State of Data Onboarding

 Newly brought to my attention, surveys and analysis, of data use. 

Flatfile

Our vision could be summed up no better than by the great poet, D.H. Lawrence:

For God’s sake, let us be humans.  Not monkeys minding machines.

Humans vs. machines

Whether you’re decreasing inefficiencies for postal workers, helping donors connect with the causes they care about, or building the next financial analysis tool - you can do little without data. The problem: getting that data, figuring out what it means, and molding it into something useful is not just a challenge, it’s a fight.

There’s a reason the term “data wrangling” exists. And if you already know what the acronyms MDM, ACID, UIMA, or ETL mean... we’re deeply sorry.

Report:  The State of  Data Onboarding

Data onboarding, a critical stage in the customer onboarding process where data is migrated into a new software application, continues to be a serious challenge for companies - no matter the size or industry. The 2020 State of Data Onboarding is the first report of its kind to shed light on data onboarding: the good, the bad, and the ugly.

We surveyed more than 100 companies and gathered additional data from more than 5,000 respondents via Twitter to learn more about how organizations today are managing the data onboarding process. ... '

Updates on Echo Auto

Have been a user of Echo Auto in the car for some time.  There were early glitches, but it is working well  now.  They are just announcing some new capabilities:

Echo Auto, Alexa in the Car

Alexa is always getting smarter, and we are excited to introduce some new features for Echo Auto to make your drive even more of a breeze.

See and do more with Auto Mode

Try Auto Mode in the Alexa App to turn your smartphone into a driver-friendly display that complements your Echo Auto. Get useful info at a glance, see what song is playing, and use easy-to-tap shortcuts to quickly access your favorite places, people, and content. To turn on Auto Mode, just press the push notification that appears after your phone connects with Echo Auto at the start of each drive. Learn more.  . 

Let Alexa help kick off your commute

Try asking, “Alexa, start my commute,” at the beginning of each drive to get the day’s weather, traffic to work, and your favorite content to keep you entertained. The first time you try it, Alexa will ask you to enable the routine. You can customize your commute in the routines section of the Alexa app to make it most useful for you.

Pay for gas with just your voice

Now you can use your Echo Auto to conveniently pay for gas at Exxon and Mobil stations nationwide. No need to swipe your card, enter your zip code, or touch the keypad the next time you fuel up your car. Secured by Amazon Pay, it's simple and safe. No additional sign-up required. Just pull up to the pump and say, "Alexa, pay for gas." Learn more.   

Things to remember for the best experience:

NIST: Quantum Bits Can be Self Correcting?

Could be a really big thing,  depending on the context and consequences.  

Error-Prone Quantum Bits Could Correct Themselves, NIST Physicists Show

National Institute of Standards and Technology

Physicists at the U.S. National Institute of Standards and Technology (NIST), the University of Maryland, and the California Institute of Technology have developed an approach that could be used to design self-correcting quantum memory switches. The researchers experimented with a photonic cavity resonator and found that constantly refreshing the supply of photons in the cavity enables the qubit's quantum information to withstand certain amounts and types of noise from the surrounding environment. The new approach accounts for the leakage of photons to the environment. Simon Lieu of the Joint Quantum Institute and the Joint Center for Quantum Information and Computer Science said, "It’s like adding fresh water. Any time the information gets contaminated, the fact that you're pushing in water and cleaning out your pipes dynamically keeps it resistant to damage. This overall configuration is what keeps its steady state strong."

TaintDroid: Information Flow on Smartphones

Brought to my attention as part of a broader study: 

 TaintDroid: an information-flow tracking system for realtime privacy monitoring on smartphones

Authors:    William Enck, Peter Gilbert, Byung-Gon Chun, Landon P. Cox, Jaeyeon Jung, Patrick McDaniel, and Anmol N. Sheth. For their paper TaintDroid: An Information-Flow Tracking System for Realtime Privacy Monitoring on Smartphones. 

In OSDI’10: Proceedings of the 9th USENIX conference on Operating systems design and implementation. “This paper was instrumental in demonstrating that taint tracking could be made both efficient and fine-grained. For unmodified smartphone applications, with minimal monitoring overhead, the authors found dozens of potential leaks of sensitive and private information. This work sparked an important research agenda on smartphone privacy that continues to this day.”

Publication:OSDI'10: Proceedings of the 9th USENIX conference on Operating systems design and implementationOctober 2010 Pages 393–407

OSDI'10: Proceedings of the 9th USENIX conference on Operating systems design and implementation

TaintDroid: an information-flow tracking system for realtime privacy monitoring on smartphones    Pages 393–407

ABSTRACT

Today's smartphone operating systems frequently fail to provide users with adequate control over and visibility into how third-party applications use their private data. We address these shortcomings with TaintDroid, an efficient, system-wide dynamic taint tracking and analysis system capable of simultaneously tracking multiple sources of sensitive data. TaintDroid provides realtime analysis by leveraging Android's virtualized execution environment. TaintDroid incurs only 14% performance overhead on a CPU-bound micro-benchmark and imposes negligible overhead on interactive third-party applications. Using TaintDroid to monitor the behavior of 30 popular third-party Android applications, we found 68 instances of potential misuse of users' private information across 20 applications. Monitoring sensitive data with TaintDroid provides informed use of third-party applications for phone users and valuable input for smartphone security service firms seeking to identify misbehaving applications.  ... '

Training AI in Four Bits

Its happening now.   Consider too that smartphones can be the source of much data that will be used to train such systems.

Tiny four-bit computers are now all you need to train AI  in TechnologyReview

Powerful neural networks could soon train on smartphones with dramatically faster speeds and less energy.   by Karen Hao    ... '

Mechanical Turk and AI: Invisible Workers

We used Mechanical Turk for a number of small projects, especially for getting initial data that was reasonable for first order testing.  Never for later predictive analysis.   Our look at it was that they were reasonably paid.    Had not seen much about MT lately, but based on this is still in considerable use.

AI needs to face up to its invisible-worker problem

Machine-learning models are trained by low-paid online gig workers. They’re not going away—but we can change the way they work, says Saiph Savage.  In Technology Review

SAIPH SAVAGE

Many of the most successful and widely used machine-learning models are trained with the help of thousands of low-paid gig workers. Millions of people around the world earn money on platforms like Amazon Mechanical Turk, which allow companies and researchers to outsource small tasks to online crowdworkers. According to one estimate, more than a million people in the US alone earn money each month by doing work on these platforms. Around 250,000 of them earn at least three-quarters of their income this way. But even though many work for some of the richest AI labs in the world, they are paid below minimum wage and given no opportunities to develop their skills. 

Saiph Savage is the director of the human-computer interaction lab at West Virginia University, where she works on civic technology, focusing on issues such as fighting disinformation and helping gig workers improve their working conditions. This week she gave an invited talk at NeurIPS, one of the world’s biggest AI conferences, titled “A future of work for the invisible workers in AI.” I talked to Savage on Zoom the day before she gave her talk.   ... " 

Predicting Railway Delays

 Interesting application.  Consider data availability, accuracy and variability of prediction.

Predicting British railway delays using artificial intelligence  by University of Illinois at Urbana-Champaign

Over the past 20 years, the number of passengers traveling on British train networks has almost doubled to 1.7 billion annually. With numbers like that it's clear how much people rely on rail service in Great Britain, and how many disgruntled patrons there would be when delays occur. A recent study from the University of Illinois Urbana-Champaign used real British Railway data and an artificial intelligence model to improve the ability to predict delays in railway networks.

"We wanted to explore this problem using our experience with graph neural networks," said Huy Tran, an aerospace engineering faculty member at UIUC. "These are a specific class of artificial intelligence models that focus on data modeled as a graph, where a set of nodes are connected by edges."  ... " 

Google AI Describes AutoML for Time Series

 Most our careers in the big enterprise involved working with time series.   Sales, Shipments delivered, Advertising Dollars,  marketing spends ... forecast plans and predictions.   A favorite quote was 'the forecast is wrong', but how wrong?   And Why?  And what are the risks involved?  So if we could do forecasts better, more data and intelligence based?  How might we do it?     

Using AutoML for Time Series Forecasting      In the GoogleBlog.

