/* ---- Google Analytics Code Below */

Monday, October 31, 2022

Shutterstock Partners with OpenAI

Generating images directly, direction you would expect as these methods evolve. 

AI News

Shutterstock partners with OpenAI to advance AI image generation 

By Ryan Daws | October 26, 2022 | TechForge Media

Categories: Applications, Artificial Intelligence, Companies, Ethics & Society,

Shutterstock is expanding its partnership with OpenAI to advance how AI image generators are trained and how contributors are rewarded.

The stock image service will offer “direct access” to OpenAI’s DALL-E 2 image generator through its website.

“The mediums to express creativity are constantly evolving and expanding. We recognize that it is our great responsibility to embrace this evolution and to ensure that the generative technology that drives innovation is grounded in ethical practices,” said Paul Hennessy, CEO at Shutterstock.

“We have a long history of integrating AI into every part of our business. This expert-level competency makes Shutterstock the ideal partner to help our creative community navigate this new technology. And we’re committed to developing best practices and experiences to deliver on our purpose, which is to empower the world to create with confidence.”

OpenAI’s DALL-E 2 is one of the most popular image generators. To train DALL-E, OpenAI licensed data from Shutterstock.

“The data we licensed from Shutterstock was critical to the training of DALL-E,” said Sam Altman, OpenAI’s CEO.

“We’re excited for Shutterstock to offer DALL-E images to its customers as one of the first deployments through our API, and we look forward to future collaborations as artificial intelligence becomes an integral part of artists’ creative workflows.”

Shutterstock is creating frameworks that will compensate artists when their intellectual property is used and when their works have contributed to the development of AI models.

At its core, DALL-E uses a process called diffusion, which starts with a pattern of random dots and gradually alters that pattern towards an image when it recognises specific aspects of that image.

Put simply, users can input what they’re looking for and DALL-E 2 will generate an often very accurate image.

Due to concerns about the societal impact, including the generation of fake images for propaganda or hate purposes, OpenAI delayed making its tool publicly available until it was able to implement “a variety of mitigations aimed at preventing and mitigating related risks”.

Other popular AI generators – most notably, Midjourney and Stable Diffusion – beat OpenAI to public availability and surged in popularity. ... ' 

Diesel Shortages Loom

Watching the supply chain: 

US diesel supplier warns companies to prepare for shortages, higher prices for consumers

Alexandra RossOctober 31, 2022  in LocalToday.News

Josh Yoder, owner of Yoder Farms, says he spends $1,500 a day on a tractor while diesel costs are skyrocketing.

Fuel supplier Mansfield Energy is taking steps to prepare for a diesel shortage on the East Coast of the United States and is issuing an advisory to companies relying on the fuel to plan rather than panic.

Last week, Mansfield Energy hoisted a red flag on the impending diesel fuel shortage in the US southeastern region, speculating that it could be due to “weak pipeline ship economies” and historically low reserve supply.

NEW JERSEY, USA – MAY 11: An aerial view of the Phillips 66 Oil Refinery is seen on May 11, 2022 in Linden, New Jersey, United States. () (Photo by Tayfun Coskun/Anadolu Agency via Getty Images / Getty Images)

On a normal day, East Coast markets stockpiled 50 million barrels, but less than 25 million barrels are available right now.

Also read: US diesel supplier warns companies to prepare for shortages, higher prices for consumers

“A tight supply of diesel will drive up prices, which will eventually make it too expensive for some people,” the company said in a press release on Monday. “High prices will lower demand enough to balance it with limited supply.”

FUEL COMPANY ISSUES DIESEL SHORTAGE WARNING, SAYS ‘RAPID DEVELOPMENT’ CONDITIONS

Mansfield warned consumers may feel pain at the pump in the US economy, but supply will still be available for businesses that rely on diesel to operate.

“That’s not to say there won’t be occasional situations where there’s a real physical shortage of product,” the company said. “Some cities could get by on diesel for a few days, at least at the terminal level. But the fuel supply chain is dynamic and suppliers will join forces to fill supply gaps.” ... ' 

Watson as an IP and Patent Advisor

Could have used this in the enterprise and with US Government work some time ago ... talked this and Watson Discovery with IBM, seemed to lay fallow for some time.   Demos?

IBMAnnounces Demonstration Project with U.S. Patent and Trademark Office to Help Make Patent Research Insights More Accessible and Usable

NEWS PROVIDED BY  IBM 

Oct 31, 2022, 09:00 ET

WASHINGTON, Oct. 31, 2022 /PRNewswire/ -- IBM (NYSE: IBM) today announces a demonstration project with the U.S. Patent and Trademark Office (USPTO) to test an artificial intelligence (AI) driven intellectual property (IP) analysis tool, the IBM IP Advisor with Watson® Demonstration System. Leveraging conversational AI technology with IBM Watson Assistant, and content insight mining and guided navigation solutions with IBM Watson Discovery, the system is designed to help users more efficiently uncover and analyze relevant patent data, allowing more time for strategic, value-added tasks.

Information known publicly before the actual filing date of a U.S. patent application, such as currently held and similar U.S. patents and published patent applications, is called prior art. When an inventor has an idea for a new product, a prior art study can be helpful in understanding the landscape of the field of invention by assessing not only prior art, but also competitors and opportunities for innovation in the field. However, with the continued rise in the number of patent applications being filed and the increasing volume of patent data, finding relevant, accurate information and analyzing prior art can be difficult and time consuming – especially for new inventors, 'Main Street' retailers and consumers not steeped in patent law. IBM aims to help address this challenge and promote greater participation in the innovation process with the IBM IP Advisor with Watson® Demonstration System.

By enabling natural language processing of queries with IBM Watson Discovery technology, the Demonstration System can help users locate relevant information using their own terminology, without the need to use specific terms and complex methods. And leveraging IBM Watson Assistant conversational AI technology as a guide, users can ask questions of embedded virtual agents to further inform and optimize their patent analysis.

The IBM IP Advisor with Watson® Demonstration System uses a subset of publicly obtainable U.S. patent data and is available for public testing and feedback on the USPTO's Open Data Portal through November 30, 2022.

"Prior art research and analysis is the foundation of a successful product or service launch, but the growth of prior art coupled with the need for appropriate keywords and the right patent classifications can be a barrier to entry for the every-day inventor," said Steve LaFleche, general manager, U.S. public and federal market for IBM. "I'm proud of our work with the USPTO in support of their objective to not just deploy smarter technology, but to build a better, more accessible IP system."  ... 

Quantum Internet Alliance

Continuing to connect to and better understand this effort.   More to follow. 

Quantum Internet Alliance

JOINING FORCES

For the future Quantum Internet.

Our mission

QIA’s mission is to build a global Quantum Internet made in Europe – by developing a full-stack prototype network validating all key sub-systems and by driving an innovative European Quantum Internet ecosystem capable of scaling all sub-systems to world-leading European technology.

Projects

The development of quantum internet technologies sets formidable challenges to which the Quantum Internet Alliance will provide unique solutions.

Partners

Europe’s leading research institutes and actors.


AI Tools for Skill Improvement

 Chess and other game playing as a means to have access to tools to improve skills.  How does this translate to more general skills?

AI Helps Humans Level Up Chess players and programmers now have access to tools that can improve skills and unleash creativity   By Harry Goldstein  In IEEEiOrg

BACK IN THE mid-1970s, IEEESpectrum senior editor Phil Ross played one of the first chess programs capable of vanquishing humans. He capitulated quickly—too quickly, it turned out: Although the program that beat him was good at openings and the middle game, it was terrible at the end game. Fast forward 50 years and the highest-ranked chess players in the world are AIs, with humans trailing far behind.

As Ross points out in his online piece “ AI’s Grandmaster Status Overshadows Chess Scandal,” the recent brouhaha involving world champion Magnus Carlsen and up-and-comer Hans Niemann highlights how chess-playing AIs, referred to as “engines” by the cognoscenti, have overtaken humans in terms of raw game-playing accuracy. The scandal also shows how AIs are being used by players at all levels to get better faster, fostering a boom in the sport.

If you were born a few decades ago, your best shot at playing and learning from grandmasters was either to be lucky enough to know one or to somehow qualify for the high-level tournaments in which they participated. Nowadays, chess newbies can log in and play engines that far exceed their own abilities, learning strategies and moves in days or weeks that in the past might have taken months or years. Engines can also help neophytes and grandmasters alike analyze their own games to give them an edge against human opponents. 

In fact, if you burrow down the rabbit holes of chess YouTube or Twitch, you’ll find grandmasters giving move-by-move analyses of games that engines played against each other. These AI tools, along with the humans who use them, are readily accessible: You can play the cybernetic versions of super grandmasters, like the open-source, current top-ranked chess engine Stockfish 14.1, not to mention tens of millions of human opponents on sites like Chess.com, a global community of some 93 million players and a central player in the cheating controversy.

