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,”
7,002 Comments
Tomi Engdahl says:
Far Out AI In Remote Locations
https://semiengineering.com/far-out-ai-in-remote-locations/
How far out can AI on the edge go? What lessons in data management and AI can be gleaned from the semiconductor industry already dealing with remote locations on Earth?
Tomi Engdahl says:
From Data Center To End Device: AI/ML Inferencing With GDDR6
How GDDR memory is ideally suited for the needs of AI/ML inferencing.
https://semiengineering.com/from-data-center-to-end-device-ai-ml-inferencing-with-gddr6/
Tomi Engdahl says:
Pandemic Accelerates MIT Machine Learning Initiative
https://www.eetimes.com/pandemic-accelerates-mit-machine-learning-initiative/
Chatbots are saving the auto insurance industry during the coronavirus pandemic. Powered by machine learning (ML), digital insurance platforms research applicants’ driving records, analyze data, apply risk metrics to coverage and pricing, and issue policies without face-to-face interaction.
Covid-19 has accelerated the development of machine learning applications that solve today’s problems, according to Professor Daniela Rus, faculty director for MIT’S Computer Science Artificial Intelligence Lab (CSAIL) research alliance. Last week, CSAIL launched, via webinar, MachineLearningApplications@CSAIL to develop applications for the latest ML technologies, research challenges limiting ML, and provide professional development for the digital workforce.
Tomi Engdahl says:
Don’t Be Misled By “Accuracy” — 3 Insights For More Useful AI Models
https://www.eetimes.com/dont-be-misled-by-accuracy-3-insights-for-more-useful-ai-models/
Tomi Engdahl says:
AI Race Comes Down to Money—and Brains
https://www.eetimes.com/ai-race-comes-down-to-money-and-brains/
Perhaps an exquisitely trained custom algorithm is required to determine just how much national treasure should be invested in the global AI competition that is taking on aspects of the Cold War space race.
Along with underlying semiconductor technology, artificial intelligence and the algorithms spawned by the dual-use technology have emerged as the focus of strategic competition between China and the west.
AI and semiconductor technology were inextricably linked during a recent DARPA virtual event dedicated to its Electronics Resurgence Initiative. “U.S. leadership in microelectronics is essential to U.S. leadership in artificial intelligence,” Gilman Louie, a member of the National Security Commission on Artificial Intelligence, told the DARPA conference.
Despite the huge sums being poured into public and private sector AI development, some industry analysts argue still more investment is needed to keep pace with China — even if the AI goals of its “Made in China” technology initiatives remain mostly aspirational.
Tomi Engdahl says:
Aspinity Raises $5.3M to Deploy Neuromorphic Analog Processing IC
https://www.eetimes.com/aspinity-raises-5-3m-to-deploy-neuromorphic-analog-processing-ic/
Tomi Engdahl says:
Artificial Brain Gives Robots Unprecedented Sensing Capabilities
Paired with artificial robotic skin, a new neuromorphic processing system can help machines ‘feel’ and manipulate objects similar to humans.
https://www.designnews.com/materials/artificial-brain-gives-robots-unprecedented-sensing-capabilities?ADTRK=InformaMarkets&elq_mid=14371&elq_cid=876648
Tomi Engdahl says:
Semiconductors Enable Human-Like Learning in Robots
https://www.designnews.com/electronics/semiconductors-enable-human-learning-robots
Scientists believe that in the future robots will be able to think and reason like humans, and researchers have developed a neurotransistor that mimics how brain neurons function using semiconductors.
Tomi Engdahl says:
https://www.theguardian.com/commentisfree/2020/sep/08/robot-wrote-this-article-gpt-3?fbclid=IwAR3Xnpe5Y2ztkxaXyQCoRt7zCYNUgvynK3JrIptUTlse51rencXhYrCKtrU
Tomi Engdahl says:
Compiling And Optimizing Neural Nets
Inferencing with lower power and improved performance.
https://semiengineering.com/compiling-and-optimizing-neural-nets/
Edge inference engines often run a slimmed-down real-time engine that interprets a neural-network model, invoking kernels as it goes. But higher performance can be achieved by pre-compiling the model and running it directly, with no interpretation — as long as the use case permits it.