Friday, December 4, 2020

Posted by Chen Liang and Yifeng Lu, Software Engineers, Google Research, Brain Team

Time series forecasting is an important research area for machine learning (ML), particularly where accurate forecasting is critical, including several industries such as retail, supply chain, energy, finance, etc. For example, in the consumer goods domain, improving the accuracy of demand forecasting by 10-20% can reduce inventory by 5% and increase revenue by 2-3%. Current ML-based forecasting solutions are usually built by experts and require significant manual effort, including model construction, feature engineering and hyper-parameter tuning. However, such expertise may not be broadly available, which can limit the benefits of applying ML towards time series forecasting challenges.

To address this, automated machine learning (AutoML) is an approach that makes ML more widely accessible by automating the process of creating ML models, and has recently accelerated both ML research and the application of ML to real-world problems. For example, the initial work on neural architecture search enabled breakthroughs in computer vision, such as NasNet, AmoebaNet, and EfficientNet, and in natural language processing, such as Evolved Transformer. More recently, AutoML has also been applied to tabular data.

Today we introduce a scalable end-to-end AutoML solution for time series forecasting, which meets three key criteria:

Today we introduce a scalable end-to-end AutoML solution for time series forecasting, which meets three key criteria:

Fully automated: The solution takes in data as input, and produces a servable TensorFlow model as output with no human intervention.

Generic: The solution works for most time series forecasting tasks and automatically searches for the best model configuration for each task.

High-quality: The produced models have competitive quality compared to those manually crafted for specific tasks.

We demonstrate the success of this approach through participation in the M5 forecasting competition, where this AutoML solution achieved competitive performance against hand-crafted models with moderate compute cost... .' 

Softbank Sells Controlling Stake in Boston Dynamics to Hyundai

Interesting development.   In TechExplore.  Had not recalled their connection with Google.

SoftBank sells controlling stake in Boston Dynamics to Hyundai

Japan's SoftBank Group will sell an 80 percent stake in robotics firm Boston Dynamics to Hyundai, the trio said Friday, in a deal that values the US company at $1.1 billion. 

Boston Dynamics has drawn huge attention with viral videos of its humanoid and dog-like robots, whose uncanny movements and impressive tricks have helped stoke fears that androids could one day become a threat to humans.

The engineering firm was founded in 1992 and bought in 2013 by Google, which sold the company on to SoftBank three years ago.

The Japanese conglomerate will keep a 20 percent stake through one of its affiliates and will work with South Korea's Hyundai to "propel development and commercialisation of advanced robots", the companies said.  ... " 

On IBM Factsheets

 On IBM Factsheets, their overview:   https://aifs360.mybluemix.net/introduction  Looking more closely.

Why IBM’s AI Fact Sheets should be the industry standard  in thenextweb

Every once in awhile an idea comes along that’s so good it makes you wonder why it took so long for someone to think of it. IBM’s AI Fact Sheets is one of those ideas.

AI Fact Sheets are a lot like packaged food nutrition labels. They contain information about an AI model’s development, capabilities, benchmark performance, and more.

Big Blue today announced its plans to “commercialize key automated documentation capabilities from IBM Research’s AI Factsheets methodology into Watson Studio in Cloud Pak for Data throughout 2021.”

In other words: businesses and developers using Watson Studio in Cloud Pak for Data will soon have access to an automated AI Fact Sheets tool to create transparency and info reports.  The tool would generate most, if not all, of the AI Fact Sheet’s information automatically.

Per IBM:

The goal of the FactSheet project is to foster trust in AI by increasing transparency an increased understanding of how AI was created and deployed and enabling governance the ability to control how AI is created and deployed. Increased transparency provides information for AI consumers to better understand how the AI model a program component that is generated by learning patterns in training data to make predictions on new data, such as a loan application. or service an executable program, deployed behind an API, that allows it to respond to program requests from other programs or services was created. This allows a consumer of the model to determine if it is appropriate for their situation.

Quick take: We love this idea. While it’s a bit more complex than we can get into in this article (research paper here), the bottom line is that anything that standardizes transparency in machine learning models is a good thing.

This comes as part of IBM’s “AI Governance” initiative along with new AI consultation services. The company also announced several interesting new Watson capabilities.   ....'

Learning Better with Shape-Shifting

 Like the example of modifying objects to address specific elements of training.   Unusual application, uses outside of sports?

Better Learning with Shape-Shifting Objects

MIT News, Adam Conner-Simons

Shape-shifting objects that can help users improve their skills is an area of investigation for Massachusetts Institute of Technology (MIT) researchers, who conceived of a basketball hoop that trains players more effectively by shrinking and raising to help them make shots more consistently. Experiments demonstrated that training on the auto-adaptive hoop improved player performance more than using a static hoop or the manually-adaptive mode. Autodesk's Fraser Anderson said, "You don't have to rely on your own sense of whether or not you've mastered a skill: the system can do that and take out the self-doubt, overconfidence, or guesswork."   ... '

IBM Delivers New AI Capabilities

New developments from IBM, and based on this it seems they are attempting to make AI easier to use and deliver.   Notably things like delivering FAQ is the simplest thing you might want to provide information quickly in a conversation.    And Explain-ability to needed convince.  Plus the ability to translate existing documents into usable intelligence.  All things we did with early AI efforts. so I  see this as good direction for Watson.

IBM announces new AI language, explainability, and automation services

Kyle Wiggers  @Kyle_L_Wiggers   In Venturebeat December 9, 2020 5:00 AM

During IBM’s virtual AI Summit this week, the company announced updates across its Watson family of products in the areas of language, explainability, and workplace automation. A new feature called Reading Comprehension surfaces answers from databases of enterprise documents in response to natural language questions, assigning a confidence score to each response. A novel module in Watson Assistant called FAQ Extraction automatically generates question-and-answer documents. And AI Factsheets automatically captures key facts on a machine learning model’s performance and generates reports to “foster transparency and ensure compliance.”

According to IBM, Reading Comprehension, which was built atop a top-performing question-answering system from IBM Research, is intended to help identify more precise answers in response to queries referring to business documents. Reading Comprehension provides scores that indicate how confident the system is in each answer and is currently in beta in IBM’s AI-powered search service Watson Discovery.    ... "

Coronavirus Apps Show Promise but Prove a Tough Sell

 Good to see an experiment of this type, at very least to see how people react to permitting monitoring , and what it takes for them to trust it and its use.

Coronavirus Apps Show Promise but Prove a Tough Sell

The New York Times, Jennifer Valentino-DeVries

Despite pilot studies demonstrating that smartphone applications can slow Covid-19 transmission, buy-in from people and states is lacking. Apple and Google's exposure-notification apps respect privacy by not tracking user locations, using Bluetooth to detect which phones have been within several feet of one another for more than a few minutes. When a user receives a positive test result, the local health system supplies code via email, text message, or phone call to enter into the app, alerting anyone who was in proximity while the person was contagious. A pilot program at the University of Arizona offered what may be the first example of an app slowing transmission; researchers estimated this fall the app sent alerts for up to 12% of transmissions. Yet such apps are only available in about a third of U.S. states, hampered by privacy issues, little awareness or interest, poor access to quick testing, and a hodgepodge of government health authorities.  ... "

Updating AI Product Performance

Nicely done piece from NVIDIA Developers.

Updating AI Product Performance from Throughput to Time-To-Solution

By Shar Narasimhan | November 23, 2020  Tags: data center, Machine Learning and AI, MLPerf, NGC

Data scientists and researchers work toward solving the grand challenges of humanity with AI projects such as developing autonomous cars or nuclear fusion energy research. They depend on powerful, high-performance AI platforms as essential tools to conduct their work. Even enterprise-grade AI implementation efforts—adding intelligent video analytics to existing video camera streams or image classification picture searches or adding conversational AI to call center support—require incredibly accurate AI models across diverse network types that can deliver meaningful results, at high throughput and low latency.

That’s the key. To be useful and productive, the model must not only have high throughput, but it also must make correct predictions at that high throughput. In everyday terms, it’s just not useful to advertise a sports car that has a top speed over 200 mph, if it can only maintain that speed for a few seconds and can never complete a journey from point A to point B.

If you’re driving from Los Angeles to New York, you want a vehicle that can speedily and reliably make the trip from start to finish. Having a fast car that sputters out in the desert or mountains shortly after setting out isn’t the type of platform to depend on. Even if you’re an avid skier who’s brought your skis with you, getting stranded in the wilderness in the Rockies at the height of winter is going to get challenging quickly.

