Industry News

Industry News

Researchers use AI and accelerometer data to predict heart rate while saving battery life, cutting it to 10%!, Alexa alternatives have a secret weapon: Privacy, How Can we Ensure that Big Data Does not Make us Prisoners of Technology?, Facial recognition technology: The need for public regulation and corporate responsibility, Algorithms are taking over – and woe betide anyone they class as a ‘deadbeat’ – seems to be a week of Woes in the land of AI and ML. Amazon AI predicts users’ musical tastes based on playback duration – this I like! Japan’s Paidy raises $55m for credit card-free online shopping service – cool but can we do even better?
Goldman Sachs Targets 25% Upside for Twitter Stock – can you write a model? But don’t use Goldman’s predictions, use one of our Sentiment data sets. How Goldman Sachs Lost the World Cup.

AI & Machine Learning News. 16, July 2018

Researchers use AI and accelerometer data to predict heart rate while saving battery life

What do the Apple Watch and Nokia Pulse Ox have in common? They’ve both got pulse oximeter sensors that measure heart rate using photoplethysmography (PPG), the expansion and contraction of capillaries based on changes in blood volume. They’re accurate to a degree, but require a fair amount of electricity because they’re light-based — they emit a signal onto the skin that reflects back to a photodiode.
One battery-saving alternative might be accelerometers, a sensor commonly found in smartphones, smartwatches, and activity trackers that measures non-gravitational acceleration. In a paper published on the preprint server, researchers at Philips Health and the University of Bristol describe a machine learning algorithm that can predict heart rate almost exclusively from the sensors, boosting the battery life of the wearable to which they’re attached.
2018-07-13 00:00:00 Read the full story.
CloudQuant Thoughts… More great applications of AI. A simple low power Accelerometer plus a little AI can detect if you seem in distress then, and only then, switching on the battery hogging PPG on your smartwatch.

Alexa alternatives have a secret weapon: Privacy

Earlier this week we learned that worldwide smart speaker sales are expected to increase sixfold within the next couple of years. This mirrors multiple studies that say the majority of U.S. households will have a smart speaker by 2022, powered by current leading intelligent assistants Google Assistant and Alexa.
At the same time, tech giants making intelligent assistants seem to want to have it both ways, selling products to both consumers and governments. For example, Microsoft, maker of Cortana, may be supplying facial recognition software to ICE, the government agency tasked with capturing and detaining immigrants who are in the United States illegally.
That’s why startups like Snips, which is bringing its own smart speaker to market, center their attention on one primary differentiator: privacy.
2018-07-14 00:00:00 Read the full story.
CloudQuant Thoughts… Alexa listening to your every word and emailing clips to random people on your email list. As a child I wanted to build my own Star Trek “computer” to talk to, heck, if they allowed me to program the trigger word myself I probably would buy one of these things (Note: Alexa does allow you to change it to one of 4 pre-selected words, one of which is “computer”!). I have seen amazing setups for home automation. A friend was going to be alone at home for two months and I recommended Alexa controlled light fittings as she already had Alexa’s in her home.. why not, it’s so easy!

How Can we Ensure that Big Data Does not Make us Prisoners of Technology?

Most of you will have interacted with several algorithms already today. In some cases, more algorithms than people. Algorithms are of course simply sets of rules for solving problems, and existed long before computers. But algorithms are now everywhere in digital services. An algorithm decided the results of your internet searches today. If you used Google Maps to get here, an algorithm proposed your route. Algorithms decided the news you read on your news feed and the ads you saw.
Three factors could come together to make an algocracy more than just science fiction:

  1. Big Data, artificial intelligence and machine learning, behavioural science.
  2. The power of Big Data corporations and their central place in providing services that are now essential in our everyday lives raise significant questions about the adequacy of global frameworks for competition and regulation.
  3. We need to anticipate the fundamental questions which Big Data, artificial intelligence and behavioral science present, and make sure that we innovate ethically to shape the answers.