At compile time, optimizations are possible that wouldn’t be available if interpreting. By quantizing automatically, merging nodes and kernels, and binding variables into constants where feasible, substantial efficiency improvements can be achieved. That means inference will run more quickly with less power.
“Running a neural network efficiently and achieving good performance on an edge device have two significant challenges — processing compute intensive convolutions and manipulating large amounts of data,” said Steve Steel, director of product marketing, machine-learning group at Arm. “Both of these challenges must be solved in order to realize a balanced system design.”
A typical configuration has the inference engine running a limited version of one of the machine-learning frameworks, like TensorFlow Lite. That engine takes a relatively abstract version of the neural-network model and interprets the required execution, calling on small, self-contained programs called “kernels” that perform various functions. A convolution would be run by a kernel. An activation function might be another kernel. The run-time interpreter invokes the kernels as it processes the network.
But interpretation adds a layer of computing to the whole problem. In addition to doing the actual inference, additional computing is needed to reduce the high level of the model to specific execution instructions. That takes time and energy to do.
Tomi Engdahl says:
A sheriff launched an algorithm to predict who might commit a crime. Dozens of people said they were harassed by deputies for no reason.
https://www.businessinsider.com/predictive-policing-algorithm-monitors-harasses-families-report-2020-9?fbclid=IwAR01_-6eRsjIQZP34-HIA60AIfpgYsvPvBujbjV56tF9ZUlcKMnms0KEpuE
A Florida sheriff’s office deployed a futuristic algorithm that uses crime data to predict who is likely to commit another crime.
In a sweeping six-month investigation published this week, the Tampa Bay Times reported that the algorithm relied on questionable data and arbitrary decisions and led to the serial harassment of people without any evidence of specific crimes.
According to the report, former sheriff’s office employees said officers went to the homes of people singled out by the algorithm, charged them with zoning violations, and made arrests for any reason they could. Those charges were fed back into the algorithm.
The report shines a light on the pitfalls of algorithm-driven policing and casts doubt on AI-powered tools meant to fight crime.
https://projects.tampabay.com/projects/2020/investigations/police-pasco-sheriff-targeted/intelligence-led-policing/
Tomi Engdahl says:
Faking It Is Almost as Good as the Real Thing
Virtual IMU measurements estimated from YouTube videos supply endless datasets for Human Activity Recognition applications.
https://www.hackster.io/news/faking-it-is-almost-as-good-as-the-real-thing-eced335ab2b6
Tomi Engdahl says:
Bet You Can’t Do That Again
AllenAct will help you get your Embodied AI act together with reproducible results and shorter startup times.
https://www.hackster.io/news/bet-you-can-t-do-that-again-6fc341d01fbd
Tomi Engdahl says:
AI Race Comes Down to Money—and Brains
https://www.eetimes.com/ai-race-comes-down-to-money-and-brains/
Tomi Engdahl says:
Muisti ja tallennus pitää optimoida tekoälyä varten
https://etn.fi/index.php?option=com_content&view=article&id=11138&via=n&datum=2020-09-11_15:46:03&mottagare=31202
Datamäärän räjähdys on johtanut tekoälyn ja koneoppimisen (ML) sovellusten valtavaan kasvuun, joissa muisti ja tallennustila ovat avainasemassa sovellusten onnistumisessa ja nopeudessa. Perinteisiä muisti- ja tallennusjärjestelmiä ei ole suunniteltu näiden suurien tietojoukkojen käsittelyyn, joten tietotekniikan tekoäly- ja ML-sovellusten keskeinen haaste on lyhentää etsintään ja käsittelyyn kuluvaa aikaa.
Tomi Engdahl says:
Konenäköjärjestelmä tunnistaa virheet tekoälyn avulla
https://etn.fi/index.php?option=com_content&view=article&id=11140&via=n&datum=2020-09-11_15:46:03&mottagare=31202
OMRON on ilmoittanut tuovansa maailmanlaajuisesti markkinoille uuden FH-sarjan konenäköjärjestelmän, joka on alan ensimmäinen tekoälytekniikkaa hyödyntävä ratkaisu, joka tunnistaa virheet ilman opetusnäytteitä. Tekoälytekniikka paikantaa luotettavasti sellaisetkin viat, joiden havaitseminen on aiemmin ollut vaikeaa.