The next time you hear a slick promotion for a sports car that has a scorching top speed, hold onto your wallet for a second and ask yourself, “Will this get me to my destination or leave me stranded in the middle of nowhere?”

The same basic principle applies in training neural networks in AI. As obvious as this may sound, the AI industry continues to pay a lot of attention to just throughput. Every new announcement on the market starts off with a speed claim on ResNet-50, a well-established image classification network. But it’s one thing to briefly run that network at high throughout (images per second in the ResNet-50 case) for a momentary period of time, but is the network converging and making accurate predictions so it can be deployed as an actual product or service? That is, does the network reach its destination?

Here’s what convergence means in a neural network. In implementing AI, you have two essential steps: training and inference. Training a network is the first step, where you teach a network how to make predictions based on your dataset. Inference is the process of deploying a trained network where you use it in production to make predictions. To train a neural network, you feed in vast amounts of data and it starts to look for patterns, like how the human brain operates.

For example, you have an image classification network being trained to classify houses. There can be hundreds or thousands of layers with billions of parameters, each comprising a series of equations that build up to the statistical probability or likelihood of a specific type of prediction. You feed in a dataset of house images and the network scans the images for patterns. This starts to update weights in probability equations that drive predictions.  ... " 

Security Flaws in IOT Devices Wont Get Fixed.

Was pointed out to me early as the classic security problem with edge devices lost in the field, they will never be updated to be secure, and they are attached to many other networks and devices, making them insecure.  

Critical Flaws in Millions of IoT Devices May Never Get Fixed

in Wired By Lily Hay Newman

Internet of Things (IoT) security firm Forescout uncovered 33 flaws, collectively labeled Amnesia:33, in seven open source TCP/IP stacks that potentially leave millions of IoT devices vulnerable. Many of the bugs were basic programming errors, like missing input validation checks that keep a system from accepting problematic values or operations. Patching these flaws is difficult if not impossible, as five stacks have been around for nearly two decades, while two have circulated since 2013; this means numerous versions and variants exist, with no central authority to issue fixes. Moreover, manufacturers who have incorporated the code into their products would have to proactively adopt the correct patch for their version and deployment, then circulate it to users. Said Forescout’s Elisa Costante, "What scares me the most is that it’s very difficult to understand how big the impact is and how many more vulnerable devices are out there."

Amazon Wants Smart Neighborhoods

Amazon has received much hammering of late on their smart home capabilities. Giving openings to hackers, listening in on private conversations  and through its neighborhoods, overpolicing the streets.  Now they are giving indications that still want to move forward to make people safer.   I will be involved,  Why not?

Amazon Sidewalk will create entire smart neighborhoods. Here's what you should know

Launching soon via Echo speakers and all sorts of other devices, Amazon's low-bandwidth IoT network lets your smart home stretch beyond Wi-Fi range.  .. '

Ry Crist in CNet

Obfuscating the Indistinguishable in Cryptography

 Below is something I had heard mentioned before, but was unclear of the implications.    Here are a number of statements and article give a good overview.   I am exploring for now.  Not saying I can completely state the implications.  Practitioners in the space should start thinking about it, 

A mouthful of terminology is potentially important to cryptography.    And apparently has now been determined to be possible.  This is written up in a recent Quanta article: 

Computer Scientists Achieve ‘Crown Jewel’ of Cryptography

A cryptographic master tool called indistinguishability obfuscation  (iO)   has for years seemed too good to be true. Three researchers have figured out that it can work.  ....

Schneier writes about it here:

   ....  Basically, obfuscation makes a computer program “unintelligible” by performing its functionality. Indistinguishability obfuscation is more relaxed. It just means that two different programs that perform the same functionality can’t be distinguished from each other. A good definition is in this paper.  

(Schneier)     https://eprint.iacr.org/2013/451.pdf    Definition 

https://eprint.iacr.org/2020/1003   Foundational paper of the most recent results

Signal Encryption Protocol

More attempts to decrease surveillance?

Hacker Lexicon: What Is the Signal Encryption Protocol?

As the Signal protocol becomes the industry standard, it's worth understanding what sets it apart from other forms of end-to-end encrypted messaging.

LAST WEEK, WITH little fanfare, Google announced a change that could soon make its 2 billion Android users worldwide far harder to surveil: The tech giant says it's rolling out a beta version of its Android messaging app that will now use end-to-end encryption by default. That level of encryption, while limited to one-on-one conversations, is designed to prevent anyone else from eavesdropping—not phone carriers, not intelligence agencies, not a hacker who has taken over the local Wi-Fi router, not even Google itself will have the keys to decrypt and read those billions of messages. .... '

Deep Learning Quality in Manufacturing?

 Interesting thoughts here.    Much more at the link.  

 Deep Learning Has Reinvented Quality Control in Manufacturing—but It Hasn’t Gone Far Enough

AI systems that make use of “lifelong learning” techniques are more flexible and faster to train  By Anatoli Gorchet

This is a guest post. The views expressed here are solely those of the author and do not represent positions of IEEE Spectrum or the IEEE.

In 2020, we’ve seen the accelerated adoption of deep learning as a part of the so-called Industry 4.0 revolution, in which digitization is remaking the manufacturing industry. This latest wave of initiatives is marked by the introduction of smart and autonomous systems, fueled by data and deep learning—a powerful breed of artificial intelligence (AI) that can improve quality inspection on the factory floor.

The benefit? By adding smart cameras to software on the production line, manufacturers are seeing improved quality inspection at high speeds and low costs that human inspectors can’t match. And given the mandated restrictions on human labor as a result of COVID-19, such as social distancing on the factory floor, these benefits are even more critical to keeping production lines running.

While manufacturers have used machine vision for decades, deep learning-enabled quality control software represents a new frontier. So, how do these approaches differ from traditional machine vision systems? And what happens when you press the “RUN” button for one of these AI-powered quality control systems?

Before and After the Introduction of Deep Learning in Manufacturing

To understand what happens in a deep learning software package that’s running quality control, let’s take a look at the previous standard. The traditional machine vision approach to quality control relies on a simple but powerful two-step process:  ... " 

Bigger Robots: Air Giants

Have taken a look at small robots for a long time,  especially as they are used in groups,  but this look at larger forms is also of interest.

Video Friday: These Giant Robots Are Made of Air, Fabric

Your weekly selection of awesome robot videos

By Evan Ackerman, Erico Guizzo and Fan Shi

Luma, is a towering 8 metre snail which transforms spaces with its otherworldly presence. Another piece, Triffid, stands at 6 metres and its flexible end sweeps high over audiences’ heads like an enchanted plant. The movement of the creatures is inspired by the flexible, wiggling and contorting motions of the animal kingdom and is designed to provoke instinctive reactions and emotions from the people that meet them. Air Giants is a new creative robotic studio founded in 2020. They are based in Bristol, UK, and comprise a small team of artists, roboticists and software engineers. The studio is passionate about creating emotionally effective motion at a scale which is thought-provoking and transporting, as well as expanding the notion of what large robots can be used for.  ... " 

Merging with our Tools

Will We, Can We ... Merge with our AI?

Well we did kind of merge with our calculators.   And we sure seem to be merging with our Smartphones.    And now ... ?

This is how we’ll merge with AI  in VentureBeat

Gary Grossman, Edelman

The relationship between humans and AI is something of a dance. We and AI come close together operating collaboratively, then are pushed away by the impossibility, only to stumble but return attracted by the potential. It is perhaps fitting that the dance community is beginning to embrace robots, with AI helping to create new movements and choreography, and with robots sharing the stage with human dancers.

The relationship between society and technology is yin and yang, with every massive enhancement accompanied by the potential for danger. AI, for example, offers the promise to end boring, repetitive jobs, enabling us to engage in higher level and more fulfilling tasks. It helps with any number of efficiency efforts, such as fraud detection, and it can even paint masterpiece artworks and compose symphonies. Sam Altman, CEO of OpenAI, hopes AI will unlock human potential and let us focus on the most interesting, most creative, most generative things.   .... '

Nathan Benaich's AI Newsletter

Just started refollowing:

Your guide to AI   December 6  2020, · Issue #48 · View online  from Nathan Benaich

Monthly analysis of AI technology, geopolitics, research, and startups.

Dear readers,

Welcome to the November 2020 issue of my newsletter, Your guide to AI. Here you’ll find an analytical narrative covering key developments in AI tech, geopolitics, health/bio, startups, research, and blogs.

