools a specific algorithm, but researchers are working on more flexible solutions
Tomi Engdahl says:
Computer scientists design way to close ‘backdoors’ in AI-based security systems
https://m.techxplore.com/news/2019-04-scientists-backdoors-ai-based.html
Tomi Engdahl says:
Brain Implant Can Say What You’re Thinking
https://spectrum.ieee.org/the-human-os/biomedical/devices/implant-translates-brain-activity-into-spoken-sentences
Tomi Engdahl says:
How to hide from the AI surveillance state with a color printout
https://www.technologyreview.com/f/613409/how-to-hide-from-the-ai-surveillance-state-with-a-color-printout/
Tomi Engdahl says:
Canny AI: Imagine world leaders singing
https://www.fxguide.com/featured/canny-ai-imagine-world-leaders-singing/
Tomi Engdahl says:
Amateurs’ Al Tells Real Rembrandts From Fakes
https://spectrum.ieee.org/tech-talk/computing/software/the-rembrandt-school-of-ai-an-algorithm-that-detects-art-forgery
Tomi Engdahl says:
Deep Learning vs. Machine Learning: Choosing the Best Approach
https://se.mathworks.com/campaigns/offers/deep-learning-vs-machine-learning-algorithm.html?s_eid=PSB_6119338527805
Tomi Engdahl says:
ARTIFICIAL INTELLIGENCE IS HELPING OLD VIDEO GAMES LOOK LIKE NEW
https://www.theverge.com/2019/4/18/18311287/ai-upscaling-algorithms-video-games-mods-modding-esrgan-gigapixel
Modders are taking advantage of AI tools to update old graphics
Tomi Engdahl says:
Getting Started with the Intel Neural Compute Stick 2 and the Raspberry Pi
Getting Started with Intel’s Movidius Hardware
https://blog.hackster.io/getting-started-with-the-intel-neural-compute-stick-2-and-the-raspberry-pi-6904ccfe963
Tomi Engdahl says:
For Better AI, Turn Up the Contrast on Reality
https://spectrum.ieee.org/computing/software/for-better-ai-turn-up-the-contrast-on-reality
Tomi Engdahl says:
Untold History of AI: How Amazon’s Mechanical Turkers Got Squeezed Inside the Machine
https://spectrum.ieee.org/tech-talk/tech-history/dawn-of-electronics/untold-history-of-ai-mechanical-turk-revisited-tktkt
Tomi Engdahl says:
Vein-pattern recognition is the latest technology driving China’s AI, robotics revolution
https://www.scmp.com/magazines/post-magazine/long-reads/article/3007619/chinese-companies-leading-ai-and-robotics
A new world order is coming, driven by Chinese companies such as DeepBlue Technology and Yitu Technology
Artificial intelligence, vein-pattern recognition and computer vision are already being adopted across China
Tomi Engdahl says:
Stanford researchers’ artificial synapse is fast, efficient and durable
https://news.stanford.edu/2019/04/25/fast-efficient-artificial-synapse-developed/
A battery-like device could act as an artificial synapse within computing systems intended to imitate the brain’s efficiency and ability to learn.
Tomi Engdahl says:
3 of the Best Budget Embedded Platforms for Predictive Analytics
https://www.designnews.com/electronics-test/3-best-budget-embedded-platforms-predictive-analytics/51099917160702?ADTRK=UBM&elq_mid=8440&elq_cid=876648
If you’re a Maker looking to experiment with predictive analytics, these three platforms will get you started for only around $30.