Nykyään valmistajilla on yhä suurempi tarve automatisoida prosesseja, jotka perustuvat kokeneiden työntekijöiden aistinvaraisiin tarkastuksiin. Erityisesti silmämääräisessä tarkastuksessa on tärkeää tunnistaa luotettavasti pienimmätkin viat myös tuotantolinjoilla.
Tomi Engdahl says:
https://semiengineering.com/ai-inference-acceleration/
Tomi Engdahl says:
Can We Trust A.I.? Modeling Tool Quantifies How Much a Machine Doesn’t Know
https://www.techbriefs.com/component/content/article/tb/stories/blog/37695?utm_source=TB_Main_News&utm_medium=email&utm_campaign=20200915&oly_enc_id=2460E0071134A8V
A self-driving car needs to make quick decisions as it detects its surroundings. But can you trust a vehicle’s ability to make sound choices within fractions of a second — especially when conflicting information is coming from the car’s cameras, LiDAR, and radar?
“There is hope,” begins the title of a USC engineering team’s research paper.
A new modeling tool from USC indicates when predictions from A.I. algorithms are trustworthy.
The important aspect of the model isn’t necessarily how it determines decisions with certainty — in fact, the model quantifies uncertainty.
“Even humans can be indecisive in certain decision-making scenarios. In cases involving conflicting information,” said lead author Mingxi Cheng . “Why can’t machines tell us when they don’t know?”
Cheng and his team, which included Shahin Nazarian and Paul Bogdan of the USC Cyber Physical Systems Group, created a system called DeepTrust. The A.I. model quantifies a level of uncertainty and uses that information to determine when human intervention is necessary.
The DeepTrust tool employs what is known as subjective logic. Subjective logic uses continuous uncertainty and belief parameters, instead of solely discrete truth values, to make decisions.
DeepTrust factors the uncertainty calculations into an assessment of the overall system architecture — not the hundreds to thousands of individual data points — of the neural networks.
“To our knowledge, there is no trust quantification model or tool for deep learning, artificial intelligence and machine learning,” said Prof. Bogdan. “This is the first approach and opens new research directions.”
Tomi Engdahl says:
Enable #MachineLearning directly from inside the #MariaDB database. Add a virtual #AI layer that allows running ML models using SQL queries
AI-Tables in MariaDB
Zoran Pandovski
https://www.mindsdb.com/blog/ai-tables-in-mariadb?utm_medium=paid+social&utm_source=facebook&utm_campaign=mariadb
Database users are the best to know what data is relevant for ML models. Virtual AI tables in MariaDB allows users to run Automated Machine Learning models directly from inside the database. This article is an overview of this integration capability
AITABLES
AITables differ from normal tables in that they can generate predictions upon being queried and returning such predictions like if it was data that existed on the table. Simply put, an AI-Table allows you to use machine learning models as if they were normal database tables, in something that in plain SQL looks like this;
SELECT FROM WHERE
Tomi Engdahl says:
Automated Machine Learning (AutoML) Libraries for Python
https://machinelearningmastery.com/automl-libraries-for-python/
AutoML provides tools to automatically discover good machine learning model pipelines for a dataset with very little user intervention.
It is ideal for domain experts new to machine learning or machine learning practitioners looking to get good results quickly for a predictive modeling task.
Open-source libraries are available for using AutoML methods with popular machine learning libraries in Python, such as the scikit-learn machine learning library.
Tomi Engdahl says:
Machine Learning Architecture: The Core Components
The 11 fundamental building blocks that make up any machine learning solution
https://datarevenue.com/en-blog/machine-learning-project-architecture?utm_source=facebook&utm_medium=cpc&utm_campaign=%5BNHD%5D+Articles+-+Website+Traffic&utm_content=ML+Architecture
Tomi Engdahl says:
https://hackaday.com/2020/09/20/boost-your-animation-to-60-fps-using-ai/
Tomi Engdahl says:
Wanna live on the edge and play with a multi-core system crammed with 5G, AI? Here’s a dev kit Qualcomm has in mind
Plus: Machine-learning software and another Power processor core opened up by IBM
https://www.theregister.com/2020/09/19/ai_roundup/
Tomi Engdahl says:
Recreating Paintings By Teaching An AI To Paint
https://hackaday.com/2020/06/27/recreating-paintings-by-teaching-an-ai-to-paint/
Tomi Engdahl says:
https://hackaday.com/2020/08/06/generate-positivity-with-machine-learning/
Tomi Engdahl says:
This is straight out of a sci-fi movie.