If you enjoyed the read, I’d appreciate you hitting forward to a couple of friends.

I’d love to hear about what you’re working on and/or your feedback on the issue, just hit reply! 

Wishing you the best for the holidays,

Nathan Benaich

These past weeks have been a positive whirlwind for the biotechnology industry and global public health. First, we had the rapid approvals of COVID-19 vaccines, both using mRNA technology (Moderna, BioNTech) and more traditional methods (Oxford). While this news wasn’t (from what I can tell) driven by AI-based designs or workflows, another news item was: Baricitinib. This drug, which was postulated in February this year by London-based Benevolent.ai as a potential treatment for patients suffering from COVID-19 was granted emergency authorization by the FDA. Patients can now receive immediate use of Baricitinib in combination with Remdesivir because patients who received this dual treatment suffered a 35% lower mortality rate than those taking Remdesivir alone. Moreover, pre-clinical experiments conducted at the University of Washington showed that computationally designed mini proteins that mimic how an antibody would bind the COVID-19 can prevent its interaction with the ACE-2 receptor. This early work shows a) that understanding 3D protein structure is key to elucidating function, and b) that computationally-driven search and optimization of protein structure can rapidly generate useful drug candidates.    ... " 

Boston Dynamics Helping Chernobyl Decommission

 Boston Dynamics is doing quite well with its mostly telepresence Spot Robotics.  Here another example, with apparent connection to specialized sensors. 

Boston Dynamics' Spot Is Helping Chernobyl Move Towards Safe Decommissioning

Legged robots are uniquely qualified for sensing around what's left of Chernobyl's Reactor   By Evan Ackerman in IEEE Spectrum

In terms of places where you absolutely want a robot to go instead of you, what remains of the utterly destroyed Chernobyl Reactor 4 should be very near the top of your list. The reactor, which suffered a catastrophic meltdown in 1986, has been covered up in almost every way possible in an effort to keep its nuclear core contained. But eventually, that nuclear material is going to have to be dealt with somehow, and in order to do that, it’s important to understand which bits of it are just really bad, and which bits are the actual worst. And this is where Spot is stepping in to help.    ... 

More:  https://chnpp.gov.ua/en/infocenter/news/5668-boston-dynamics-robot-at-chnpp

Folding Proteins

 This is a very famous problem with broad solution implications in biochemistry, has it been solved now via Deep tech?     Not clear to me by this.  In NewScientist.   Protein Folding.  By Michael Le Page   See: Alphafold

Word from Cognixion on Brain Computer Interfaces

Update from colleague at Cognixion, developers of Brain Computer Interfaces (BCI) for  people with severe disabilities. 

Hi Franz, 

I wanted to share some great news. We recently closed a $1million seed angel round in less than 72 hours. The angel group was assembled and led by Tom Washing of Sequel Venture Partners. The use of proceeds is to accelerate our BCI (Brain Computer Interface) solution that enables people with severe disabilities to speak using only their brainwaves. We are also changing our company name from Smartstones to Cognixion™, to better reflect our vision and value proposition of becoming the cognitive layer to the internet - translating bio-signals like brainwaves into meaning and communication. 

“Andreas has assembled a great team and Cognixion’s technology is on the cutting edge of the quickly developing BCI space. Our group is very excited to join this highly promising venture,” said Washing.

Almost $1 billion has been invested into brain computer interface startups in the last 6 months to a year, and because many these big bets are to build high fidelity hardware and sensing technology, our platform is perfectly suited to integrate brainwaves from ANY BCI company. We will be integrating other BCI device partners with our brainwave pattern recognition software to unleash their potential. Further, we have begun building out our A.I./data science and development team in Toronto, Ontario, and are developing key academic, scientific, clinical and commercial partnerships soon to be announced.

As our vision has evolved, our mission has remained the same. Bring the brightest minds together to create solutions that can unlock a chorus of 370 million nonverbal humans - allowing them to express themselves, their ideas and engage with society in ways never before thought possible.

Cognixion is operating at the intersection of Neuro-technology, Artificial Intelligence and Augmented Humanity/Reality, and is helping increase human capacity and capabilities with augmentative technology.

We will be sharing more information as appropriate, but are operating in a relatively quiet/stealth mode as we fortify our platform, partnerships and distribution.

Feel free to contact me if you have any questions.

Andreas Forsland

www.speakprose.com 

2017 Roddenberry Prize Winner

Closer to Realistic Quantum Computing?

A move closer?  Technical.

Hitting the Quantum 'Sweet Spot': Researchers Find Best Position for Atom Qubits in Silicon

University of New South Wales Sydney Newsroom

December 1, 2020

Researchers from Australia's Center of Excellence for Quantum Computation and Communication Technology (CQC2T) at the University of New South Wales Sydney (UNSW), working with colleagues at Australia’s Silicon Quantum Computing firm, identified the ideal positioning for qubits in silicon in order to scale up atom-based quantum processors. Precise placement of phosphorus atoms in silicon is necessary to produce robust interactions between qubits, which CQC2T's Sven Rogge said are needed "to engineer a multi-qubit processor and, ultimately, a useful quantum computer." CQC2T's Benoit Voisin said scanning tunneling microscope (STM) lithography techniques developed at UNSW can find the "special angle, or sweet spot, within a particular plane of the silicon crystal where the interaction between the qubits is most resilient."   .... ' 

What Makes Robust AI?

 Much enjoyed Gary Marcus' writing on the essense of robust AI, most recently from his book, 'Rebooting AI'.  Now see more from him on 'Four Steps Towards Robust AI'.   See more about this in ZDNet.   See more also in Garymarcus.com

Police Drones Arriving

Expect that this kind of autonomous approach will continue to expand, and will be regulated in the coming years.   Here a good example.

Police Drones Starting to Think for Themselves

The New York Times,  By Cade Metz

Police agencies in four U.S. cities are participating in the Drone as First Responder program, launching unmanned aerial vehicles in response to emergency calls. The Chula Vista, CA, police dispatches drones, with a certified pilot federally on the roof of the Police Department to oversee launches and pilot the drones upon their return; a special drone from Silicon Valley’s Skydio avoids obstacles on its own and can follow a particular person or vehicle. The latest drone technology would allow police to operate autonomous drones relatively inexpensively, although civil liberties proponents are concerned. Greater police use of drones could eliminate any expectation of privacy outside the home, as the drones collect and store more video footage. The American Civil Liberties Union’s Jay Stanley said, "It could allow law enforcement to enforce any area of the law against anyone they want."

Moving AI Recognized Things

Interesting generalization of a set of common tasks using AI identification.

Robotics Researchers Propose AI That Locates, Safely Moves Items on Shelves

Venture Beat   By Kyle Wiggers in CACM

Two new robotics studies detail methods for locating occluded objects on shelves and solving "contact-rich" manipulation tasks. Researchers at the University of California, Berkeley developed the Lateral Access maXimal Reduction of occupancY support Area (LAX-RAY) system, which predicts an object's location even when only a portion of it is visible. LAX-RAY achieved 87.3% accuracy in a simulation, which translated to about 80% for a real-world robot. Meanwhile, Google developed the Contact-aware Online COntext Inference (COCOI), which uses video footage and readings from a robot-mounted touch sensor to encode dynamics information into a representation, which then permits a reinforcement learning algorithm to plan with “dynamics-awareness,” increasing its robustness in difficult environments.  ... ' 

Data Storytelling

'Data Storytelling', never heard it described that way, but when building process maps, we often described it in the form of a flow, or journey, or 'story'  to make it clear how changes occured.  So I like the idea of telling it as a story to make it easy to understand. 

Exploring 'Data Storytelling'  in CACM

By Joe Dysart    Commissioned by CACM Staff

Firms specializing artificial intelligence (AI) that generates written text are developing tools that will allow you to have sophisticated conversations with common business databases like Microsoft Excel, Microsoft Power BI, Microstrategy, Qlik, Spotfire, SAP, and Tableau. Instead of typing at your computer to gain insights, you'll be able voice-chat with your computer about sales forecasts, investment options, and five-year plans.

"It is totally possible," says Robert Weissgraeber, chief technology officer at AX Semantics, a firm specializing in natural language generation (NLG). Weissgraeber says getting to the point where you can easily converse with your database is mainly a problem of collecting and unifying the data into a source that can be connected to the right tools.