While it is certainly true that bad actors use AI to cheat at chess—potentially posing an existential threat to the game, as Carlsen has suggested—it is equally true that the chess world has openly embraced AI and has been thriving as a result. Similar risk/reward calculations will need to be made in other domains.  .... ' 

IBM and RedHat

Following this for some time.  Here Marketwatch calls it a big win. 

Opinion: IBM’s expensive and risky pivot to the cloud is finally paying off

Published: Oct. 25, 2022 at 3:10 p.m. ET,  By Alex Salkever

Red Hat is a bright spot for the company, which is starting its third transformation in history

International Business Machines agreed to buy open-source enterprise-software company Red Hat for a whopping $33 billion in 2019, its biggest acquisition ever.

Critics scoffed that the deal was folly. Everyone knew it was a big bet, even for a behemoth such as IBM IBM, +2.78%. The purchase also represented a one-way door that IBM chose to go through — a shift to become primarily an open-source software company.

Red Hat’s business was built on selling services and support for software that anyone could freely download and deploy. Yes, the rising popularity of open-source meant Red Hat would likely continue to be a critical component in enterprise computing.

Read: This one stock stands out when looking at key data from the FAANG+ group

But those critics thought Big Blue would struggle to be one of the open-source cool kids and that its old-school culture would kill Red Hat’s vibe. Most importantly, Red Hat’s business model could never hope to touch the profitability of Big Blue’s tidy proprietary software and hardware businesses. .... ' 

(Much More) 

Apple Not 'Fully Patching' all OS's

Implications?  And to what degree?  Only most serious, and how often?  So what then is an upgrade? seems perilous for users.   Also in Schneier with comments, 

Apple clarifies security update policy: Only the latest OSes are fully patched  in ArsTechnica

New document confirms what security researchers have observed for a few years.

ANDREW CUNNINGHAM - 10/27/2022, 12:18 PM

Earlier this week, Apple released a document clarifying its terminology and policies around software upgrades and updates. Most of the information in the document isn't new, but the company did provide one clarification about its update policy that it hadn't made explicit before: Despite providing security updates for multiple versions of macOS and iOS at any given time, Apple says that only devices running the most recent major operating system versions should expect to be fully protected.

FURTHER READING

PSA: Apple isn’t actually patching all the security holes in older versions of macOS

Throughout the document, Apple uses "upgrade" to refer to major OS releases that can add big new features and user interface changes and "update" to refer to smaller but more frequently released patches that mostly fix bugs and address security problems (though these can occasionally enable minor feature additions or improvements as well). So updating from iOS 15 to iOS 16 or macOS 12 to macOS 13 is an upgrade. Updating from iOS 16.0 to 16.1 or macOS 12.5 to 12.6 or 12.6.1 is an update.

"Because of dependency on architecture and system changes to any current version of macOS (for example, macOS 13)," the document reads, "not all known security issues are addressed in previous versions (for example, macOS 12)."

In other words, while Apple will provide security-related updates for older versions of its operating systems, only the most recent upgrades will receive updates for every security problem Apple knows about. Apple currently provides security updates to macOS 11 Big Sur and macOS 12 Monterey alongside the newly released macOS Ventura, and in the past, it has released security updates for older iOS versions for devices that can't install the latest upgrades.  ... ' 

Data Sharing with UK

Not sure to what extent and detail it did exist before.

A data-sharing agreement between the US and UK is now in effect   in Engaget

The countries say the pact will help combat serious crimes, but privacy advocates have raised concerns.

By Kris Holt

As of today, a data-sharing pact between the US and the UK is in effect, five years after it was first floated. The two sides claim that the Data Access Agreement, which was authorized by the Clarifying Lawful Overseas Use of Data (CLOUD) Act in the US, will help law enforcement to combat serious crimes in both countries. The Department of Justice called the initiative the first of its kind, adding that it would enable investigators "to gain better access to vital data" to fight serious crimes in a manner that's "consistent with privacy and civil liberties standards."

Under the agreement, authorities in one country can request data from ISPs in the other country, as long as it's related to preventing, detecting, investigating and prosecuting serious crimes including terrorism, transnational organized crime and child exploitation. US officials can't submit data requests targeting people in the UK and vice-versa — presumably the requests can either be used to assist domestic investigations or investigations into foreign nationals. Authorities also need to adhere to certain requirements, limitations and conditions when they access and use data.

The UK Home Office's Investigatory Powers Unit will oversee the Data Access Agreement in the UK, while the DOJ's Office of International Affairs (OIA) will handle matters in the US. The OIA has put together a CLOUD team that will review and certify orders on behalf of federal, state, local and territorial authorities. It will directly submit orders to ISPs in the UK and ensure data is transferred to authorities who requested it.

Privacy advocates have blasted the initiative and the CLOUD Act. In 2018, just after the bill was introduced, the Electronic Frontier Foundation said it "creates a dangerous precedent for other countries who may want to access information stored outside their own borders, including data stored in the United States." Fight for the Future argued that it would threaten user privacy. ... '

Amazon has a new plan for its home robot Astro: to guard your life

Will robots soon be protecting us? 

HUMANS AND TECHNOLOGY  In TechnologyReview

Amazon has a new plan for its home robot Astro: to guard your life

The cute home assistant could be the most powerful, invasive home robot we’ve seen thus far.

By Tanya Basu  archive page,  September 28, 2022

Amazon's home robot, Astro, will be getting a slew of major updates aimed at further embedding it in homes—and in our daily lives, the firm announced on Wednesday.

Broadly speaking, the new features offer more home monitoring. The capabilities include some standard fare: Astro will be able to watch pets and send a video feed of their activities to users, for example. But Astro will also be able to wander around the house to keep an eye on rooms and entry points.

“This will start with doors and windows, so that Astro can alert you if something was left open that shouldn’t have been,” Ken Washington, vice president of consumer robotics at Amazon, said in a presentation on new Amazon devices and services. 

Amazon also announced a new collaboration between Astro and the Ring home security camera system, called Virtual Security Guard, which would protect areas outside the home from possible break-ins. Amazon, which bought Ring in 2018, pitched the pairing as a way to further guard small businesses from break-ins, by videotaping intrusions and calling the authorities (though it seems like homeowners should be able to use that capability as well). 

Ring’s approach to surveillance hasn’t been without controversy. As my colleague Eileen Guo reported last year, Ring marketed itself as a tool to protect domestic violence survivors, but it simultaneously provided access into survivors’ lives. Ring has also been called out for racial profiling and privacy violations. It’s reasonable to ask whether combining Astro’s ability to roam around a house with Ring’s established surveillance system, might create even more surveillance problems than either product did in their previous iterations.

Astro’s evolution as a security guard is a notable one. Astro was introduced nearly a year ago, and since then, reviews of the home robot have been limited. That’s because potential customers had to be invited to test the robot (Amazon offered Astro for $999 to early invitees, but it has since raised the cost to $1,450). The robot, with its big eyes and R2-D2-like structure, was undeniably cute.  But even those limited user reviews were mixed. Many people found that Astro was incapable of doing much more than delivering objects between rooms. It also hit snags in mapping rooms, and some users found its intense focus in following a person around almost creepy.

Related Story

Amazon’s Astro robot is stupid. You’ll still fall in love with it.

From Jibo to Aibo, humans have a long track record of falling for their robots. Except this one’s sold by Amazon.

Still, last year’s pitch was that Astro would make home life more comfortable and entertaining—a sort of adorable, bumbling sidekick. This year, Amazon has recast Astro as a device with a much more serious mission: to provide another set of eyes on our pets, our homes, and our livelihoods. 

As with any surveillance technology you invite into your life, using Astro will require an element of trust. This time, however, you would be extending that trust to a robot capable of moving through your home. That might seem like a small change, but makes a huge difference not only in how people interact with and view home robots but also in how involved Amazon is in our private lives. (In a statement, Amazon said that privacy was folded into Astro's design: "We know that customer trust is paramount, which is why we’ve taken a privacy-by-design approach with Astro since the beginning of its development.”)

Nevertheless, Maya Cakmak, an assistant professor at the University of Washington and head of the Human-Centered Robotics Lab, says Astro’s compatibility with other aspects of Amazon’s home surveillance ecosystem, Ring and Alexa, could well set it up for success. “Astro can provide these services seamlessly,” says Cakmak.

This piece has been updated with comment from Amazon.

by Tanya Basu in TechnologyReview

Sunday, October 30, 2022

Tesla Humanoid Robotics

 Was asked to comment on this, on status, direction and effort involved.  Pointers? 

Tesla is borrowing Autopilot team for Optimus humanoid robot project  in Electric

Fred Lambert - Sep. 20th 2022 3:14 am PT

The CEO has said that the Tesla Bot project has become a priority.

Tesla’s approach to selling Autopilot, and now its Full Self-Driving (FSD) package features before they are ready, is creating some pressure to deliver, but Tesla has missed several timelines already.

CEO Elon Musk has talked about Tesla turning its fleet into 1 million “robotaxi” vehicles by the end of the year, but the goal has now changed to expanding its FSD Beta program, which is still far from being able to perform a robotaxi service.

Many customers are getting impatient to the point that some have filed a class action lawsuit against Tesla over its claims about Autopilot and the FSD package.