New AI Can Create Images Based On Your Thoughts
https://www.iflscience.com/technology/new-ai-create-images-based-your-thoughts/
A team of researchers from the University of Helsinki have created a novel brain-computer interface that can sort of figure out what a person is thinking and generate an image based on what that individual has in mind. By interpreting brain signals, the new artificial intelligence (AI) system was able to generate general pictures of faces that matched the characteristics that study participants were thinking of.
In a statement, study author Tuukka Ruotsalo explained that “if you want to draw or illustrate something but are unable to do so, the computer may help you to achieve your goal. It could just observe the focus of attention and predict what you would like to create.”
“The technique does not recognise thoughts but rather responds to the associations we have with mental categories. Thus, while we are not able to find out the identity of a specific ‘old person’ a participant was thinking of, we may gain an understanding of what they associate with old age,” said Senior Researcher Michiel Spapé.
Tomi Engdahl says:
https://hackaday.com/2020/07/30/art-generated-from-the-dubious-comments-section/
Tomi Engdahl says:
Microsoft Gets Exclusive License For OpenAI’s GPT-3 Language Model
https://tech.slashdot.org/story/20/09/22/1852222/microsoft-gets-exclusive-license-for-openais-gpt-3-language-model?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+Slashdot%2Fslashdot%2Fto+%28%28Title%29Slashdot+%28rdf%29%29
Microsoft gets exclusive license for OpenAI’s GPT-3 language model
https://venturebeat.com/2020/09/22/microsoft-gets-exclusive-license-for-openais-gpt-3-language-model/
Microsoft today announced that it will exclusively license GPT-3, one of the most powerful language understanding models in the world, from AI startup OpenAI. In a blog post, Microsoft EVP Kevin Scott said that the new deal will allow Microsoft to leverage OpenAI’s technical innovations to develop and deliver AI solutions for customers, as well as create new solutions that harness the power of natural language generation.
“We see this as an incredible opportunity to expand our Azure-powered AI platform in a way that democratizes AI technology, enables new products, services and experiences, and increases the positive impact of AI at scale,” Scott wrote. “The scope of commercial and creative potential that can be unlocked through the GPT-3 model is profound, with genuinely novel capabilities — most of which we haven’t even imagined yet. Directly aiding human creativity and ingenuity in areas like writing and composition, describing and summarizing large blocks of long-form data (including code), converting natural language to another language — the possibilities are limited only by the ideas and scenarios that we bring to the table.”
Tomi Engdahl says:
#TBT: With his AI-powered brick-sorting machine, Jacques Mattheij discovered there’s no such thing as too much Lego.
How I Built an AI to Sort 2 Tons of Lego Pieces
https://spectrum.ieee.org/geek-life/hands-on/how-i-built-an-ai-to-sort-2-tons-of-lego-pieces
For many years as a child, I did nothing but play with Lego. Eventually I had children of my own, who had a nice Lego collection themselves, but nothing you’d need machinery to sort. That changed after a trip to Legoland in Denmark.
I noticed adults at the park buying Lego in vast quantities, despite its high price. Even second-hand Lego isn’t cheap, sold as it is by the part on specialized websites, or by the boxed set and in bulk on eBay. I noticed that bulk unsorted Lego sells for roughly €10 per kilogram (about US $11/kg), boxed sets go for €40/kg, and collections of rare parts and Lego Technic pieces (the sort used to build complex mechanical creations) go for hundreds of euros per kilo. Consequently, there exists a cottage industry of people who buy new sets and bulk Lego and manually sort all the pieces into more valuable groupings.
figured this would be fun to get into. I put in some eBay bids on locally available large lots of Lego and went to bed. The next morning, I woke up to a rather large number of congratulatory emails from eBay sellers (eBay lesson one: If you win that many auctions, you are bidding too high).