"The big thing that has changed in the past two years is that, thanks to commoditization of natural language generation tools, a conversational business database is not a special project with a million-dollar price tag anymore. It can be set up for a far lower price point, allowing even smaller businesses to implement and benefit from it."

Many consumers have used basic conversational databases without even realizing it. When you ask Alexa or asks a question like "What happened today in history?" or "How does the intercom work?" or "How many ounces are in a pound?" you're using a conversational database.

Natural language generation firms like AX Semantics and Arria NLG to take this technology and apply it to a business database, so businesses can verbally ask questions of their databases, like:

"How do our sales for Q4 this year compare with Q4 sales for last year?"

"What are the five top-selling products in our product line this year?"

"Who are the top three salespeople in our corporation in each of our eight regions?"

Software from NLG companies can respond to such questions by relying on pre-programmed code templates that are activated by the questions, which instruct it how to obtain the information in a company's database, then report its findings verbally.

The level of detail such software is able to produce is limited by the number of questions companies want answered. Some companies may be satisfied if the software can produce verbal answers to 10 crucial questions, while others may want their managers' top 100 questions answered.

The Art of the Barcode

Our group was one of the very first users of barcodes in retail, so have followed it and related information and compression techniques for a long time.   Here an update, have not read it but will.

[eBook] The Art of the Code   in RetailWire

An essential guide to the latest in barcode data capture technology

Since its first scan, the barcode has revolutionized the retail industry, bringing improved store productivity and inventory management as well as upstream benefits throughout the supply chain. At this point in its evolution, there are dozens of barcode symbologies, each with individual strengths, weaknesses and application-specific designs.

This Resource Guide is a valuable reference tool for barcode beginners and IT veterans alike. Learn about:

Different bar code symbology and their applications;

The right symbology for your business/application;

The right printing technique for your selected symbology. ... '

Digitizing StoryCorps Audio Interviews

 Also similar to work we did in the enterprise to understand consumer interviews.  We spent some time looking for solutions of this type, but they were not available.  Today it would seem that adaptations of the below would work.

AI to digitize StoryCorps’ massive archive of audio interviews   By  Mike Wheatley in SiliconAngle

Google LLC today shared how it has worked with the nonprofit organization StoryCorps Inc.      to digitize an enormous archive of human story audio recordings using some of its most advanced artificial intelligence technologies. 

StoryCorps’ mission is to record, preserve and share the stories of Americans from all backgrounds and beliefs. The project was founded back in 2003, and since then it has recorded one-on-one interviews with more than 600,000 people. Those recordings are stored at the American Folklife Center in the U.S. Library of Congress, where they are being preserved for future generations.

The StoryCorps archives are believed to be the largest collection of human voices on the planet, but up until recently those stories have been inaccessible to most people. To change that, StoryCorps approached Google to ask if it could help make its rich archive of history universally accessible to anyone who’s interested in listening to them.   ... 

.... StoryCorps’ entire archives can now be accessed from its online platform  and its mobile application.   ... 

Navigating Business Turbulence

Non-technical points on developing with rapidly changing data and contexts.   Ultimately good requirements for adapting business process. 

How to use data and analytics to navigate business turbulence   by 7wData

The ability to quickly gather pertinent facts and data points about your business is key to making informed decisions and overcoming any crisis, such as the COVID-19 pandemic we're going through now.

This is a period of great anxiety and uncertainty for businesses around the globe. But by tapping into reliable data and real-time analytics, organizations can take action to improve their situation. They can quickly course-correct if they see their business going sideways, reallocate resources to areas where they'll do the most good, and make effective decisions to survive and succeed in these challenging times.

Periods of crisis will always happen, whether it's a terrorist attack, a global conflict, a market crash or another pandemic. But humans adjust. It's what we do best. We usually bounce back better and stronger than before-and this time should not be any different.

Here are four ways that the effective use of data and analytics can help your company as you enter the recovery phase and plan a brighter future.

Years ago, it was possible to fly your business by hand, because information arrived at a manageable pace. You could take in data, evaluate it and adjust your course at a deliberate pace, with a key spreadsheet or three keeping track of the numbers.   .... ' 

But today you have so much data coming at you in real time that it's hard to gather it, organise it and put it to good use. A lot of companies simply get overwhelmed, throw up their hands, and keep flying straight into the bad weather ahead. That's a big mistake because you will have a rough flight and could very well crash.

Instead, get the tools you need to gather and analyse data. This way you can see turbulence before it arrives, navigate through the clouds avoiding the worst of the storm and resume a smooth flight path on the other side.

Airlines long ago recognised that instrumentation is vital to the safe and efficient operation of their aircraft. The cockpit instruments provide pilots with crucial data like altitude, airspeed and heading. They also can show the impact of headwinds on fuel consumption or calculate the fuel margin available after a significant detour-so much more than they can gather with the naked eye. By the same token, the best corporations now rely on data and analytics to manage their operations, rather than fly blind from one hair-raising incident to the next.  Breaking down silos between your front and back office functions lets you know not just how much fuel is left in the tank, but what's the weather like between you and your target, and should you be speeding up, slowing down, or trying a different path to dodge the worst of the storm.   .... " 

Light-Based Quantum Computer Exceeds Fastest Classical Supercomputers

More indications of increases in quantum computing speed,  when applied to particular kinds of problems.  But this does not mean this is available yet,  but in a quantum future when fault tolerant quantum computers of sufficient capacity are available 

Light-Based Quantum Computer Exceeds Fastest Classical Supercomputers

Scientific American  by Daniel Garisto

Researchers at the University of Science and Technology of China (USTC) have for the first time coaxed a quantum computer composed of photons to outperform the fastest classical supercomputers. The Jiuzhan system executed Gaussian boson sampling, detecting 76 photons versus classical supercomputers' previous record of five. Jiuzhan combines lasers, mirrors, prisms, and photon detectors, and its achievement is only the second demonstration of quantum primacy to date. UTSC's Chao-Yang Lu said independent corroboration that quantum computing principles can enable primacy on totally different hardware "gives us confidence that in the long term ... useful quantum simulators and a fault-tolerant quantum computer will become feasible."

Driverless Cars Are Still Coming

 More comments on the likely general introduction of driverless vehicles.

Driverless Cars Are Coming, But Not Yet to Take Over

The Wall Street Journal, By Stephen Wilmot

Driverless cars are making their way onto U.S. roads, but it will take time for them to reach their full potential. Alphabet's Waymo began offering "robotaxis" with no backup driver in Phoenix suburbs in October, and by year’s end, General Motors' Cruise will no longer require backup drivers in its autonomous test cars in California. However, after safely operating driverless-cars in one urban district in good weather, companies must adapt to new areas and other driving conditions. It will take years, and billions of dollars in capital, to enable self-driving vehicles to adapt to multiple geographies and weather conditions, prompting partnerships between tech giants and the auto industry. For instance, Amazon has taken a stake in Aurora Innovation, a startup looking to take over Uber's operation, and acquired Aurora's rival Zoox for $1.3 billion this summer. Waymo's Larry Burns predicts that freight, rather than taxis, will fuel the commercialization of driverless technology.  .... '

Last Year in AI, Analytics, Machine Learning and Data Science ....

Good end of the year piece from KDNuggets that was instructive.

AI, Analytics, Machine Learning, Data Science, Deep Learning Research Main Developments in 2020 and Key Trends for 2021

Tags: 2021 Predictions, AI, Ajit Jaokar, Analytics, Brandon Rohrer, Daniel Tunkelang, Data Science, Deep Learning, Machine Learning, Pedro Domingos, Predictions, Research, Rosaria Silipo

2020 is finally coming to a close. While likely not to register as anyone's favorite year, 2020 did have some noteworthy advancements in our field, and 2021 promises some important key trends to look forward to. As has become a year-end tradition, our collection of experts have once again contributed their thoughts. Read on to find out more.

By Matthew Mayo, KDnuggets.

To the chagrin of absolutely no one, 2020 is finally drawing to a close. It has been a rollercoaster of a year, one defined almost exclusively by the COVID-19 pandemic. But other things have happened, including in the fields of AI, data science, and machine learning as well. To that end, it's time for KDnuggets annual year end expert analysis and predictions. This year we posed the question:

What were the main developments in AI, Data Science, Machine Learning Research in 2020 and what key trends do you see for 2021?