Some were also worried that Tesla was spreading itself thin when adding its Optimus humanoid robot project to its plate, and Musk even said that the robot has become a “top priority” at Tesla.

Tesla planned to leverage its work on computer vision for self-driving to use in the robot, which raised concerns considering the automaker has yet to deliver on its self-driving promises despite selling the feature to customers.

Now when an owner asked Musk about a fix in Autopilot on Twitter last night, the CEO seems to suggest that it would have to wait because the Autopilot team is working on Optimus:

Note, Autopilot/AI team is also working on Optimus and (actually smart) summon/autopark, which have end of month deadlines., The deadline at the end of the month for Optimus is the Tesla AI Day 2, during which the company is expected to unveil its first working prototype of the robot.

Tesla apparently also plans to unveil an updated version of its “Smart Summon” and Autopark features at the vent.  Is Tesla spreading itself too thin with this robot project? Should Tesla focus on delivering on its self-driving promises? Let us know what you think in the comments section below.

Robotic Falcons to Deter Bird Strikes

 Novel Application.  might also deter birds eating key crops. 

Robotic Falcon Designed to Help Prevent Bird Strikes, By CNN, October 28, 2022

Collisions between birds and aircraft, known as bird strikes, can result in aircraft damage, as well as delays and cancellation of flights, costing the International Civil Aviation Organization a reported $1.4 billion each year.

Researchers at the University of Groningen (UG) in the Netherlands have developed a robotic peregrine falcon in a bid to prevent birds from colliding with aircraft.

The ground-controlled RobotFalcon imitates a falcon's movements to ward off birds, using propellers on its wings and a camera on its head to facilitate a "first-person view while steering."  ... 

UG's Rolf Storms said tests near the Dutch city of Workum showed the robot cleared flocks from fields within five minutes of taking off, with half the sites cleared within 70 seconds; a drone only managed to clear 80% of the birds from fields in the same time.  ... 

From CNN

View Full Article    

Whats Your Pacebo?

Good Thoughts. 

What's Your Placebo? (links here) 

By Ryan Bockmon, Stephen Cooper

Communications of the ACM, October 2022, Vol. 65 No. 10, Pages 31-33   10.1145/3528085

A tale of woe: Our institution, like many others, has a high attrition rate in introductory computer science (CS1), our first programming class for majors. We often use the term "DWF rate," as those students who earn Ds or Fs, or who withdraw from the course are ineligible to continue further in taking other CS classes as part of the major. Beyond the DWF rate, students who earn Cs in their CS1 course, while technically allowed to continue taking CS classes, tend to struggle in those later classes. We do offer a pre-CS1 programming class to help students who are not ready to jump directly into CS1. Many researchers have shown students with prior programming experience tend to do better in the initial CS1 course. We have published research on this,3 as have many others.

We were tasked with developing an assessment our department could use to recommend whether students should sign up for our pre-CS1 course or our CS1 course. We decided to use an instrument that combined two recently validated instruments that measured a student's programming ability. One was a computing concepts inventory1 and the second was a programming comprehension inventory.4 ... ' 

AI in Your Ears, Enhancing Speech with Neural 'Clearbuds'

New to me and interesting, out of ACM

AI In Your Ears  By R. Colin Johnson

Commissioned by CACM Staff, August 25, 2022

"ClearBuds" is the code-name of the first "end-to-end hardware-software neural-network based binaural system using wireless synchronized earbuds," according to hardware engineer Maruchi Kim at the University of Washington.

Kim and his colleagues demonstrated a prototype of their speech-enhancing/noise-reducing devices at the ACM International Conference on Mobile Systems, Applications, and Services (ACM MobiSys2022, held in Portland, OR June 27-July 1 ).

The "first" claimed by the researchers is the pairing of binaural (dual) microphones—one in each ear's ClearBud—with two neural networks in an app on a smartphone, resulting in a superior user-experience of voice isolation and noise cancellation during telephone conversations, according to test subjects.

"While neither dual mics nor neural network software is unique or innovative, the combination has value since it reportedly provides an experience that the users liked," said Fan Gang Zeng, a professor of otolaryngology and director of the Hearing and Speech Laboratory at the University of California, Irvine. A researcher in auditory science and technology who was not involved with the research, Zeng added, "Also, there is no technical barrier for others to develop or use the same combo."

To assist other researchers and even commercial telephony equipment providers to use the ClearBud approach, the researchers open-sourced their hardware, software, and neural network architectures. Details are provided in their paper, as well as in their audio demonstrations (which also contain links to the open-source hardware, including the printed circuit board layout, the software code for binaural transmission over Bluetooth, and the code and architectures of the neural networks).

McKinsey Blueprints the Metaverse

 A how and why is presented. 

Your blueprint to building in the metaverse  in McKinsey

October 29, 2022  Is the metaverse is coming for all of us? COVID-19 ushered in a new normal in so many ways—for Ken Wee, chief strategy officer at Activision Blizzard, the pandemic accelerated the degree to which everyone engages in digital experiences. “The total addressable market for interactive entertainment is increasing dramatically,” he told partner Hamza Khan—and that’s just scratching the surface of the metaverse’s true potential. In Metaverse Talks, a new series, Khan and others interview key metaverse players to set out a framework for how organizations can navigate these uncharted waters. For more from Wee and others on why the metaverse is the next generation of the internet, check out the collection below.

Appealing to a broader audience

Exploring the business promise of the metaverse

Building a safer metaverse

Building an open metaverse

Gaming for a cause

More on the metaverse

Digital twins: The foundation of the enterprise metaverse

What is the metaverse? | McKinsey Explainers

Value creation in the metaverse

Probing myth and reality in the metaverse ... ' 

Microsoft: Hackers are Using Open Source Software, Fake Jobs in Phishing Attacks

 Not very new, except in the details. seen this before.  

ACM NEWS

Microsoft: Hackers are Using Open Source Software, Fake Jobs in Phishing Attacks

By ZDNet

September 30, 2022

Microsoft is warning that hackers are using open source software and bogus social media accounts to dupe software engineers and IT support staff with fake job offers that in reality lead to malware attacks.

A phishing-happy hacking crew linked to North Korea's armed forces has been using trojanized open-source apps and LinkedIn recruitment bait to hit tech industry employees, according to threat analysts from Microsoft's advanced persistent threat (APT) research group.

The Microsoft Threat Intelligence Center (MSTIC, pronounced 'Mystic') has seen the group using PuTTY, KiTTY, TightVNC, Sumatra PDF Reader, and the muPDF/Subliminal Recording software installer for these attack since late April, according to MSTIC's blogpost....

A security team at Microsoft-owned LinkedIn said the same actors created fake profiles to impersonate recruiters from companies in the technology, defense, and media entertainment sectors.

From ZDNet

View Full Article  

Saturday, October 29, 2022

Open-Source Fleet Management Tools for Autonomous Mobile Robots

Are I am working in, soon? 

At ROSCon 2022, NVIDIA announced the newest Isaac ROS software release, Developer Preview (DP) 2. This release includes new cloud– and edge-to-robot task...

Open-Source Fleet Management Tools for Autonomous Mobile Robots

NVIDIA Technical Blog by Gerard Andrews / October 19, 2022 at 09:12PM//keep unread//hide

task management and monitoring software for autonomous mobile robot (AMR) fleets, as well as additional features for ROS 2 developers.

NVIDIA Isaac ROS consists of individual packages (GEMs) and complete pipelines (NITROS) for hardware-accelerated performance. In addition to performance improvements, the new release adds the following functionality:

Mission Dispatch and Client: An open-source CPU package to assign and monitor tasks from a fleet management system to the robot. Mission Dispatch is a cloud-native microservice that can be integrated as part of larger fleet management systems.

FreeSpace Segmentation: A hardware-accelerated package for producing a vision AI–based occupancy grid in the proximity of the robot to be used as an input to the navigation stack.

H.264 Video Encode and Decode: Hardware-accelerated packages for compressed video data recording and playback. Video data collection is an important part of training AI perception models. The performance of these new GEMs on the NVIDIA Jetson AGX Orin platform measured at 2x 1080p stereo cameras at 30 fps (>120 fps total), reducing data footprint by ~10x.

Mission Dispatch and Client ... 

Collaborative Creativity in AI

Discovered of late.    Excerpt of article below. Some key issues outlined.

Collaborative creativity in AI

In Nature Machine Intelligence volume 4, page 733 (2022)  

The public release of ‘Stable Diffusion’, a high-quality image generation tool, sets new standards in open-source AI development and raises new questions.

Generative AI tools provide creative inspiration for artists, illustrators and writers, but also for scientists — for example, in the discovery of drugs, materials and even the design of quantum experiments. Powerful capabilities arise in particular with very large neural network models that have billions or even trillions of parameters and are trained on vast amounts of data on the internet. For example, language models such as GPT-3 from OpenAI can generate original text as if written by humans. More recently, such algorithms have been trained on text–image pairs and developed into image generation models. DALL-E (from OpenAI), Imagen (from Google) and Midjourney produce stunning images in any style, from photorealistic portraits and newspaper cartoons to medieval tapestry and much more, given text or other images as prompts.