Tomi Engdahl says:
Trained in 10 minutes on a system with an NVIDIA RTX 2080 Ti GPU, MAIL decreases localization errors by 36 percent for better navigation.
MAIL Positioning System Provides a Localization Boost for Indoor Navigation Using Magnetic Readings
https://www.hackster.io/news/mail-positioning-system-provides-a-localization-boost-for-indoor-navigation-using-magnetic-readings-9001ef1afdb8
Trained in 10 minutes on a system with an NVIDIA RTX 2080 Ti GPU, MAIL decreases localization errors by 36 percent for better navigation.
“Knowing accurate indoor locations of pedestrians has great social and commercial values, such as pedestrian heat-mapping and targeted advertising,” the team explains in the paper’s abstract. “Location estimation with sequential inputs (e.g., geomagnetic sequences) has received much attention lately, mainly because they enhance the localization accuracy with temporal correlations.”
Tomi Engdahl says:
https://techcrunch.com/2020/09/24/spectrum-labs-raises-10m-for-its-ai-based-platform-to-combat-online-toxicity/?tpcc=ECFB2020
Tomi Engdahl says:
3 Ways Artificial Intelligence Will Change Healthcare
http://on.forbes.com/6184Gx7YA
It’s no secret that healthcare costs have risen faster than inflation for decades. Some experts estimate that healthcare will account for over 20% of the US GDP by 2025. Meanwhile, doctors are working harder than ever before to treat patients as the U.S. physician shortage continues to grow. Many medical professionals have their schedules packed so tightly that much of the human element which motivated their pursuit of medicine in the first place is reduced.
In healthcare, artificial intelligence (AI) can seem intimidating. At the birthday party of a radiologist friend, she gently expressed how she felt her job would be threatened by AI in the coming decade. Yet, for most of the medical profession, AI will be an accelerant and enabler, not a threat. It would be good business for AI companies as well to help, rather than attempt to replace, medical professionals.
In a previous article, I expressed three ways in which I consistently see AI adding value: speed, cost and accuracy. In healthcare, it’s no different. Here are three examples of how AI will change healthcare.
https://www.forbes.com/sites/konstantinebuhler/2020/02/19/levels-and-limits-of-ai/#3678045030ee
Tomi Engdahl says:
METIER Model Combines Wearable Sensor Activity and User Recognition Tasks to Excel at Both
Designed to share information between the activity and user recognition tasks, METIER can do both at the same time.
https://www.hackster.io/news/metier-model-combines-wearable-sensor-activity-and-user-recognition-tasks-to-excel-at-both-cdc42cd4a07d
Tomi Engdahl says:
https://blogs.microsoft.com/blog/2020/09/22/microsoft-teams-up-with-openai-to-exclusively-license-gpt-3-language-model/
Tomi Engdahl says:
Array programming with NumPy
https://www.nature.com/articles/s41586-020-2649-2
Tomi Engdahl says:
Chinese propaganda network on Facebook used AI-generated faces
https://techcrunch.com/2020/09/22/facebook-gans-takes-down-networks-of-fake-accounts-originating-in-china-and-the-philippines/
Tomi Engdahl says:
DeepMV Captures Multiple Wi-Fi, Ultrasonic Data Sources to Accurately Classify Human Activity
https://www.hackster.io/news/deepmv-captures-multiple-wi-fi-ultrasonic-data-sources-to-accurately-classify-human-activity-3b5c84e4fdce
Designed to sense activity without the subject needing to carry a dedicated device, DeepMV is more accurate than existing solutions.