Last year's noted main developments and predictions included continued advancements in many research areas, NLP in particular. While there can be debate as to whether 2020's big NLP advancement was as formidable as some may have originally thought (or continue to think), there is no doubt that there was a continued and intense focus on NLP research in 2020. It should not be difficult to surmise that this continues into 2021 as well.  ... "

Samsung Does New Bixby Interface in Major Update

 Been a long time since we had an unimpressed look at Bixby.   Some good details here.

Samsung Reveals New Bixby Interface and Features in Major Update By Eric Hal Schwartz   in Voicebot.ai

Samsung has officially announced plans to revamp the Bixby virtual assistant a little over a month after it quietly began releasing the update to some Samsung smartphones. The changes bring a new, simplified interface and several new features to Bixby, with an eye toward making Bixby more accessible and personalize the voice commands it suggests.

BACKGROUND AI

Bixby’s new look is all about discretion and consolidation. Talking to Bixby no longer requires all of the screen’s real estate. Bixby runs from a section at the bottom of the screen, with a horizontal bar and faded background letting the user know the voice assistant is listening. The app on the smartphone is condensed as well, with everything related to Bixby on a single home screen instead of using separate pages for the Bixby Capsules. Bixby is also keen to help people use its features more often, providing suggestions for commands based on popular trends. The AI can even track what a user has done before and provide ideas for things to do with Bixby that they might not realize are available.

“Using Bixby feels more personal in the latest update, as the service now offers customized voice command suggestions based on your usage patterns and other devices you have registered with Bixby, Samsung explained in its announcement. “With recommendations for a variety of apps and services, the new tailored experience helps you discover more ways to use Bixby and improve your Galaxy experience.”

BIXBY’S FUTURE

The updated Bixby certainly confirms Samsung’s faith in the voice assistant. That matters more than it might normally after months of rumors that Samsung was dropping Bixby in favor of Google Assistant. Samsung shutting down Bixby Vision, the augmented reality feature of its voice assistant that can measure objects and simulate objects in a room, and reassigning the Bixby developer marketing team to other areas at the end of September didn’t help. And that news followed stories of other Bixby departures, including Voicebot’s exclusive report from July that Adam Cheyer, Viv Labs co-founder, and Samsung vice president leading Bixby development, had left the company. ... " 

Machine Learning Underspecification

 Yes, often a problem, often  mainly for lack of sufficient data in context.   Seems Google know this too.  Key issue.  Credibility/value is the key.  Note COVID example.  Technical. 

Googlers Speak Out on the Scourge of ML Underspecification

Oliver Peckham    in Datanami

A few days ago, 40 authors (all but a handful hailing from Google) published a 59-page paper. https://arxiv.org/pdf/2011.03395.pdf   The topic at hand: why so many machine learning models, borne out by internal testing, proceed to then fail spectacularly in real-world applications. The answer, the Googlers say, is underspecification – a blight on machine learning that, they stress, requires substantive solutions.

“An ML pipeline is underspecified when it can return many predictors with equivalently strong held-out performance in the training domain,” they write. In plain language: an underspecified model can think of a bunch of reasonably accurate explanations for why a dataset looks the way it does. The problem comes in when researchers assume that all of those explanations are equivalently valid based solely on the model’s training results, without accounting for real-world factors that may have escaped the model’s training process. In those situations, the authors say, “ML models often exhibit unexpectedly poor behavior when they are deployed in real-world domains.”

By way of illustration, the Googlers highlight examples spanning “computer vision, medical imaging, natural language processing, clinical risk prediction based on electronic health records, and medical genomics.” In epidemiology, for instance, they discuss how early data from an epidemic (such as the COVID-19 pandemic) is easily explained by a variety of models that do not substantively account for major factors – such as the gradually diminishing number of susceptible people in an area as an epidemic infects (and then renders immune) larger and larger portions of the populace.

“Importantly, during the early stages of an epidemic … the parameters of the model are underspecified by this training task,” they write. “This is because, at this stage, the number of susceptible is approximately constant at the total population size (N), and the number of infections grows approximately exponentially.”

As a result, they say, “arbitrary choices in the learning process” determine which parameters are deemed most predictive by the model, despite different models predicting “peak infection numbers, for example, that are orders of magnitude apart.” .... '

The paper, titled “Underspecification Presents Challenges for Credibility in Modern Machine Learning,”

Podcasts via RSS Feed

New way to connect to podcasts

Google Podcasts lets you add shows using their RSS feeds

Subscribe to your favorite niche podcast

by Jon Fingas  In Engadget    Google Podcasts just became more alluring if you tend to listen to shows that aren’t always available on major podcast networks. 9to5Google reports that Google has added support for subscribing to podcasts using their RSS feeds. You just have to paste the raw address into the Android, iOS and web apps to have access to episodes across your devices.

You can add a podcast in the mobile app by tapping “Activity,” then the “Subscriptions” tab, and choosing “Add by RSS feed” from the overflow menu. We’ve given you the direct web link above, but you can always find it again by looking for “Add by RSS feed” in the navigation drawer.   ... '

The Risk and Uncertainty of it All

A space we played in early on, determining measures of uncertainty as we built AI and expert system based models.   Its part of any decision driving system,  That includes measures of uncertainty and risk.  Glad to see efforts to include it in machine learning systems today.  Such measures should lead to model improvements.  Now can these models also learn what aspects of the model design and data directly create problems of certainty?   

A Neural Network Learns When It Should Not Be Trusted    By MIT News

 

The advance could enhance safety and efficiency in artificial intelligence-assisted decision-making. 

Massachusetts Institute of Technology researchers have developed a way for deep learning neural networks to rapidly estimate confidence levels in their output.

Researchers at the Massachusetts Institute of Technology (MIT) and Harvard University have enabled a neural network to rapidly process data, yielding both predictions and confidence levels based on the quality of the available data.

This deep evidential regression technique, which estimates uncertainty from a single run of the neural network, could lead to safer results.

The team designed the network with bulked-up output, generating not only a decision but also a new probabilistic distribution capturing the evidence supporting that decision; these evidential distributions directly capture the model's confidence in its forecast.

Included is any uncertainty within the underlying input data and the model's final decision, which indicates whether uncertainty can be reduced by modifying the network itself, or whether the input data is merely noisy.

MIT's Daniela Rus said, "By estimating the uncertainty of a learned model, we also learn how much error to expect from the model, and what missing data could improve the model."

From MIT News

Small Sat Launching by Drone

 Been watching the numerous satellite launches going on, like primarily to improve communications to remote areas. Agree with some of the worries for the future,  will we be able to see the stars in future years?   Will kids of the future see the sky?     Here is another that is now happening, drone plane launching.

Huge drone set to launch satellites in mid-air  by Peter Grad , Tech Xplore

We all have images in our mind of rocket launches from Cape Canaveral Air Force Station hurtling astronauts into space and satellites into orbit.

But those launches may be a thing of the past as a new generation of drones that can do the same job cheaper, safer and better steps into play.

Alabama-Based Aevum unveiled its Ravn X Autonomous Launch Vehicle Wednesday that it says is the world's largest unmanned aircraft system.

The Ravn X fleet, with aircraft at 80 feet long and 60 feet wide, can deliver satellite payloads with precision accuracy every 180 minutes around the clock. Best of all, company officials say, it can be done safer than current space delivery systems because the crafts are unmanned.  ... "

Shrinking BERT Networks to Model Language

Considerable shrinking of neural networks, more likely for applications at the Edge.

Shrinking massive neural networks used to model language

A new approach could lower computing costs and increase accessibility to state-of-the-art natural language processing.

Daniel Ackerman | MIT News Office

Researchers at the Massachusetts Institute of Technology (MIT), the University of Texas at Austin, and the MIT-IBM Watson Artificial Intelligence Laboratory identified lean subnetworks within a state-of-the-art neural network approach to natural language processing (NLP). These subnetworks, found in the Bidirectional Encoder Representations from Transformers (BERT) network, could potentially enable more users to develop NLP tools using less bulky and more efficient systems, like smartphones. BERT is trained by repeatedly attempting to fill in words omitted from a passage of writing, using a massive dataset; users can then refine its neural network to a specific task. By iteratively trimming parameters from the BERT model, then comparing the new subnetwork's performance to that of the original model, the team found effective subnetworks that were 40% to 90% leaner, and required no task-specific fine-tuning to identify "winning ticket" subnetworks that executed tasks successfully.  .... 