2018-07-13 10:29:48+00:00 Read the full story.
CloudQuant Thoughts… Is it fair that Insurers use your eating and drinking records to set the price of your health insurance? Is it fair for credit card companies to limit your credit if you make a payment to a marriage guidance counselor? We should really have dealt with these questions before allowing corporate giants to run rampant with our data. Privacy must be on by default, the release of data must be opt-in only and must include regular clear weekly summaries of who the data was given to and why. With these simple rules in place, we would have avoided the FB Russia issue.

Facial recognition technology: The need for public regulation and corporate responsibility

All tools can be used for good or ill. Even a broom can be used to sweep the floor or hit someone over the head. The more powerful the tool, the greater the benefit or damage it can cause. The last few months have brought this into stark relief when it comes to computer-assisted facial recognition – the ability of a computer to recognize people’s faces from a photo or through a camera. This technology can catalog your photos, help reunite families or potentially be misused and abused by private companies and public authorities alike.
2018-07-13 00:00:00 Read the full story.
CloudQuant Thoughts… So many stories this week on privacy and security around AI and ML. This is an interesting, detailed and thus long BlogPost by Microsoft in reaction to their employees’ response to Microsoft’s Computer Vision technology being used by ICE. But as people freak out about computer vision and the ability of government to watch you wherever you go, they ignore the fact that you do not need complicated AI and ML to achieve this. We all now carry tracking devices in our pockets.

Algorithms are taking over – and woe betide anyone they class as a ‘deadbeat’

The radical geographer and equality evangelist Danny Dorling tried to explain to me once why an algorithm could be bad for social justice. Imagine if email inboxes became intelligent: your messages would be prioritised on arrival, so if the recipient knew you and often replied to you, you’d go to the top; I said that was fine. That’s how it works already. If they knew you and never replied, you’d go to the bottom, he continued. I said that was fair – it would teach me to stop annoying that person.
If you were a stranger, but typically other people replied to you very quickly – let’s say you were Barack Obama – you’d sail right to the top. That seemed reasonable. And if you were a stranger who others usually ignored, you’d fall off the face of the earth. “Well, maybe they should get an allotment and stop emailing people,” I said. “Imagine how angry those people would be,” Dorling said. “They already feel invisible and they [would] become invisible by design.”
2018-07-12 00:00:00 Read the full story.
CloudQuant Thought… So many negative articles this week. Are we running before we can even walk?

Amazon AI predicts users’ musical tastes based on playback duration

Engineers at Amazon have developed a novel way to learn users’ musical tastes and affinities with artificial intelligence: by using song duration as an “implicit recommendation system.” Bo Xiao, a machine learning scientist and lead author on the research, today described the method in a blog post ahead of a presentation at the Interspeech 2018 conference in Hyderabad, India.
Distinguishing between two similar songs — for instance, Lionel’s Richie’s “Hello” and Adele’s “Hello — can be a real challenge for voice assistants like Alexa. One way to resolve this is by having the assistant always choose the song that the user is expected to enjoy more, but as Xiao notes, that’s easier said than done. Users don’t often rate songs played back through Alexa and other voice assistants, and playback records don’t necessarily provide insight into musical taste.
2018-07-14 00:00:00 Read the full story.
CloudQuant Thoughts… You do not need AI to do this, you can do it with iTunes. It is easy to detect how many times you have pressed skip on a song compared to how many times you have not. It would be very simple for Apple and their ilk to add hard-coded logic to determine what we will like and what we will not like. But do we want to, we seem to be in a world of echo chambers. Listen to different things, go to different places, read different books/opinions. If you do the same as everyone else you are part of the herd. Writing profitable models requires the ability to step back from the herd and observe.