Tomi Engdahl says:
https://www.uusiteknologia.fi/2020/09/21/kone-ennustaa-ajatuksesi-ja-luo-niista-kuvia/
Tomi Engdahl says:
OnHW Has Useful Written All Over It
OnHW is a large and diverse online handwriting recognition dataset that is publicly available.
https://www.hackster.io/news/onhw-has-useful-written-all-over-it-8780293653ce
Tomi Engdahl says:
https://hackaday.com/2020/09/26/twitter-its-not-the-algorithms-fault-its-much-worse/
Tomi Engdahl says:
Amsterdam and Helsinki Launch Algorithm Registries To Bring Transparency To Public Deployments of AI
https://tech.slashdot.org/story/20/09/29/2116227/amsterdam-and-helsinki-launch-algorithm-registries-to-bring-transparency-to-public-deployments-of-ai?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+Slashdot%2Fslashdot%2Fto+%28%28Title%29Slashdot+%28rdf%29%29
Amsterdam and Helsinki this week launched AI registries to detail how each city government uses algorithms to deliver services, some of the first major cities in the world to do so. From a report:
An AI Register for each city was introduced in beta today as part of the Next Generation Internet Policy Summit, organized in part by the European Commission and the city of Amsterdam. The Amsterdam registry currently features a handful of algorithms, but it will be extended to include all algorithms following the collection of feedback at the virtual conference to lay out a European vision of the future of the internet, according to a city official. Each algorithm cited in the registry lists datasets used to train a model, a description of how an algorithm is used, how humans utilize the prediction, and how algorithms were assessed for potential bias or risks. The registry also provides citizens a way to give feedback on algorithms their local government uses and the name, city department, and contact information for the person responsible for the responsible deployment of a particular algorithm. A complete algorithmic registry can empower citizens and give them a way to evaluate, examine, or question governments’ applications of AI.
Amsterdam and Helsinki launch algorithm registries to bring transparency to public deployments of AI
https://venturebeat.com/2020/09/28/amsterdam-and-helsinki-launch-algorithm-registries-to-bring-transparency-to-public-deployments-of-ai/
Tomi Engdahl says:
Making a relatively minor adjustment to existing deep learning architectures could give AI applications a performance boost
A Simple Neural Network Upgrade Boosts AI Performance
https://spectrum.ieee.org/tech-talk/artificial-intelligence/machine-learning/simple-neural-network-upgrade-boosts-ai-performance
A huge number of machine learning applications could receive a performance upgrade, thanks to a relatively minor modification to their underlying neural networks.
If you are a developer creating a new machine learning application, you typically build on top of a existing neural network architecture, one that is already tuned for the kind of problem you are trying to solve—creating your own architecture from scratch is a difficult job that’s typically more trouble than it’s worth. Even with an existing architecture in hand, reengineering it for better performance is no small task.
Tomi Engdahl says:
https://etn.fi/index.php/13-news/11227-jo-yli-1200-suomalaisyritysta-hyodyntaa-tekoalya
https://www.uusiteknologia.fi/2020/10/01/raportti-tekoalyn-jarruna-datan-ja-osaamisen-puute/
Tomi Engdahl says:
Computing is warming up to memristors, literally. Researchers at Hewlett Packard Laboratories, Texas A&M, and Stanford have invented a device based on a special kind of temperature-controlled memristor that acts like a neuron and opens the door to copying the human brain’s power efficiency and quirky computational skill.
Memristor Breakthrough: First Single Device To Act Like a Neuron
https://spectrum.ieee.org/nanoclast/semiconductors/devices/memristor-first-single-device-to-act-like-a-neuron
Tomi Engdahl says:
Yet Again, OpenAI’s GPT-3 Proved To Be A Doom For Humanity
https://analyticsindiamag.com/yet-again-openais-gpt-3-proved-to-be-a-doom-for-humanity/
With last week’s major news, we had learnt that in an attempt to extend its partnership with OpenAI, Microsoft acquired an exclusive license to GPT-3. Although the company hasn’t stopped the API access of the model to other users, it indeed provided the access of its underlying codes and mechanism to Microsoft. This would allow the company to leverage the state-of-the-art technical innovations of OpenAI to create advanced products for its customers
Tomi Engdahl says:
https://hackaday.com/2020/10/04/vizy-ai-camera-wants-to-make-machine-vision-less-complex/
Tomi Engdahl says:
https://hackaday.com/2020/10/04/hackaday-links-october-4-2020/
As if 2020 hasn’t dealt enough previews of various apocalyptic scenarios, here’s what surely must be a sign that the end is nigh: AI-generated PowerPoint slides. For anyone who has ever had to sit through an endless slide deck and wondered who the hell came up with such drivel, the answer may soon be: no one. DeckRobot, a startup company, is building an AI-powered extension to Microsoft Office to automate the production of “company compliant and visually appealing” slide decks. The extension will apparently be trained using “thousands and thousands of real PowerPoint slides”. So, great — AI no longer has to have the keys to the nukes to do us in. It’ll just bore us all to death.