Illustration produced with the generative AI tool Stable Diffusion. The authors of a News & Views article in this issue, about a bio-inspired robotics approach, sent us a sketch of an idea for an illustration: a robot dog looking into a mirror and seeing a real dog in the reflection. This idea was turned into a Nature-style image to accompany the News & Views article. We tested what Stable Diffusion would come up with when shown the same sketch and given text prompts, and share here one example output image.

This editorial is illustrated with an image generated with a new text-to-image generation tool known as ‘Stable Diffusion’. This model has been making a substantial impact worldwide in the last month, as it is completely open-source and free to use, unlike the aforementioned image generation models. Its public release on 22 August 2022 was accompanied with guidelines and a license that focus on responsible and ethical use and re-use of the model.

Plenty of ethical concerns come to mind with regard to generative models trained on data from the internet, which contains much harmful content and amplifies biases inherent in society. Moreover, generative algorithms could be used to produce unlawful ‘deepfakes’ that are damaging to individuals, and questions arise about copyright and undervaluing the work of professional artists. A concern that was in particular on the mind of OpenAI researchers in 2019 was the potential for malicious use of their language model (then GPT-2) in the form of the large-scale spread of misinformation. As a remedy, they released a smaller version of the model but this caused a backlash in the community. Many were unhappy with the fact that ‘hype’ was generated around GPT-2 while only few had access to the full model, making replication experiments challenging. It was argued that openness and engagement with the community would be a better way to examine any concerns and to promptly work on countermeasures when problems arise. GPT-3 is currently made available to researchers upon request, and for others there is controlled access via an API. Filters have been introduced to tackle misuse and problems with bias. DALL-E, from OpenAI, is available upon request and also has been endowed with various filters.

In an initiative to give the world a high-quality text-to-image generation large neural network model, Stable Diffusion was developed by a collaboration between several groups, including start-up stability.ai, non-profit AI platforms Eleuther AI and LAION, and researchers from the Machine Vision & Learning research group at LMU Munich. The model is built on the latent diffusion model published by the LMU group last year. In the approach, noise is added to a low-dimensional latent representation of example images, and a neural network is trained to revert this diffusion process. The de-noising process is guided by a text prompt. The algorithm was trained on a subset of the LAION-5B image database, using a cluster of 4,000 GPUs.  ...  

AI Model Transferability in Healthcare

Very good look at this issue,  considering potential application. 

AI Model Transferability in Healthcare: A Sociotechnical Perspective

By Nature Machine Intelligence, October 24, 2022  (Was Introduced to Nature Machine Intelligence) 

Predictive model transferability is gaining more attention as healthcare organizations attempt to implement artificial intelligence (AI)-based prediction tools. Although some machine learning (ML)-based models fail when subjected to retrospective validation across institutions and patient populations, technical improvements show promise for addressing this model efficacy problem. To address the engineering challenges, a technical subfield labelled MLOps has emerged.

However, the focus of MLOps on technical transferability may be obscuring a larger set of obstacles to sociotechnical transferability: organizational, social, and individual challenges of deploying models at scale across contexts, whether institutions, teams or individual roles....

To deliver value in healthcare AI and ML models must be integrated not only into technology platforms but also into local human and organizational ecosystems and workflows.

From Nature Machine Intelligence

View Full Article  

Imaginable

 Now reading game designers Jane McGonigal's just published:  " Imaginable: How to see the Future Coming and feel ready for anything - even things that seem impossible today"    Had met her at a meeting some time ago and was impressed.   We were looking at how to integrate games with company process decisions.  I see she works with IFTF  , which we also worked with at the time.

Amazon's description:

  An innovative guide to living gamefully, based on the program that has already helped nearly half a million people achieve remarkable personal growth.

In 2009, internationally renowned game designer Jane McGonigal suffered a severe concussion. Unable to think clearly or work or even get out of bed, she became anxious and depressed, even suicidal. But rather than let herself sink further, she decided to get better by doing what she does best: She turned her recovery process into a resilience-building game. What started as a simple motivational exercise quickly became a set of rules for “post-traumatic growth” that she shared on her blog. These rules led to a digital game and a major research study with the National Institutes of Health. Today nearly half a million people have played SuperBetter to get stronger, happier, and healthier.

But the life-changing ideas behind SuperBetter are much bigger than just one game. In this book, McGonigal reveals a decade’s worth of scientific research into the ways all games - including videogames, sports, and puzzles - change how we respond to stress, challenge, and pain. She explains how we can cultivate new powers of recovery and resilience in everyday life simply by adopting a more “gameful” mind-set. Being gameful means bringing the same psychological strengths we naturally display when we play games - such as optimism, creativity, courage, and determination - to real-world goals.

Drawing on hundreds of studies, McGonigal shows that getting superbetter is as simple as tapping into the three core psychological strengths that games help you build:

Your ability to control your attention, and therefore your thoughts and feelings

Your power to turn anyone into a potential ally, and to strengthen your existing relationships

Your natural capacity to motivate yourself and super-charge your heroic qualities, like willpower, compassion, and determination

SuperBetter contains nearly 100 playful challenges anyone can undertake in order to build these gameful strengths. It includes stories and data from people who have used the SuperBetter method to get stronger in the face of illness, injury, and other major setbacks, as well as to achieve goals like losing weight, running a marathon, and finding a new job.

As inspiring as it is down to earth, and grounded in rigorous research, SuperBetter is a proven game plan for a better life. You’ll never say that something is “just a game” again.  ... ' 

3D Quantum Accelerometer

 Was unaware of such an application:

First 3D quantum accelerometer could let ships navigate without GPS

A device that measures acceleration very precisely using quantum effects could be used for navigation when GPS is unavailable

PHYSICS 28 October 2022  in NewScientist

By Karmela Padavic-Callaghan

A quantum device that can determine its position in three dimensions is more accurate than non-quantum versions. Vehicles could use it to navigate even if GPS stopped working.

One way to keep track of something’s position is with an accelerometer, which is a small device that is found in everything from phones to drones. Accelerometers work by detecting changes in movement and therefore position.

It has been known for decades that quantum effects could be used make more accurate accelerometers, but most …

AI and Intelligence

 Musing on future directions. 

AI's true goal may no longer be intelligence

Some scholars of AI warn that the present technologies may never add up to "true" intelligence or "human" intelligence. But much of the world may not care about that.

Written by Tiernan Ray, Contributing Writer on Oct. 28, 2022

Illustration of a person with their face lifted out as a mask to show the computer inside, and wearing a nametag saying Hi, I'm AI

AI has been rapidly finding industrial applications, such as the use of large language models to automate enterprise IT. Those applications may make the question of actual intelligence moot.

The British mathematician Alan Turing wrote in 1950, "I propose to consider the question, 'Can machines think?'" His inquiry framed the discussion for decades of artificial intelligence research.

For a couple of generations of scientists contemplating AI, the question of whether "true" or "human" intelligence could be achieved was always an important part of the work. 

AI may now be at a turning point where such questions matter less and less to most people. 

The emergence of something called industrial AI in recent years may signal an end to such lofty preoccupations. AI has more capability today than at any time in the 66 years since the term AI was first coined by computer scientist John McCarthy. As a result, the industrialization of AI is shifting the focus from intelligence to achievement.

Also: OpenAI's Dall•E 2 may mean we never need stock photos again

Those achievements are remarkable. They include a system that can predict protein folding, AlphaFold, from Google's DeepMind unit, and the text generation program GPT-3 from startup OpenAI. Both of those programs hold tremendous industrial promise irrespective of whether anyone calls them intelligent. 

Among other things, AlphaFold holds the promise of designing novel forms of proteins, a prospect that has electrified the biology community. GPT-3 is rapidly finding its place as a system that can automate business tasks, such as responding to employee or customer queries in writing without human intervention.

That practical success, driven by a prolific semiconductor field, led by chipmaker Nvidia, seems like it might outstrip the old preoccupation with intelligence. 

In no corner of industrial AI does anyone seem to care whether such programs are going to achieve intelligence. It is as if, in the face of practical achievements that demonstrate obvious worth, the old question, "But is it intelligent?" ceases to matter.  ... ' 

Fairer Search Results

 Ranking, Fairness. 

Fairer Ranking System Diversifies Search Results

Cornell Chronicle

Patricia Waldron,  September 19, 2022

Cornell University researchers Yuta Saito and Thorsten Joachims have developed a fairer ranking system for recommendations that prevents search results from only highlighting a few top hits. Conventional recommender systems try ranking items based on what users want to see, so Saito designed the enhanced ranking system according to the economic principles of "fair division." The researchers used synthetic and real-world data to test the system's feasibility. The system returns viable results that rank items' benefits better than random discovery, does not easily improve items' impact, and does not confer an advantage to any item by switching how it is ranked versus other items in a series of searches. Saito said the framework "can be applied to any type of two-sided ranking system."