Japan’s Paidy raises $55m for credit card-free online shopping service

Paidy launched its post-pay credit account for ecommerce in 2014, and now claims more than 1.4 million accounts, bringing online shopping to people who do not have, or do not want to use, credit cards.
Once registered, customers make purchases using a mobile phone number and email address with a four digit SMS or voice verification code, before settling a single monthly bill for all their purchases, either at a convenience store, by bank transfer or auto debit.
The firm says that its proprietary models and machine learning mean that transactions are underwritten in seconds, with guaranteed payment to merchants – increasing their revenues.
2018-07-12 15:41:00 Read the full story.
CloudQuant Thoughts… Many of us do not think about those who are excluded from this brave new world of online shopping and easy fast purchases but there are significant numbers of people in the western world who are unable to gain credit. But where does that leave those that Paidy’s algorithm rejects? To quote from the “Algorithms are taking over article” above… “The Chinese government is working towards assigning its citizens a social score by 2020, an algorithm will rate citizens as a desirable employee, reliable tenant, valuable customer – or a deadbeat, shirker, menace, and waste of time”. If you want to see an entertaining dystopian view of this version of the future I would suggest Black Mirror, Season 3 Episode 1 “Nosedive”.

Goldman Sachs Targets 25% Upside for Twitter Stock

Shares of Twitter Inc (NYSE:TWTR) are up 2.6% to trade at $44.95, after Goldman Sachs lifted its price target on the social media stock to $55 from $40 — a more than 25% premium to last night’s close at $43.87. The brokerage firm said the company’s efforts to delete fake accounts are “contributing to incremental ad dollars flowing onto the platform.”
2018-07-12 00:00:00 Read the full story.
CloudQuant Thoughts… Twitter’s high was $74.73 in December 2013, its low $13.725 in May 2016. Can you write a model to predict these kind of movements, perhaps using News reports, perhaps using Sentiment data. Try now on But I would not put much sway in Goldman Sachs’ predictions because…

How Goldman Sachs Lost the World Cup

Goldman Sachs’ statistical model for the World Cup sounded impressive: The investment bank mined data about the teams and individual players, used artificial intelligence to predict the factors that might affect game scores and simulated 1 million possible evolutions of the tournament. The model was updated as the games unfolded, and it was wrong again and again. It certainly didn’t predict the final opposing France and Croatia on Sunday.
2018-07-14 00:00:00 Read the full story.
CloudQuant Thoughts… I’m sorry, I watched this disaster unfold, it was hilarious!

Below the Fold

Traditional statistical methods often out-perform machine learning methods for time-series forecasts

It is impossible today to sit in a meeting in an analytics environment and discuss a methodological approach to a problem without a machine learning (ML) based solution being suggested. There is merit to this; ML techniques from SVM, CART regression trees, to the suite of neural networks (BNN, RNN, LSTM) offer superior predictive capabilities. When turning this predictive capability to time-series forecasting, it would be natural to think these ML algorithms should be the first choice. Well, perhaps not. A recent paper from 3 forecasting experts at the National Technical University of Athens would suggest otherwise; that when it comes to time-series forecasting, traditional statistical techniques such as ARIMA or ETS may in fact offer superior forecasting performance.
2018-07-09 Read the full story.

AI drug discovery startup Verge Genomics raises $32 million

Developing new drugs is a challenging enterprise. It costs pharmaceutical companies an average of $2.7 billion to bring medicine to store shelves, according to the Tufts Center for the Study of Drug Development, and as much as 90 percent of treatments in late-stage trials never come to market because they’re deemed ineffective or unsafe.
Verge Genomics, run by 29-year-old Alice Zhang, is trying to address these problems by making drug discovery faster and cheaper. “Drug companies are looking at one gene at a time. That works for certain diseases, but more complex ones can be caused by hundreds of genes,” she said. “They [also] aren’t typically using human data until they get into clinical trials. We use that data from day one … [those are] some of the ways we’re decreasing the [drug] failure rate.”
2018-07-16 00:00:00 Read the full story (at Venture Beat).
2018-07-16 00:00:00 Read the full story (at Business Insider).