Robotic Hands Closer to Human

New robotic hands with lots of possibilities for more dexterous manipulation 

Robot Hands One Step Closer to Human, Thanks to WMG AI Algorithms  By University of Warwick (U.K.)

Artificial intelligence algorithms enable the Shadow Robot Dexterous Hand to manipulate objects the same ways humans do.

Artificial intelligence (AI) algorithms developed by researchers at the Warwick Manufacturing Group (WMG) academic department of the U.K.'s University of Warwick enable the Shadow Robot Dexterous Hand to manipulate objects like humans do.

The Hand reproduces all of the degrees of freedom of a human hand.

The algorithms permit the Hand to learn how to coordinate movements and execute tasks such as throwing a ball and spinning a pen; the researchers said the algorithms can learn any task, as long as it can be simulated.

Said Warwick's Giovanni Montana, "The future of digitalization relies on AI algorithms that can learn autonomously, and to be able to develop algorithms that give Shadow Robot’s hand the ability to operate like a real one is without any human input is an exciting step forward.   ... " 

IRS to Make ID Protection PIN Open to All

Good direction.   Now how will it be effectively used?    See more detail at the link.

IRS to Make ID Protection PIN Open to All in Krebs On Security

The U.S. Internal Revenue Service (IRS) said this week that beginning in 2021 it will allow all taxpayers to apply for an identity protection personal identification number (IP PIN), a single-use code designed to block identity thieves from falsely claiming a tax refund in your name. Currently, IP PINs are issued only to those who fill out an ID theft affidavit, or to taxpayers who’ve experienced tax refund fraud in previous years. .... " 

Considering Orbit in Developer Relations

Read about this in Andreessen Horowitz.  About better online communities.  Involved in an effort that will need some of this. 

Investing in Orbit by Martin Casado

Great communities make great companies, and community-building is now part of a broader shift in B2B businesses towards bottom-up growth. As part of that shift, companies are relying on communities as a key method for growing and engaging their user base, and moving from focusing on traditional marketing and direct sales to community growth and post-sales. Post-sale because a community member often already uses the product, so maturing them with the business is more about cultivating the user experience and turning them into community advocates than trying to convert them from a lead to a paying customer.    .... 

Why Orbit is Better Than Funnel for Developer Relations

DevRel teams need tools and models created specifically for our discipline, and not just those adopted from other fields.

Josh Dzielak and Patrick Woods.   .... 

Google Discusses How it Organizes Information

Below the introduction.   Informative.

How Google organizes information to find what you’re looking for

By Nick Fox in the Google Blog

Vice President of Product & Design, Search and Assistant

When you come to Google and do a search, there might be billions of pages that are potential matches for your query, and millions of new pages being produced every minute. In the early days, we updated our search index once per month. Now, like other search engines, Google is constantly indexing new info to make accessible through Search.

But to make all of this information useful, it’s critical that we organize it in a way that helps people quickly find what they’re looking for. With this in mind, here’s a closer look at how we approach organizing information on Google Search.

Organizing information in rich and helpful features

Google indexes all types of information--from text and images in web pages, to real-world information, like whether a local store has a sweater you’re looking for in stock. To make this information useful to you, we organize it on the search results page in a way that makes it easy to scan and digest. When looking for jobs, you often want to see a list of specific roles. Whereas if you’re looking for a restaurant, seeing a map can help you easily find a spot nearby. 

We offer a wide range of features--from video and news carousels, to results with rich imagery, to helpful labels like star reviews--to help you navigate the available information more seamlessly. These features include links to web pages, so you can easily click to a website to find more information. In fact, we’ve grown the average number of outbound links to websites on a search results page from only 10 (“10 blue links”) to now an average of 26 links on a mobile results page. As we’ve added more rich features to Google Search, people are more likely to find what they’re looking for, and websites have more opportunity to appear on the first page of search results.   ..." 

Texting to Alexa

 Noted for using voice in specific context.   Say for the need of privacy in conversation. 

Amazon now lets you text Alexa on iOS to ask for things instead of only using your voice  By Chaim Gartenberg    in MSN.com  via Walter Riker

Amazon is testing a new feature for its iOS Alexa app: the ability to type out Alexa commands, instead of having to ask questions or requests by speaking out loud.

“Type with Alexa is a Public Preview feature available to iOS Alexa app customers allowing you to interact with Alexa without using voice, meaning everything you can currently say to Alexa can now also be typed using your Alexa mobile app. Type with Alexa is available to iOS customers in the U.S.,” said an Amazon spokesperson. ... " 

Are Smart Things Really Smart?

Interesting challenge.  Depends on what 'smart' means.  Like what I have often said, its in a  'context'.   Smart can be exactly what it needs to be to provide value ... or not. Marketing tries to enhance the link.  Sometimes it works, sometimes not.  Thought  provoking piece. 

 

Are Smart Things Really Smart or is it Just Smart Marketing?

Denyse Drummond-Dunn  in CustomerThink

Earlier this year I wrote about the impact of AI and ML on digital marketing. The article is called AI and ML are Taking Digital Marketing to the Next Level.” In it, I compared the positive and negative implications of technology for customers and companies alike.  So this week I wanted to write about the impact of smart choices for business in general.

We seem to be surrounded by smart things: smartwatches, smart clothing, smart cars, smart houses and smart appliances. But are they really that smart? 

The reason for my question is that an article entitled “Taking ‘Smart’ Out Of Smart Things” by Chuck Martin made me think about whether “smart things” really are that smart, or whether it’s something else that’s making them appear smart?

So here are my views on it. Feel free to add your own opinions in the comments below, I would love to start a discussion on “smartness”.

The Age of the Customer and the Fourth Industrial Revolution

In one of their older Customer Experience reports Forrester claimed that we are now in the “Age of the Customer“. This was music to my ears when I first read it, because as you know I’m a customer champion. However, The World Economic Forum reported a few years ago that we are now on the brink of the “ Fourth Industrial Revolution” which is blurring the lines between physical, digital and biological spheres.

In their article, they explain that

“The First Industrial Revolution used water and steam power to mechanize production. The Second used electric power to create mass production. The Third used electronics and information technology to automate production. Now a Fourth Industrial Revolution is building on the Third, the digital revolution that has been occurring since the middle of the last century. It is characterized by a fusion of technologies that is blurring the lines between the physical, digital, and biological spheres.”   ... " 

Google General Voice Accessibility

In  the Google Blog, a means of more general voice accessibility via voice on the android phone.  Now updated. 

ACCESSIBILITY

Use Voice Access to control your Android device with your voice

Easily use and navigate your phone by speaking out loud with Voice Access. Image shows two cartoon figures, one who uses a wheelchair and has their phone mounted in front of them, and another standing behind the wheelchair as they take a selfie.

Tom Hume, Senior Product Manager, Google Research

In 2018, we launched Voice Access, an Android app that lets you control your phone using your voice. The ability to use your phone hands-free has been helpful to people with disabilities, and also those without.

Today, on International Day of Persons with Disabilities, we’re rolling out an updated version of Voice Access, available in Beta, that is easier to use and available to more people. This version of Voice Access, which was previously available on Android 11, is now available globally to devices running Android 6.0 and above.   ... " 

Using VisiRule to Create Chatbots

Had been looking for an easy way to create chatbots that include FAQ's. Recalled a company called  VisiRule that supported rule-based expert systems.  Decision tree style chatbots.    We looked at them some time ago.    Below are some examples of their work, with a number of chatbot examples.  VisiRule has been mentioned many times in this blog, see their tag below.

ChatBot Demos

This page contains various ChatBot demos which have been automatically generated from a VisiRule chart. This allows authors to quickly draw the intended conversation structure visually and easily. In addition, common questions can be answered by attaching a FAQ KB in the form of question/response. VisiRule contains a NLU component to help match user input text to these known questions. The combination of a structured decision tree flowchart conversation with general purpose information retrieval makes for a more rounded and holistic user experience.  ... "

Alphabet's Loon Balloons for Internet Connectivity

Application of the general idea is expanding.  Here with autonomous navigation. 

Alphabet's Loon Balloons in Internet Connectivity

Alphabet’s Loon hands the reins of its internet air balloons to self-learning AI.  The company’s new AI flight control system outperforms its human-made one

By Nick Statt@nickstatt

Alphabet’s Loon, the team responsible for beaming internet down to Earth from stratospheric helium balloons, has achieved a new milestone: its navigation system is no longer run by human-designed software. ... " 

Detecting Scientific Conflicts of Interest

 With other kinds of application in compliance.   But see the note about subjective aspects. 