Friday, October 28, 2022

Microsoft Admits its Patches Broke OneDrive

I did notice that OneDrive was acting oddly, have used it since my corporate days.    When it was standard for us.     Also it was using my personal pix backup as a means to advertise to me, in a way with little useful flexibility.     It had also started to act strangely beyond that, which made me think it would become a dangerous liability.   Please fix it with safe usefulness as THE most important thing.    Else its out of here.  

It’s not just you — Microsoft admits its patches broke OneDrive

By Alan Truly, October 28, 2022 7:34AM  in DigitalTrends

If you’ve been experiencing OneDrive crashes and error messages, before digging too deep for a solution, note that it might be Microsoft’s fault. Common solutions like restarting, or signing out and back in won’t help because the issue is with the latest Windows 10 update.

Apparently, the problem begins after installing the 22H2 update for Windows 10 that was released on October 18, 2022. Today, Microsoft confirmed that after updating Windows 10, OneDrive might “unexpectedly close,” a nice way to describe a crash. This problem isn’t affecting Windows 11 computers and it’s still possible to use OneDrive via a browser.

If you try to sign out or unlink your account in the Microsoft OneDrive sync app, you’ll get an error message. It will also be impossible to unlink sites or folders from syncing to your device from Microsoft SharePoint or Microsoft Teams sites.

Related

OneDrive is the premier cloud storage solution for Windows computers, a standard feature that Microsoft includes to make backing up and sharing files more convenient. That means it’s a top priority for Microsoft to get OneDrive back up and running on Windows 10. The Windows 10 update that broke the OneDrive sync app also affects OneDrive for Business.

Microsoft’s cloud storage solution is compatible with most operating systems and there are apps available for Android, iPhone, Mac, and Linux. The OneDrive app is working correctly on those devices and also on Windows 11.

Microsoft notes that it’s working on a resolution of the OneDrive sync app problem that was caused by Windows 10 22H2 and expects to have a solution soon. A patch should be available in the coming week.   ... ' 

Singapore, Germany to Mutually Recognize IoT Cybersecurity Labels

IOT security direction particularly important.     Why 

Singapore, Germany to Mutually Recognize IoT Cybersecurity Labels

By ZDNet,  October 27, 2022

Singapore and Germany have agreed to mutually recognize their respective smart computer product cybersecurity rating labels, with Germany the second European company to do so after Finland.

The pact stipulates that products issued with Germany's Federal Office for Information Security (BSI) label would be designated as satisfying Level 2 of the Cyber Security Agency of Singapore (CSA)'s cybersecurity labeling system.

CSA's system rates smart devices into four asterisk-demarcated levels, with each asterisk indicating an additional tier of completed testing and assessment; products rated Levels 2 through 4 must meet "enhanced" cybersecurity mandates, including device and data requirements.

The mutual CSA-BSI recognition would apply to consumer Internet of Things devices like smart TVs, smart toys, health trackers, smart lighting, and smart thermostats.

CSA said it would add more product categories under the bilateral agreement in cooperation with BSI.

The mutual recognition will apply to consumer Internet of Things devices including smart televisions, smart toys, health trackers, smart lighting, and smart thermostats.

Full article:   

Autonomous, Wireless Robots Could Dance on a Human Hair

 Antbots, more on micro robots and their potential uses.

Autonomous, Wireless Robots Could Dance on a Human Hair

September 23, 2022

The Antbots incorporate a photovoltaic cell to accept light as power, a tiny integrated circuit for controlling and directing that power, and a set of hinged legs to scoot itself around.

Cornell University researchers have created wireless robots small enough to sit on a human hair that operate autonomously, using light as a power source.

The Antbots combine a photovoltaic cell, an integrated circuit for controlling and directing power, and hinged legs to provide locomotion.

The researchers manufactured the Antbots' integrated circuits using a 180-nanometer silicon fabrication process.

They said said the Antbots could be employed for environmental cleanup and monitoring, targeted drug delivery, cellular monitoring or stimulation, or microscopic surgery.

From TechCrunch   View Full Article  

Software Tools Aim to Aid 3D Content Creation for Metaverse

Seems useful, if needed. Makes sense out of Adobe. Note apparel example.

Software Tools Aim to Aid 3D Content Creation for Metaverse

By Reuters, October 27, 2022

A set of software tools rolled out by Adobe Inc. aims to facilitate the development three-dimensional (3D) digital objects for marketing campaigns, video games, and the metaverse.

With the 3D Capture tool, photos of real-world objects taken with virtually any camera can be combined to create a 3D digital object.

Adobe's François Cottin gave an example of a shoe seller using the tool to allow users to try on sneakers via augmented reality technology.

Another tool will allow artists editing 3D objects to jump between their desktop computers and manipulating the object using their hands and a virtual reality headset, with a feeling like sculpting clay.

From Reuters

View Full Article   

Apple Forms Security Research Hub

New emphasis on Security Apple Security Research

Apple Launches New Security Research Hub

Apple engineers share technical details about the team's work on memory safety features on the new Apple Security Research site. 

dr_staff_125x125.jpg  Dark Reading Staff

Apple's work on hardening the memory allocator has made it harder for attackers to exploit certain classes of software vulnerabilities on iOS and Mac devices, the company's security engineers wrote on a new website Apple launched to share technical details behind iOS and MacOS security technologies.

The new initiative, Apple Security Research, also offers tools to help security researchers report issues to Apple, get real-time status updates for submitted reports, communicate securely with Apple engineers investigating the issue, and provides information about the Apple Security Bounty program. The intent behind the new security hub is to share with the research community how Apple engineers approach security challenges, and also to invite researcher contributions and feedback.

Memory safety is a key area of focus, especially since memory safety violations are the most widely exploited class of software vulnerabilities. On Apple platforms, improving memory safety includes "finding and fixing vulnerabilities, developing with safe languages, and deploying mitigations at scale," the engineers wrote in a technical post on XNU memory safety.  ... ' 

Batteries Could be 3D Printed

 Evolving battery tech

The future of solid-state batteries could be 3D-printed  in TheVerge

That means batteries could have customized shapes

By LIZZIE PHILIP, Oct 28, 2022, 9:00 AM  

The race to create a solid-state battery that could compete with today’s lithium-ion cells is heating up. Lithium-ion batteries are everywhere: in your phone, car, camera, and more. Since their debut in the 1990s, they’ve become a leader in energy storage. But they have one major flaw: safety. Lithium-ion batteries have a tendency to catch fire, especially when damaged or at high temperatures.

Solid-state batteries replace a flammable liquid electrolyte in lithium-ion batteries with a more stable solid one. They also could have more power, faster charging, and a longer lifespan. Right now, lots of startups are trying to get their first batteries out of the lab and into a factory and hope to prove that solid-state batteries can be commercially viable. 

One of those companies is California-based startup Sakuu, and it’s taking on an even bigger task: 3D-printing these next-gen batteries. Sakuu claims that 3D printing allows it to fit more battery layers in the same amount of space, boosting the capacity of its batteries compared to those made by traditional manufacturing. In theory, the batteries could take on more customized shapes, which could change how batteries are integrated into product design. But the company has yet to 3D-print a full battery using its prototype. Check out our video to learn more about how this new technology could reinvent the way batteries  ... ' 

New X-Ray Technique Could Help Detect Explosives, Tumors

 Shape detection leveraged.  

New X-Ray Technique Could Help Detect Explosives, Tumors

By Adrianna Nine on September 15, 2022 at 1:02 pm 

A new X-ray technique that combines conventional equipment with a deep-learning algorithm might find its way into both security settings and the healthcare industry.

Researchers from the United Kingdom’s University College London (UCL) recognized that X-ray security systems, though good at detecting shapes, weren’t so great at recognizing textures. Identifying textural abnormalities could be the key to locating explosives and other harmful items—especially those hidden away within larger objects. So they set about devising a system that could be paired with existing equipment to detect concerning textures. ... ' 

Steve Nouri of the Forbes Technology Council talks AI

In this episode, Rashmi Mohan hosts Steve Nouri, Founder of AI4Diversity, Founding Member of Hackmakers, and Chief AI Evangelist at Wand. He’s an award-winning technical leader, data scientist, academic, entrepreneur, and global leader on artificial intelligence. Nouri sits on the Forbes Technology Council, is a committee member at the International Organization for Standardization (ISO), and was named ICT Professional of the Year Gold Disruptor in 2019 by the Australian Computer Society (ACS). With more than 1 million followers on LinkedIn, he is one of the most influential voices in AI and Data Science.

The podcast is available in the ACM Learning Center  , where you can also subscribe to an RSS feed as well as download a full transcript of each episode, and on popular podcast platforms including Apple Podcasts, Google Podcasts, Spotify, and Stitcher. ...    From  ACM

Thursday, October 27, 2022

Smelling in VR

A space we examined in retail.