A Gentle Introduction to Credit Risk Modeling with Data Science — Part 2

In our last post, we started using Data Science for Credit Risk Modeling by analyzing loan data from Lending Club.
We’ve raised some possible indications that the loan grades assigned by Lending Club are not as optimal as possible.
Over the next posts, our objective will be using Machine Learning to beat those loan grades.
2018-07-15 17:37:51.531000+00:00 Read the full story.

How Cellular Features Improve ML Accuracy In Phenotyping

Research and discoveries in cell biology have come a long way. Improvements in biological equipments, especially in the area of microscopy, have risen to a cutting-edge level. The precision in obtaining images on a microscopic scale is astonishing. These advances have now presented a challenge of obtaining a vast amount of image data in crispy-clear quality.
Although machine learning (ML) has resolved this problem with quick efficacy by using automation, it fails to utilise information from microscopic elements such as cells and tissues. ML only considers the properties or features surrounding the data. It does not dig in deep about the cellular features that determine or influence the extrinsic (environmental) factors have on humans.
2018-07-16 05:44:17+00:00 Read the full story.

Top 10 Kaggle Data Notes of the last week…

  1. S&P 500 Simple Forecasting with Prophet  (Facebook’s library for time series forecasting)
  2. arXiv Data Analysis: Computation and Language Papers
  3. Dysonian SETI with Machine Learning
  4. Content-based Recommender Using Text Mining
  5. Predicting Cast in TV’s Frasier Based on Dialog
  6. Shift in Votes Between Political Parties
  7.  Data Science Glossary on Kaggle
  8. Kaggle Dataset #1: Stanford Cars Dataset
  9. Kaggle Dataset #2: Columbia Object Image Library
  10. Kaggle Dataset #3: CelebFaces Attributes Dataset

2018-07-12 00:00:00 Read the full story.

Particle Swarm Optimisation in Machine Learning – Towards Data Science

“Gradient Descent will not make you an expert at Machine Learning”
Most of the articles you would have come across must have talked about Gradient Descent whether it is a Simple Linear Regression or Neural Networks. In this article I will introduce a technique i.e. Particle Swarm Optimisation (PSO) to you. No doubt that Gradient Descent is a good optimisation technique which works great for convex functions & low dimensional space but you can expect bizarrely good results with PSO.
2018-07-15 14:51:52.071000+00:00 Read the full story.

New dog, old tricks: Fintec data management in the cloud

Every cloud has a silver lining, at least that’s what our elders drummed into us, an early example of expectation management and how to deal with life’s challenges. Other teaching proxies included looking after the family dog, a somewhat more practical and physical learning experience, involving across the board responsibilities of duty of care, maintenance and welfare for another living being.
Cloud computing for data management offers similar canine inspired opportunities for curating what can turn out to be a beast or a docile pet in equal measure. So let’s run through the many considerations of responsible cloud ownership and their interpretation
2018-07-16 00:00:00 Read the full story.

AI has arrived – don’t leave data strategy behind

Over the last century computers have sped up exponentially, according to Moore’s law, whilst their production costs have halved every 18 months. That pace is increasing even more with quantum computers that are already running at 50 qubits. By comparison, a 100 qubit computer could theoretically be more powerful than all the supercomputers in the world combined. That’s a lot of computing power.
Meanwhile, the amount of data produced in the world is advancing at a staggering pace. Today we are generating 16.3 zetabytes of data per year and that number is set to grow to 163 zetabytes by 2025, according to IDC. That’s a 10-fold increase in just seven years.
For this reason, Gartner bills AI as the most disruptive technology to emerge over the next decade, because it has the power to process vast pools of data and turn them into critical insights that enhance lives. Although AI has been around since the 1960s, progress has been slow – mainly due to the lack of data and poor computer power. Both are now available in abundance.
2018-07-13 00:00:00 Read the full story.