DeckRobot Raises $1.5 Million To Generate PowerPoint Slides Using AI
https://www.forbes.com/sites/frederickdaso/2020/10/01/deckrobot-raises-15-million-to-generate-powerpoint-slides-using-ai/
Tomi Engdahl says:
Unis turn to webcam-watching AI to invigilate students taking exams. Of course, it struggles with people of color
Plus: IBM shares its ML know-how in schizophrenia fight
https://www.theregister.com/2020/10/05/in_brief_ai/
In brief AI software designed to monitor students via webcam as they take their tests – to detect any attempts at cheating – sometimes fails to identify the students due to their skin color.
Tomi Engdahl says:
Thomas Nicholson / WIRED UK:
YouTubers are using AI software like DeOldify to upscale and “enhance” historic footage into 4K, but historians say the result undermines the footage
YouTubers are upscaling the past to 4K. Historians want them to stop
https://www.wired.co.uk/article/history-colourisation-controversy
YouTubers are using AI to bring history to life. But historians argue the process is nonsense
The first time you see Denis Shiryaev’s videos, they feel pretty miraculous. You can walk through New York as it was in 1911, or ride on Wuppertal’s flying train at the turn of the 20th century, or witness the birth of the moving image in a Leeds garden in 1888.
Shiryaev’s YouTube channel is a showcase for his company Neural Love, based in Gdansk, Poland, which uses a combination of neural networks and algorithms to overhaul historic images. Some of the very earliest surviving film has been cleaned, unscuffed, repaired, colourised, stabilised, corrected to 60 frames per second and upscaled to vivid 4K resolution.
For viewers, it almost feels like time travel.
But these vivid videos and images haven’t wowed everyone. Digital upscalers and the millions who’ve watched their work on YouTube say they’re making the past relatable for viewers in 2020, but for some historians of art and image-making, modernising century-old archives brings a host of problems. Even adding colour to black and white photographs is hotly contested.
“The problem with colourisation is it leads people to just think about photographs as a kind of uncomplicated window onto the past, and that’s not what photographs are,” says Emily Mark-FitzGerald, Associate Professor at University College Dublin’s School of Art History and Cultural Policy.
Peck says Neural Love makes clear to clients the huge difference the company sees between “the restoration aspect and the enhancement aspect”. They see the removal of scratches, noise, dust or other imperfections picked up during processing as a less ethically fraught process to upscaling and colourising. “You’re really returning the film to its original state,” she says.
That’s not a view many academics hold, however. Luke McKernan, lead curator of news and moving images at the British Library, was particularly scathing about Peter Jackson’s 2018 World War One documentary They Shall Not Grow Old, which upscaled and colourised footage from the Western Front. Making the footage look more modern, he argued, undermined it. “It is a nonsense,” he wrote. “Colourisation does not bring us closer to the past; it increases the gap between now and then. It does not enable immediacy; it creates difference.”
DeOldify and Neural Love, though, see their tools as a means of bridging the gap of understanding opened up by a century of technological advancement. Their tech is a means of making jerky, jittery images seem suddenly modern, but for historians, the distance between now and then is the whole point. “It’s the effort that creates the understanding,” McKernan writes. “Without that there is no true sympathy, only false sentiment. Film that looks like it was shot last week belongs only to last week.”
Tomi Engdahl says:
John McCormick / Wall Street Journal:
AI-based tools are helping California firefighters to monitor fires, evacuate threatened areas, and judiciously steer resources to where they are most needed
California Firefighters Tap AI for an Edge in Battling Wildfires
https://www.wsj.com/articles/california-firefighters-tap-ai-for-an-edge-in-battling-wildfires-11601544600?mod=djemalertNEWS
Stretched by longer, deadlier fire seasons, officials are using artificial intelligence to more closely track blazes and more judiciously steer resources