Smelling in VR is Possible with Gaming Technology

Stockholm University (Sweden)

October 12, 2022

Scientists at Sweden's Stockholm and Malmö universities have created a three-dimensionally (3D)-printable "olfactometer" that can generate odors in virtual reality via a gaming controller. The researchers deployed the scent machine in a game where players guess types of wines in a virtual cellar by their aromas. When the players lift virtual glasses, the olfactometer attached to the controller discharges an odor. The olfactometer has four valves that are each connected to a channel, which under the player's control open to a different extent and release unique fragrance mixtures. Said Malmö University's Simon Niedenthal, "The possibility to move on from a passive to a more active sense of smell in the game world paves the way for the development of completely new smell-based game mechanics based on the players' movements and judgments."

Full Article


Thinking Step by Step

Makes sense don't have the full article as yet ...

AIs become smarter if you tell them to think step by step

Artificial intelligence models can outperform humans at tasks AIs normally struggle with if they are told to think a certain way, but it doesn’t help them grasp sarcasm  ... ' 

TECHNOLOGY , 26 October 2022.  By Chris Stokel-Walker  in  NewScientist

Artificial intelligence can get better at tasks if told to think about things in steps

Telling artificial intelligence models to “think” step by step when carrying out a task can improve their performance so much that they can outperform humans at jobs AIs usually struggle with.

Using the phrase “let’s think step by step” to cajole AIs into taking more logical decisions was first suggested in a May study presented at a computational neuroscience conference. Such “chain-of-thought” prompting encourages these models, which include GPT-3, a text-generating AI developed by ... '   

HVIDIA Funds UF For AI

As mentioned formerly, now with a impressive building.  A report from a former school of mine. Now with an impressive building: Malachowsky Hall for Data Science & Information Technology.  Funded by NVIDIA with future plans for a hub there. 

A name that has put UF on the AI map! 

With AI taking center stage in the curriculum across all colleges, it is only fitting that AI take center stage on campus as well. 

AI and other technologies will have a home in the heart of campus starting in 2023 when Malachowsky Hall for Data Science & Information Technology opens. The 263,000-square-foot building, anchored by a gift from Chris Malachowsky (BS ’80) and his company, NVIDIA, will create a hub for advances in computing, communications and cyber-technologies with the potential for profound societal impact. 

The building provides space for researchers and students in medicine, engineering, pharmacy, informatics and others to collaborate across disciplines. In making the gift, Malachowsky noted the possibility for AI and data sciences to be “life-changing” for students and scientists.

#GoGreater #GoGators #NamedSpaces #AIatUF

Microscopic Robots, Walking

 Autonomous movement, brains.

Microscopic robots walk autonomously using simple 'brains'

Robots could one day treat health issues from inside your body.

J. Fingas, @jonfingas, September 22, 2022 9:35 AM

Autonomous microrobot with a simple electronic 'brain'

Noël Heaney/Cornell University

It's long been possible to make extremely small robots, but they usually need some form of direct external control just to operate. Cornell scientists may have solved that problem on a basic level, however. They've created microrobots (no more than 250 micrometers across) with basic electronic "brains" that let them walk autonomously. Two- and six-legged robots move relatively simply, while a four-legged "dogbot" changes speed when an operator sends laser pulses.

The trick was to build a complementary metal-oxide-semiconductor (or CMOS, as computer enthusiasts know it) clock circuit whose signal produces phase-shifted square wave frequencies that set the gait of the robot's platinum-based legs. Photovoltaics control both the legs and the circuit. The design is far from complex at just 1,000 transistors (for context, a GeForce RTX 4090 has 76.3 billion), and it's still large enough that it effectively serves as the robot's body. However, even that is an achievement — the exceptionally low power demands saved Cornell from having to use relatively gigantic photovoltaics.

These inventions are a far cry from the more sophisticated large-scale autonomous robots you see today. They can move forward, but not much else. The researchers see this as just a beginning, though. They believe future microrobots could be crucial to healthcare, where they could perform internal surgery and clean your arteries. Elsewhere, they could detect chemicals and eliminate pollutants. Any such bots are likely years away, but this project suggests they're technically possible.  ... ' 

More: Is AI Becoming Sentient?

I say no, even the term itself is flawed, 

ACM NEWS

Is AI Becoming Sentient?

By Gregory Goth, Commissioned by CACM Staff, October 27, 2022

Insights into human cognition need to become more granular in order to accurately evaluate AI’s capabilities – or to accurately compare them to those of humans, said Konstantinos Voudouris of the U.K.'s University of Cambridge.

There has been more than a little sensational speculation in recent months regarding the ability of artificial intelligence (AI) to attain sentience, but Bruce McNaughton, distinguished professor of neurobiology and behavior at the University of California at Irvine, does not waste much time pondering the possibility.

In fact, McNaughton doesn't ascribe much to the idea of sentience, period. He discourages his students from using the term, as well as "consciousness."

"I discourage them from using the term because as a scientist, I think of the brain as a physical system that obeys the laws of physics," McNaughton said. "We just didn't understand that implementation of the laws of physics could become so incredibly complex through the laws of evolution, and I think most people have only 

Konstantinos Voudouris, a psychologist and graduate student researcher at the Leverhulme Centre for the Future of Intelligence of the U.K.'s University of Cambridge, shares McNaughton's sentiments to a fair degree. "It's beguiling almost to be anthropomorphic about how these systems are behaving," said Voudouris, first author of a recent study that directly compared the cognitive abilities of AI agents and children age 6-10. "It's almost a quality of human psychology to anthropomorphize things, but the psychologist can come in and scientifically evaluate that hypothesis against the many alternatives that exist."

Neither McNaughton nor Voudouris are computer scientists or engineers, yet their recent work on artificial intelligence is emblematic of a surge in multidisciplinary research; the mechanics of human cognition, which have served as the theoretical underpinnings of AI development for decades, are receiving greater attention in the development of AI systems that are subject to ever-greater expectations of what they will be able and need to do — and vice versa.

McNaughton said there is more back-and-forth between how cognitive science and artificial intelligence can affect each other because the principles, the mathematics, and the fundamental approaches of machine learning are the same principles and problems the brain confronts: how to make efficient generalizable knowledge that can be flexible and used in different situations.

"In order to capture the statistical structure of the world as we experience it or as an artificial network experiences it, it takes many, many trials in order to gain a good statistical representation of the domain of the data," he said. "The brain has that problem, and artificial neural networks have that problem."

For instance, he said, cognitive scientists are still searching for how the brain actually does the equivalent of backpropagation if, indeed, it doesn't do backpropagation itself: "That is a central problem in computational neuroscience. But that's where neuroscientists who have some ability to follow the machine learning literature can gain insight from that field."

Jay McClelland, director of Stanford University's Center for Mind, Brain, Computation, and Technology, who has published numerous influential explorations of cognitive science and AI with McNaughton, said he also sees a burgeoning dialogue between machine learning scholars and cognitive scientists.

"We do have a two-way street, in the sense that work from AI is at least leading to ways in which people in neuroscience can see how they can engage in the discussion with other people in other fields," he said. "And computer science can offer hypotheses and alternative ways of thinking about exactly what the brain is doing and how it is solving problems, or raise questions we need to answer as brain scientists."

Lifelong learning in machines and humans

One of the major factors driving this computer science/cognitive science dialogue is the Lifelong Learning Machines (L2M) research program launched by the U.S. Defense Advanced Research Projects Agency (DARPA) in 2017, with University of Massachusetts AI expert Hava Siegelmann as the project's first program manager.

The L2M program's core goal was to create AI systems that could take new data, leverage previously learned information, and learn on the fly. Traditional AI architectures, Siegelmann said, fall far short of that. "If you train a network to separate cats and dogs, then you use the same network to separate elephants from tigers, if you use just regularizers, your system won't be able to separate elephants from cats, because it was never a task that it learned."

Siegelmann convened a cross-discipline pool of computer scientists, neuroscientists, biologists, and others to fundamentally change the depth to which research into machine and human cognition interacted.

"They really stepped outside the box and tried to incorporate a range of ideas and thinking in the field," McNaughton said, "and part of that was a subset of neuroscientists who were also interested in these problems. Suffice to say that connection has been strengthened, at least from the perspective of the interests of the machine learning community in neuroscience."

"I wanted the biologists and neuroscientists to tell me the mechanism of how learning works in the brain," Siegelmann said. "I didn't want them to just tell me it goes from the hippocampus to the cortex. We know that. I wanted them to give me a mechanism in such detail I could actually write equations and program them."

The latest research by McNaughton's lab, published in the Proceedings of the National Academy of Science, hewed closely to Siegelmann's stipulations by addressing a persistent problem in artificial neural networks. Termed catastrophic interference or catastrophic forgetting, it is the rapid loss of previously acquired knowledge if new information is introduced too quickly, essentially because the new information re-weights the network to an extent that the system virtually forgets what it has previously learned. Traditionally, artificial network architectures try to alleviate this by re-introducing everything the system has learned as new information is introduced, but this approach becomes both time- and compute-resource impractical, especially if a system is expected to function successfully on the fly.