Oval Money to link automated savings app to Twitter postings

Oval Money, the automated savings platform, has enabled users to double their monthly savings by tapping into their social media habits. The average UK user was putting away just £63 per month on the platform when it launched 12 months ago. In just a year, the figure has leapt to £130, equivalent to an annual jump from £756 to £1,560.
The firm has attributed the increase in part to users linking the app with their Facebook activity, so they made automated savings every time they posted.
2018-07-16 10:13:00 Read the full story.

Fidelity’s Welo: Forget Tech; AI to Have Greater Impact on Industrials

Technology companies may believe that they have cornered the market for artificial intelligence, but investors may want to set their sights toward industrial companies to get some AI exposure.
…Incorporating AI into industrial companies’ processes will offset higher inflation, notably wage inflation. After all, downtime is extremely costly for industrial companies, and if AI-enabled sensors and software can run some processes more efficiently, it could help companies minimize this issue. The fund manager pointed to a gas turbine as one example. If a company can use AI to anticipate a potential failure, that could bring a massive benefit to that business.
2018-07-12 12:06:00-06:00 Read the full story.

Intel Adds Structured ASICs to Product Lineup With eASIC Acquisition

Intel over the past several years has adapted to an increasingly data-centric world by diversifying its product lines beyond CPUs for PCs and servers to offer a broader range of computing systems including programmable chips, ASICs and, soon discrete GPUs.
“Customers designing for high-performance, power-constrained applications in market segments like wireless, networking and the internet of things (IoT) sometimes begin deployments with FPGAs for fast time-to-market and flexibility,” McNamara wrote in a post on the company blog.
“They then migrate to devices called structured ASICs, which can be used to optimize performance and power-efficiency. A structured ASIC is an intermediary technology between FPGAs and ASICs. … The addition of eASIC will help us meet customers’ diverse needs of time-to-market, features, performance, cost, power and product life cycles.”
2018-07-12 00:00:00 Read the full story.

Banks stand to reap $512 billion revenue boost from ‘intelligent automation’

With banks the world over exploring the business case for AI and robotic automation, a new report estimates that hundreds of billions of dollars in additional revenue may be up for grabs as the focus moves away from costs savings to income generation. To date, automation technologies, such as RPA (Robotic Process Automation), have been implemented by the financial services industry to drive down costs and create efficiencies. But a new front is opening, with the deployment of machine learning tools for customer-facing interactions seen as new weapon in the armoury of banks facing the threat of competition from Big Tech players like Amazon and Alphabet.
2018-07-12 09:32:00 Read the full story.

BattleFin Announces Pending Launch of Alt Data Assessment & Testing Platform

The Ensemble platform allows hedge funds and investment firms to test alternative data sets that they found at the BattleFin Alternative Data Discovery Days or that are part of the BattleFin Alternative Data Accelerator. Standardized NDAs, Testing Agreements, & Compliance checks will be completed once.
One interesting feature of the Ensemble platform is that fundamental asset managers looking to add alternative data to their research process will be able to leverage BattleFin’s data scientist community by scoping out projects to track and predict specific Key Performance Indicators using specific alternative data sets.
2018-07-12 08:36:10+00:00 Read the full story.

Finra Taps AI to Stop ‘Mini Manipulation’

Artificial intelligence may help market manipulators skirt the rules, but the same technology is helping regulators detect the once-hidden behavior. “The machine is learning the activity and identifying it for us as well as the changes in activity so that we can stay slightly ahead of it,” said Gene DeMaio, senior vice president, market regulation at Financial Industry Regulatory Authority, during a recent call hosted by the STA Foundation, the educational arm of the Security Traders Association. “As we train the program a little bit more as time goes by, we are getting better and better exceptions.”
The technology has helped the self-regulatory organization to identify potential instances of “mini-manipulation,” which is also known as cross-asset manipulation that uses equities and options. Mini-manipulation occurs when a trader holds an options position and attempts to move the underlying equity to change the price of the equity.
2018-07-13 21:40:39-04:00 Read the full story.