Do You Have a Conflict of Interest? This Robotic Assistant May Find It First

The New York Times, Dalmeet Singh Chawla

Frontiers, a Swiss publisher of open-access journals, has rolled out the Artificial Intelligence Review Assistant (AIRA) to check for potential conflicts of interest. The software flags whether the authors of a manuscript, as well as the editors and peer reviewers handling it, have been co-authors on previous papers. Said Frontiers' Kamila Markram, "AIRA is designed to direct the attention of human experts to potential issues in manuscripts." Other publications are using similar artificial intelligence (AI) tools, but some researchers note that conflicts of interest can be subjective and difficult to unravel. Canada's McGill University computer scientist Kaleem Siddiqi noted, "There's no real solution."  .... 

Detecting Likelihood of Suicide

Have consulted and spoke at the Veterans Administration.  Here is another pattern recognition type problem of interest.  

Can an Algorithm Prevent Suicide?

The New York Times, Benedict Carey

The U.S. Department of Veterans Affairs (VA) is using machine learning algorithms to help identify veterans at risk of suicide. The VA's Reach Vet algorithm is the first such program used in daily clinical practice, and is designed to generate a new list of high-risk veterans every month. When someone is flagged at risk, their name appears on the computer dashboard of the local clinic's Reach Vet coordinator, who contacts them to set up a meeting. The algorithm is based on analysis of thousands of veteran suicides; it considers numerous factors in veteran medical records to focus on those with the strongest cumulative association with suicide risk. Initial results indicate that over six months, high-risk veterans more than doubled their use of VA services—and had a lower mortality rate—with Reach Net installed, compared to a control group. ... 

SAS: Four Principles of Analytics

As usual, SAS does a good job of precisely outlining analytics. Here including AI and Big Data with statistics. 

It's no secret that technology's changing. A change accelerating so fast, it's hard to keep up.

Big data and AI are exploding. Industries are reinventing themselves. And breakthroughs seem to redefine our world every day. Yet thankfully, some things do stay the same.

The four principles of analytics are steadfast truths that inform your approach to data and analytics. They're truths because, well, they work – helping you make the best decisions no matter what else changes. Read this blog to learn why:

Analytics follows the data.

Analytics is more than algorithms.

Data and analytics should be available to everyone.

Analytics is a differentiator.

 .... After all, change isn't the only thing that's constant. So are good ideas.

Read the Blog

AI Battling AI

 Had always thought it would come to this.  Mentioned it in a talk I gave in the 90s.   But how well and how soon?  As long as there are malicious developers out there, sooner than we thought.

This Company Uses AI to Outwit Malicious AI  in Wired

Robust Intelligence is among a crop of companies that offer to protect clients from efforts at deception.

IN SEPTEMBER 2019, the National Institute of Standards and Technology issued its first-ever warning for an attack on a commercial artificial intelligence algorithm.

Security researchers had devised a way to attack a Proofpoint product that uses machine learning to identify spam emails. The system produced email headers that included a “score” of how likely a message was to be spam. But analyzing these scores, along with the contents of messages, made it possible to build a clone of the machine-learning model and craft spam messages that evaded detection. ... " 

Germany's Energiewende

 As a native, a keen watcher of German energy efforts.   I remember all the windmills sprouting in the 90s.   Here a good overview in IEEE Spectrum. 

Germany's Energiewende, 20 Years Later

Germany's far-reaching program to reduce the share of fossil fuels in energy has achieved almost exactly what the United States achieved, but at greater expense

By Vaclav Smil

In 2000, Germany launched a deliberately targeted program to decarbonize its primary energy supply, a plan more ambitious than anything seen anywhere else. The policy, called the Energiewende, is rooted in Germany’s naturalistic and romantic tradition, reflected in the rise of the Green Party and, more recently, in public opposition to nuclear electricity generation. These attitudes are not shared by the country’s two large neighbors: France built the world’s leading nuclear industrial complex with hardly any opposition, and Poland is content burning its coal.

The policy worked through the government subsidization of renewable electricity generated with photovoltaic cells and wind turbines and by burning fuels produced by the fermentation of crops and agricultural waste. It was accelerated in 2011 when Japan’s nuclear disaster in Fukushima led the German government to order that all its nuclear power plants be shut down by 2022.  .... " 

Blackberry and IVY

 A long time examiner of how automobiles, today or in a more autinomos future, will link to more sensors and the web.   Also an early user of Blackberry devices in the enterprise, before they were driven under by the iPhone.    Is Blackberry back?   What are the key apps to watch for? 

BlackBerry shares rocket upwards on AWS deal to integrate sensor data in vehicles

By Jonathan Shieber in TechCrunch

BlackBerry shares shot up in early trading on news that the company will partner with Amazon Web Services to jointly develop and market its vehicle data integration and monitoring platform, IVY.

BlackBerry stock was up 35%, or $2.11, at the opening bell on the New York Stock Exchange. It’s a sign of both the potential market for smart vehicle services and the ability of Amazon businesses to boost the fortunes of businesses with its attention.

The former undisputed heavyweight of the smartphone market, BlackBerry has transformed itself into a provider of business security and information integration services, and it’s through this transformation that the company attracted the attention of Amazon’s web services business.  ... " 

IKEA Smart Home

 Have not seen too much of a move by IKEA as yet, especially in linking to other capabilities.  But here is one example.

IKEA's smart home system now supports scenes

Long-awaited shortcut buttons appear to be coming soon, too.

Jon Fingas in Engadget

IKEA is filling an important feature gap in its smart home system. The Verge reports that IKEA is adding scene support to Home Smart through a firmware update (1.12.31) for its Tradfri gateways, making it easy to control multiple devices at once without having to rely on other companies’ platforms. You can quickly dim the lights and silence your Symfonisk speakers when you’re getting ready for bed, for instance.... " 

Microsoft Teams to be Enhanced for Business Plays

Been looking for ways that the now numerous collaborative systems will really improve task and goal oriented business process.   Its still mostly the basics.  Very useful but hardly 'intelligence' in any sense.  For example a system that could analyze the essence of a  meeting and extract decisions like who agreed to do what and what support they need.  Schedule follow ups.  The below is not that, but I get the sense they are considering new things of real value.   We brainstormed some things like that in the early 00s.  Lets see it now.

Microsoft Teams gets an overhauled calling interface, CarPlay support, and more

Microsoft Teams will soon let you transfer calls easily between desktop and mobile

By Tom Warren@tomwarren in TheVerge 

Microsoft is overhauling its calling features inside Microsoft Teams today. A new calling interface will now show contacts, voicemail, and calling history in a single location. It’s designed to allow Microsoft Teams to more easily replace your desk phone, with built-in spam call protection, reverse number lookup, and the ability to merge calls.

Microsoft Teams users will also be able to transfer calls between mobile and desktop soon, allowing people to quickly move locations in the middle of an audio or video call. The Teams app will let people join without audio on an additional device, or simply transfer the call and end it automatically on other devices. Microsoft says this particular feature will be available in early 2021.  ... '

Amazon Trainium Chip For Cloud ML

 Cute name, new to me.  Another example of the hardware linking with the software methods for driving AI.  What improvements can we expect, and where?

Amazon debuts Trainium, a custom chip for machine learning training in the cloud

By Kyle Wiggers  @Kyle_L_Wiggers n VentureBeat

Amazon today debuted AWS Trainium, a chip custom-designed to deliver what the company describes as cost-effective machine learning model training in the cloud. It comes ahead of the availability of new Habana Gaudi-based Amazon Elastic Compute Cloud (EC2) instances built specifically for machine learning training, powered by Intel’s new Habana Gaudi processors.

“We know that we want to keep pushing the price performance on machine learning training, so we’re going to have to invest in our own chips,” AWS CEO Andy Jassy said during a keynote address at Amazon’s re:Invent conference this morning. “You have an unmatched array of instances in AWS, coupled with innovation in chips.”   .... " 

Holiday Shopping in 2020

Based on a number of studies.  Will be interesting to see how this plays out after some normalcy returns.

Holiday shopping in 2020

Consumers, longing for normalcy and eager for something to celebrate, plan to spend money while holiday shopping--but differently from the way they have in the past.   Sent from McKinsey Insights, available in the App Store and Play Store.  ...  "