McNaughton's group, led by the study's first author, graduate student Rajat Saxena, refined a learning system introduced by McNaughton and McClelland in 2020 called Similarity Weighted Interleaved Learning (SWIL). The SWIL theory suggests that learning in artificial networks can be made more efficient by introducing only a subset of old items that share substantial representational similarity with the new information: "By using such similarity-weighted interleaved learning, artificial neural networks can learn new information rapidly with a similar accuracy level and minimal interference, while using a much smaller number of old items presented per epoch," the group concluded.

In their original paper, McNaughton, McClelland, and Andrew Lampinen concluded SWIL performed similarly to networks that interleaved every old item with the new ones to be learned, but used 40% fewer items. They did not find, though, that it scaled beyond a simple neural network.

The latest paper successfully scaled SWIL to work on traditional classification datasets (Fashion-MNIST, CIFAR10, and CIFAR100) as well or better than existing schemes such as Fully Interleaved Learning (FIL), Focused Learning (FoL), and Equally Weighted Interleaved Learning (EqWIL). The team concluded that SWIL's future, at least in terms of AI, probably lays in complementing other learning techniques, such as generative replay or elastic weight consolidation. And, while McNaughton called the latest SWIL research the "evolution" of a breakthrough concept rather than a breakthrough in itself, he did say it stimulated questions about human cognition.  .... ' 

Deep Learning Makes X-Ray CT Inspection of 3D-Printed Parts Faster, More Accurate

 Inspection applications for Deep Learning, 

Deep Learning Makes X-Ray CT Inspection of 3D-Printed Parts Faster, More Accurate

Oak Ridge National Laboratory

S. Heather Duncan, October 14, 2022

Researchers at the U.S. Department of Energy (DOE) Oak Ridge National Laboratory (ORNL) have developed a deep-learning framework that uses X-ray computed tomography (CT) to improve the speed and accuracy of inspecting three-dimensionally (3D) printed metal parts. The framework is being incorporated into software commercial partner ZEISS uses in its equipment at DOE’s Manufacturing Demonstration Facility at ORNL, where companies get help perfecting their 3D printing. ZEISS's Pradeep Bhattad said, "With this, we can inspect every single part coming out of 3D-printing machines. Currently CT is limited to prototyping. But this one tool can propel additive manufacturing (3D printing) toward industrialization."

Wednesday, October 26, 2022

AI in Medicine Is Overhyped?

 Caution in implementation.

AI in Medicine Is Overhyped

By Scientific American, October 25, 2022

Doctor wears virtual reality goggles depicting futuristic healthcare scenario.

Adopting a more holistic approach to developing and testing clinical AI models will lead to more nuanced discussions about how well these models can work and their limitations.

Mistakes by artificial intelligence (AI) models that support doctors’ clinical decisions can mean life or death. Therefore, it is critical that we understand how well these models work before deploying them. Published reports of this technology currently paint a too-optimistic picture of its accuracy, and the scientific papers detailing such advances may become foundations for new companies, new investments and lines of research, and large-scale implementations in hospital systems.

However, in most cases, the technology is not ready for deployment. Why? As researchers feed data into AI models, the models are expected to become more accurate, or at least not get worse. However, our work and the work of others has identified the opposite, where the reported accuracy in published models decreases with increasing dataset size.

From Scientific American

View Full Article   (May Require Paid Registration)

Are We Getting Closer to Artificial Life?

Implications for Health Care?  Health-Technical. 

In Nature:  Article, Published: 14 September 2022

Living material assembly of bacteriogenic protocells

Can Xu, Nicolas Martin, Mei Li & Stephen Mann 

Nature volume 609, pages 1029–1037 (2022)   Cite this article

Abstract

Advancing the spontaneous bottom-up construction of artificial cells with high organizational complexity and diverse functionality remains an unresolved issue at the interface between living and non-living matter1,2,3,4. Here, to address this challenge, we developed a living material assembly process based on the capture and on-site processing of spatially segregated bacterial colonies within individual coacervate microdroplets for the endogenous construction of membrane-bounded, molecularly crowded, and compositionally, structurally and morphologically complex synthetic cells. The bacteriogenic protocells inherit diverse biological components, exhibit multifunctional cytomimetic properties and can be endogenously remodelled to include a spatially partitioned DNA–histone nucleus-like condensate, membranized water vacuoles and a three-dimensional network of F-actin proto-cytoskeletal filaments. The ensemble is biochemically energized by ATP production derived from implanted live Escherichia coli cells to produce a cellular bionic system with amoeba-like external morphology and integrated life-like properties. Our results demonstrate a bacteriogenic strategy for the bottom-up construction of functional protoliving microdevices and provide opportunities for the fabrication of new synthetic cell modules and augmented living/synthetic cell constructs with potential applications in engineered synthetic biology and biotechnology.  ... ' 

Drones Work Like Bees

ACM TECHNEWS

3D Printing Drones Work Like Bees

By Imperial College London (U.K.)

September 23, 2022An international team of scientists led by the U.K.'s Imperial College London (ICL) and the Swiss Federal Laboratories for Materials Science and Technology (EMPA) developed a fleet of three-dimensional (3D) printing drones that can build and repair structures in flight.

The bee-inspired Aerial Additive Manufacturing (Aerial-AM) fleet operates cooperatively from a blueprint, adjusting on the fly.

The fleet includes material-depositing BuilDrones and quality-controlling ScanDrones, as well as a controller that monitors their progress and can intervene based on real-time feedback from the drones.

Said ICL's Mirko Kovac, "We believe our fleet of drones could help reduce the costs and risks of construction in the future, compared to traditional manual methods."

From Imperial College London (U.K.)  


Measuring Muscle Movement with Magnets

https://www.youtube.com/watch?v=sdDeRgAUsWo

MIT Media Lab,  56.1K subscribers

MagnetoMicrometry

Nearly all movement in the animal kingdom is driven by muscles, but can we more accurately measure how our muscles move during natural activity? Researchers at MIT Media Lab have validated that a strategy for tracking muscle movement can do that — track muscle motion outside the lab environment with through the use of implanted magnetic beads. These particular magnetic beads are also biocompatible, necessary for applying this technology toward improved control over prostheses and exoskeletons.... ' 

Researchers Change Trajectory of Drones

 Better speeds and paths for drones via optimization.

Researchers Change Trajectory of Drones for the Better

University of Illinois at Urbana-Champaign

Aaron Seidlitz, October 20, 2022

The University of Illinois at Urbana-Champaign's Kris Hauser, Gao Tang, and Weidong Sun have improved unmanned aerial vehicles' ability to navigate around obstacles via multi-level optimization. Said Hauser, "The basic idea is that you can separately optimize the shape of the path and the speed along it, but since how fast you can fly depends on the shape of the path, these two components are intimately coupled. Our novel formulation treats this as a bilevel optimization problem, in which the speed optimization is nested inside the shape optimization as an inner subproblem." Hauser said experiments proved the method "is faster and significantly more numerically stable than optimizing both the shape and the speed simultaneously." .... '

Tuesday, October 25, 2022

Duke Energy Smart Healing Grids

Been following this for some time, impressive results so far.  Much more to be done. 



Smart, self-healing technology in Florida helped to automatically restore more than 160,000 customer outages during powerful storm

ST. PETERSBURG, Fla., Oct. 24, 2022 /PRNewswire/ -- As Hurricane Ian made its way across Florida, Duke Energy's grid improvements were already on the job helping to combat power outages from the storm.

Smart, self-healing technology helped to automatically restore more than 160,000 customer outages and saved nearly 3.3 million hours (nearly 200 million minutes) of total lost outage time.
"Hurricane Ian is a strong reminder of the importance of grid hardening and storm preparedness to help keep the lights on for our customers," said Melissa Seixas, Duke Energy Florida state president. "Self-healing technology is just one of many grid improvements that Duke Energy is making to avoid outages, restore service faster and increase reliability for our customers."

Much like the GPS in your car can identify an accident ahead and reroute you around the incident to keep you on your way, self-healing technology is like a GPS for the grid. The technology can quickly identify power outages and alternate energy pathways to restore service faster for customers when an outage occurs.

Additionally, self-healing technology provides a smart tool to assist crews in the field with power restoration after a major storm like Ian, helping reduce outage impacts and freeing up resources t help restore power in other locations.

Three days after Hurricane Ian exited the state, Duke Energy Florida wrapped up restoration of approximately 1 million customers. This progress enabled the company to deploy more than 550 Duke Energy workers from throughout Florida, as well as contractors from across the country, to help restore power for Lee County Electric Cooperative customers.

Crews worked in Cape Coral and Pine Island, one of the hardest-hit areas in the storm's path and completed power restoration for the majority of customers on Pine Island within approximately one week after arriving to the island.

Prior to Ian in 2022, smart, self-healing technology had helped avoid nearly 250,000 extended customer outages in Florida, saving around 285,000 hours (17.1 million minutes) of total lost outage time.
Duke Energy currently serves around 59% of customers in Florida with self-healing capabilities on its main power distribution lines, with a goal of serving around 80% over the next few years.
Self-healing technology video.

Self-healing grid system animation video.