AI doctor app Babylon fails to diagnose heart attack, complaint alleges

An artificial intelligence app that claims to be able to diagnose medical conditions better than human doctors fails to properly identify heart attacks, it has been claimed.
Babylon, which lets people book virtual appointments with GPs and receive prescriptions through its app, also operates a tool that uses artificial intelligence to automatically diagnose health problems. Customers can type their symptoms and receive a diagnosis through the app.
2018-07-13 00:00:00 Read the full story (PAYWALL)

Forget the sex, the hot new book about Google is an important reminder of what Sergey and Larry are really after

A new book has generated headlines over Sergey Brin’s ‘playboy’ period during Google’s early days, but there’s another big takeaway in the book that readers in the tech industry should not ignore. The book explores a deep-rooted and intense desire that’s driven Google’s founders all these years.
Brin and cofounder Larry Page stumbled onto the magic search business but that was never their main interest. From the start, Google was always intended to be an AI company. They’re now closer to that vision than they have ever been, and what comes next could make search look like a footnote in Google’s history.
2018-07-14 00:00:00 Read the full story (PAYWALL).

22 books Wall Streeters think everyone should read this summer

Everyone needs a good beach read. We asked three Wall Streeters for the books they are having trouble putting down as the weather warms up. Along with titles about finance and successful figures, we also received recommendations for thrillers, novels, and historic non-fiction that make the list of perfect page-turners.

  • ‘Sapiens: A Brief History of Humankind’ by Yuval Noah Harari
  • ‘Bad Blood’ by John Carreyrou
  • ‘Extreme Ownership – How U.S. Navy SEALS Lead and Win’ By Jocko Willink and Leif Babin
  • ‘A Life Well Played’ by Arnold Palmer
  • ‘Lords of Finance: The Bankers Who Broke The World’ by Liaquat Ahamed
  • ‘How Not to Be Wrong: The Power of Mathematical Thinking’ by Jordan Ellenberg
  • ‘Elon Musk: Telsa, SpaceX and the Quest for a Fantastical Future’ by Ashlee Vance
  • ‘Hillbilly Elegy: A memoir of a family and culture in crisis’ by JD Vance
  • ‘Spymaster’ by Brad Thor
  • ‘The Orphan Master’s Son’ by Adam Johnson
  • ‘King Leopold’s Ghost’ by Adam Hochschild
  • ‘Prediction Machines: The Simple Economics of Artificial Intelligence’ by Ajay Agrawal, Joshua Gans, Avi Goldfarb
  • ‘The Cuban Affair’ by Nelson DeMille
  • ‘Principles’ by Ray Dalio
  • ‘Radical Candor’ by Kim Scott
  • ‘Medium Raw: A Bloody Valentine to the World of Food and the People Who Cook’By Anthony Bourdain
  • ‘The Other Woman’ by Daniel Silva
  • ‘The Hard thing about Hard Things’ by Ben Horowitz
  • ‘The Perfect Weapon’ by David E. Sanger
  • ‘The Underground Railroad’ by Colson Whitehead
  • ‘The Republic of Pirates’ by Colin Woodard
  • ‘The Soul of America’ by Jon Meacham

2018-07-16 00:00:00 Read the full story.

This news clip post is produced algorithmically based upon CloudQuant’s list of sites and focus items we find interesting. If you would like to add your blog or website to our search crawler, please email We welcome all contributors.
This news clip and any CloudQuant comment is for information and illustrative purposes only. It is not, and should not be regarded as “investment advice” or as a “recommendation” regarding a course of action. This information is provided with the understanding that CloudQuant is not acting in a fiduciary or advisory capacity under any contract with you, or any applicable law or regulation. You are responsible to make your own independent decision with respect to any course of action based on the content of this post.