June 15, 2017

Issue 236: CognitionX Data Science, AI and Machine Learning Briefing


June 14, 2017

Issue 235: CognitionX Data Science, AI and Machine Learning Briefing


June 14, 2017

Issue 234: CognitionX Data Science, AI and Machine Learning Briefing


June 12, 2017

Issue 233: CognitionX Data Science, AI and Machine Learning Briefing

Alphabet sells Boston Dynamics. AI + the gender gap. Ocado’s robot-packed factory. http://cognitionx.com/news-briefing/

If you haven’t seen Boston Dynamics animal-like, sort-of-strange robots, then go see them now….what’d you think? Pretty crazy, no? Alphabet has just announced that they are selling Boston Dynamics to Japan’s Softbank (who we’ve seen in the news recently for their legendary $100B tech fund).

As to why they decided to do this, it is not entirely clear. Some suggest that it is in line with the fiscal philosophy of Alphabet CFO, Ruth Porat, who is “sweeping out moonshot projects and units that, even if they are cool, stand very little chance of kicking out a practical, profitable product anytime soon.”

Also, congrats to CognitionX advisor and investor, Brent Hoberman, who has unveiled a new $60 million European startup fund.

Want to know which AI companies are hot and ripe for investment? We’ve got just the list for you…

Ladies and gentleman, we are pleased to announce (drum roll, please…) the finalists for the inaugural CogX AI Innovation Awards. The AI Innovation Awards have been judged by a panel of experts from the advisory board of CognitionX. Entrants were judged on their product idea, the team and the long-term sustainable advantage of the product. The result is a shortlist of finalists representing the most innovative and disruptive businesses in the world of AI. Check out the finalists here and don’t forget to buy your tickets to the Awards Gala Dinner (June 21st) before they sell out.

To explore what the next major AI investment opportunities will be and how to get the timing right, come to hear about Investing in AI at CogX on Tuesday June 21st.


Tabitha UntilTheBotsTakeOver Goldstaub

P.S. 8 days till CogX.  Be there or be intelligence artificial.

Deal of the Day

Alphabet sells off robot maker Boston Dynamics to Softbank

Google’s parent company Alphabet has sold robotics firm Boston Dynamics to Japan’s Softbank for an undisclosed sum, over a year after putting it up for sale. As part of the deal, Softbank also purchased Schaft, a lesser known Alphabet robotics subsidiary.

Boston Dynamics is best known for its “BigDog” line of quadrupedal robots, first developed to serve as artificial pack animals as part of a military contract with America’s Darpa research agency. The robots, alongside the humanoid “Atlas” machine and wheeled hybrid “Handle”, have become viral sensations online, thanks to their eerily lifelike movements, unnerving high-pitched motor sounds and the videos themselves, which feature Boston Dynamics staff torturing their creations with pushes, kicks, shoves and heavy weights to demonstrate their versatility and reliability.

+Like robots? Come to one of CogX’s most popular panel, Impact of AI on Design, Manufacturing, Robotics & IoT.


Artificial intelligence can now predict suicide with remarkable accuracy

Colin Walsh, data scientist at Vanderbilt University Medical Center, and his colleagues have created machine-learning algorithms that predict, with unnerving accuracy, the likelihood that a patient will attempt suicide. In trials, results have been 80-90% accurate when predicting whether someone will attempt suicide within the next two years, and 92% accurate in predicting whether someone will attempt suicide within the next week.

The prediction is based on data that’s widely available from all hospital admissions, including age, gender, zip codes, medications, and prior diagnoses. Walsh and his team gathered data on 5,167 patients from Vanderbilt University Medical Center that had been admitted with signs of self-harm or suicidal ideation. They read each of these cases to identify the 3,250 instances of suicide attempts. You can find the paper here (paywall).

Future of Health

Robotic tool to reduce the consequences of a stroke

Researchers have developed a new robotic tool that may in combination with standard rehabilitation programme improve the mobility of patients surviving a stroke. Stroke, known as a leading cause of long-term disability, is a sudden loss of brain function, caused by the interruption of blood flow to the brain or the rupture of blood vessels in the brain and an estimated 1,50,000 people die from it each year. As a consequence of stroke, the survivors are often left with muscle over-activity, including spasticity, characterised by tight or stiff muscles and an inability to control those muscles.

The study showed that the new robotic tool may help in assessing muscle overactivity and movement dysfunction in survivors of stroke. The rehabilitation robotic system was found to quantitatively measure the three degree-of-freedom (DOF) impedance of human forearm and wrist in minutes.

Business Impact of AI

City analysts disappointed by Ocado’s robot-packed factory

“It looks like a hybrid of dodgems and Pac Man,” says Shore Capital’s George Mensah. Analysts who made the trip to Andover say they were impressed by the ingenuity on display. The system aims to be more efficient than humans. It marks an upgrade from the conveyor belts used at Ocado’s two other CFCs, at Dordon in Warwickshire and Hatfield in Hertfordshire.

But if the showcase was meant to inspire investor confidence in Ocado’s ability to roll out its business model across Europe, it failed. The day after the tour, Ocado shares slumped by 7.9pc, their biggest one-day fall for nine months. The Andover facility, which opened last November, is only running at about 10pc capacity and still has obvious glitches that needed ironing out, analysts told clients on their return.