Duke Energy Florida

Duke Energy Florida, a subsidiary of Duke Energy, owns 10,300 megawatts of energy capacity, supplying electricity to 1.9 million residential, commercial and industrial customers across a 13,000-square-mile service area in Florida.

Duke Energy (NYSE: DUK), a Fortune 150 company headquartered in Charlotte, N.C., is one of America's largest energy holding companies. Its electric utilities serve 8.2 million customers in North Carolina, South Carolina, Florida, Indiana, Ohio and Kentucky, and collectively own 50,000 megawatts of energy capacity. Its natural gas unit serves 1.6 million customers in North Carolina, South Carolina, Tennessee, Ohio and Kentucky. The company employs 28,000 people.  ... ' 

Further Upheld: AIs Cant Hold Patents

Recall we examined this for a Govt client, her some additional implications.

AI News

US court upholds ruling that AIs can’t be patent holders

By Ryan Daws | October 21, 2022 | TechForge Media

Categories: Artificial Intelligence, Ethics & Society, Legislation & Government,

The US Court of Appeals has upheld previous rulings that AIs cannot hold patents for inventions.

AIs are increasingly being used to make new discoveries but, under most patent laws, a human must be listed as the patent holder for inventions.

Dr Stephen Thaler created a device called DABUS that consists of neural networks and has been used to invent an emergency warning light, a food container that improves grip and heat transfer, and more. Thaler believes that AIs should be patent holders and has launched numerous cases in at least 15 countries to argue the case.

So far, the cases in the UK, US, and New Zealand have all been rejected.

In August 2021, an Australian federal court ruled that AI systems can be credited as inventors. However, in April 2022 that decision was reversed.

Ryan Abbott, a professor at the University of Surrey, filed the applications in Australia on behalf of Dr Thaler., “If I teach my PhD student and they go on to make a final complex idea, that doesn’t make me an inventor on their patent, so it shouldn’t with a machine,” Abbott told the Wall Street Journal in 2019.

In the US, the case was submitted by the Artificial Inventor Project (AIP) which argued that Thaler has no experience in consumer goods and would not have created them himself. Once again, it was rejected.

“This is a case in which the question of statutory interpretation begins and ends with the plain meaning of the text … [T]here is no ambiguity: the Patent Act requires that inventors must be natural persons; that is, human beings,” wrote Judge Leonard P. Stark in the court’s ruling.

At the time, Thaler and Abbott said they’d continue appealing the decisions.

“It ignores the purpose of the Patent Act and the outcome that AI-generated inventions are now unpatentable in the United States,” Abbott told Bloomberg Law.

“That is an outcome with real negative social consequences.”

While the appeal rejections haven’t deterred Thaler or Abbott, most courts appear set on maintaining the principle that humans can only be listed as inventors. The primary argument of the courts is that no AI is creating a new invention by itself, it’s a tool that’s ultimately acting on the instruction of a human. Until we achieve general artificial intelligence, it’s a position that’s hard to dispute.

Of the latest rejection, Thaler is requesting a rehearing at the Federal Circuit level.   ... ' 

YouTube's Dislike Button Rarely Shifts Recommendations

Been following feedback aspects of YouTube

YouTube's Dislike Button Rarely Shifts Recommendations

By The New York Times, September 21, 2022

Mozilla Foundation researchers found pressing the "dislike" button on a YouTube video does little to improve the videos chosen by the YouTube platform's recommender algorithm.

The researchers analyzed over 567 million YouTube video recommendations, and used Mozilla's RegretReporter tool to examine the recommendation algorithm. They found use of the dislike button reduced similar, unwanted recommendations by just 12%.

Selecting "Don't recommend channel" was 43% effective in reducing unwanted recommendations, while pressing "not interested" was shown to be 11% effective, and removing videos from the user’s watch history was 29% effective.

Mozilla's Jesse McCrosky said YouTube should incorporate greater transparency, and should provide users with more control over what they see.

From The New York Times    

Quantum Computing Modalities Blog

 This blog was brought to attention, by Russ Fein.  examining further.   Mention of companies involved.  How up to date is as yet unclear. comments

Blog by Russ Fein:   Useful Links: 

https://quantumtech.blog/home/

https://quantumtech.blog/blog/

https://quantumtech.blog/2021/08/08/example-post-3/

https://quantumtech.blog/links/

        Quantum Computing Modalities – A Qubit Primer Revisited

Posted by Russ Fein , October 20, 2022, Posted in NISQ, Quantum Advantage, Quantum Computers, qubitsTags: Quantum Computing, qubits

In December 2021, in an early iteration of this Blog, I described the various qubit modalities in use by some of the Quantum Computing (QC) hardware players.  A lot has happened since that post, so I thought it would be constructive to revisit the topic.

When that earlier post was published (click here if interested in reviewing), it described 10 leading quantum hardware companies focusing on four core qubit types (superconducting, trapped ions, photonics and quantum dots).  Today there are dozens of quantum hardware companies, a few additional common modalities (notably neutral atoms) and significant advances made across the spectrum.

Qubit Dynamics

While many articles describing and comparing QCs focus on the number of qubits, this core number belies the complexity in comparing actual QC performance due to additional limitations described below.  Qubit count is the equivalent of only using horsepower to describe a car.  While horsepower is an important metric, most car buyers are equally if not more focused on comfort, handling, fuel economy, styling, etc.  Some effort has been made to “consolidate” these variables for QC into a single performance metric (such as Quantum Volume, CLOPS (circuit layer operations per second) or QED-C’s Benchmarks), although no single measurement has yet been adopted by the broad QC ecosystem.  For the casual reader, I’d caution you to not focus too much on the number of qubits a given QC has.  While “more is better” is generally a useful mantra, as you’ll see below, it is not that simple.

As you may know or recall, placing qubits in a superposition (both “0” and “1” at the same time) and entangling multiple qubits where one is dependent on the status of the other (entanglement) are two fundamental quantum properties which help empower Quantum Computers and allow them to perform certain calculations that can’t easily be executed on traditional computers.  Before we review the various types of qubits (i.e., quantum hardware platforms), it may be helpful to summarize some of the limitations faced when placing qubits in superposition and/or entangling multiple qubits, and discuss the key metrics used to measure these properties.   ... ' 

Monday, October 24, 2022

Shuffling Cards

 Some time ago when doing simulation projects, we discovered some problems in code based random numbers.    We even tested them vs some physical machines.  To feed analytical simulation ofprocess.  Humans are well known to be terrible generators of randomness.  The article below points out that the problem is yet to be completely solved. 

On the Randomness of Automatic Card Shufflers   by Bruce Schneier

Many years ago, Matt Blaze and I talked about getting our hands on a casino-grade automatic shuffler and looking for vulnerabilities. We never did it—I remember that we didn’t even try very hard—but this article  shows that we probably would have found non-random properties:

…the executives had recently discovered that one of their machines had been hacked by a gang of hustlers. The gang used a hidden video camera to record the workings of the card shuffler through a glass window. The images, transmitted to an accomplice outside in the casino parking lot, were played back in slow motion to figure out the sequence of cards in the deck, which was then communicated back to the gamblers inside. The casino lost millions of dollars before the gang were finally caught. ... ' 

Its interesting that the comments point to further experience and research, including using a magician to detect regularity in in choice!


Tech Advances or Creeping Madness?

 Soo to be drones Everywhere? 

Casino Developers Want to Fill Times Square With Surveillance Drones

"If the city makes this high-stakes bet on casino surveillance, I worry they’ll gamble away the future of our public streets," said one privacy expert.   By Thomas Germain in Gizmodo

There’s great news if you’re a fan of gambling and flying, spying robots.

Times Square, fifth circle of hell and home of the world’s dumbest ads, could be the site of New York City’s next casino. Caesars Entertainment has teamed up with developer SL Green Realty Corporation in a bid for a casino license, and they’re trying to sweeten the pot with a proposal that is going to drive you wild if you hate privacy.

According to a document obtained by the New York Times:

If You Live in One of These Cities, There's Probably a Security Camera Pointed at Your Face Right Now

In their letter seeking support for the casino, SL Green and Caesars said that gambling revenues could be used to more than double the number of “public safety officers” in Times Square and to deploy surveillance drones.

The letter said a new casino would result in more than 50 new artificial intelligence camera systems “strategically placed throughout Times Square, each capable of monitoring 85,000+ people per day.” ... ' 

Chula Chatbot

ACM CAREERS

Chula Chatbot Serves As an Education and Career Coach

By Chulalongkorn University, September 26, 2022

A chatbot developed by lecturers at Chulalongkorn University is designed to serve as an education and career coach for learners and students.

The EmpowerMe chatbot "is like having a student counselor or personal coach who can offer advice on how to develop career skills in the digital age so that students can find the job for which they are well suited," says Professor Jintavee Khlaisang. It will also suggest lessons to help a user develop the types of skills needed in a particular career. It assesses a learner's performance and motivates them with rewards in the form of medals.  

The application received a gold medal award from the 2021 Seoul International Invention Fair. It has been patented for user-interface design.

From Chulalongkorn University

View Full Article