Ethics Question for the Day

AI and bridging the gender gap

Debrah Lee Charatan, co-founder, principal, and president of BCB Property Management presents 4 essential steps for giving women entrepreneurs a leg-up:

  1. Have a dream and stick to it
  2. Don’t be afraid to ask for help
  3. Learn to communicate
  4. Know your worth

Also, a new AI platform from The Circular Board in partnership with Dell and Pivotal claims to be the missing puzzle piece needed for women to finally receive their due. Alice connects women with mentors and potential funding partners, along with resources like financial planning advice and relevant event listings. The goal is to provide female entrepreneurs with everything they need to eliminate the gender gap and succeed in a workforce that is actively pushing back against them.


DeepLoco: dynamic locomotion skills
using hierarchical deep reinforcement learning 

Learning physics-based locomotion skills is a difficult problem, leading to solutions that typically exploit prior knowledge of various forms. In this paper, the authors aim to learn a variety of environment-aware locomotion skills with a limited amount of prior knowledge.

Podcasts We Love

Intelligent Autonomous Robots with Ilia Baranov – TWiML

In this episode of TWiML,  Ilia Baranov, engineering manager at Clearpath Robotics discusses the field of autonomous robots. Ilia is responsible for setting the engineering direction for all of Clearpath’s research platforms. Ilia likes to describe his role at the company as “both enabling and preventing the robot revolution.” He’s a longtime contributor to the Open Source Robotics Community and ROS, an open source robotic operating system. He is the also the managing engineer of the PR2 support team at Clearpath and leads the technical demonstration group.

In their conversation, they cover a lot of ground, including what it really means to field autonomous robots, the use of autonomous robots in research and industrial environments, the different approaches and challenges to achieving autonomy, and much more!

I’ve been making some changes based on Feedback. Would love to hear from more of you. Please do click to share your thoughts!


June 12, 2017

Issue 232: CognitionX Data Science, AI and Machine Learning Briefing

Microsoft buys Israeli cyber firm for $100M. IoT for bedbugs. IBM’s Science for Good. http://cognitionx.com/news-briefing/

Data is everything in elections. Whether that be using your social media data to send you targeted ads or using data to understand the political parties positions better, you just can’t ignore it.

With a hung parliament, data could will be even more important to help us decipher what’s going on. Thankfully Represent.me analysed the main party manifestos across over 160 statements and policies. Their purpose was to help voters find their best match and have their say on the key issues. In doing, they also show how much the parties agreed or disagreed with each other. You can compare them individually here.

You can still participate if you’d like to have your say on the issues and guide the new MPs towards the world you’d like to see. Represent.me will be sending the results to all the new MPs in the next few weeks to show what voters care about in detail. You can find the survey here.

Go forth and get involved,

Tabitha UntilTheBotsTakeOver Goldstaub

P.S. 11 days till CogX.  Be there or be intelligence artificial.

CogX Highlights

The latest from companies coming to CogX

HR- Beamery: Companies are moving away from traditional candidate management and applicant tracking systems, instead, turning to machine-learning enabled ones to take “smarter” hiring decisions. A case in point is Beamery, a Candidate Relationship Management (CRM) software that helps enhance the candidate experience.

Health- Biobeats: A study has found that one in two City workers in London were in fact more stressed when they got home than they were at work. It was commissioned by Biobeats, which calls itself “a digital health and artificial intelligence business”, in association with the University of Sussex and was based on “heart rate variability”.

Legal- Luminance has just been chosen by top-ranked Stockholm law firm Cederquist which will use its market-leading AI technology to improve the efficiency of its M&A due diligence. Founded in 2010 by mathematicians from the University of Cambridge, Luminance is quickly expanding its client base across Europe and further afield, “demonstrating the benefits of our technology for lawyers in all jurisdictions,” according to Luminance CEO Emily Foges.

Deal of the Day

Microsoft agrees to buy US-Israeli cyber firm Hexadite for an estimated $100M

Microsoft confirmed yesterday that it has acquired Hexadite, an Israeli startup that uses AI to identify and protect against attacks. The idea is to expand Microsoft’s existing security portfolio with an infusion of new technology based around new innovations in areas like AI and machine learning.

“Our vision is to deliver a new generation of security capabilities that helps our customers protect, detect and respond to the constantly evolving and ever-changing cyberthreat landscape,” said Terry Myerson, executive vice president, Windows and Devices Group, Microsoft, in a statement. “Hexadite’s technology and talent will augment our existing capabilities and enable our ability to add new tools and services to Microsoft’s robust enterprise security offerings.”

Products We Love

The bedbugs won’t bite thanks to this IoT device

From smart thermostats that intelligently control the temperature in a room to smart devices like the Roomba vacuum cleaner that hoover up dirt, connected devices are capable of carrying out all sorts of functions in the places we sleep. Here’s one we’ve not come across before, though: A Wi-Fi enabled system smart device that’s designed to alert you of any bedbug outbreaks in the making.

Developed by the company Delta Five, the currently available Automated Insect Monitoring System takes the form of a small, 3-inch box that monitors for insects around the clock. “Given our initial focus on hospitality, and our core competencies in robotics, unattended sensors, computer vision, and machine intelligence, we recognized a significant gap between available products and actual need as it relates to bedbugs,” Jason Janét, Delta Five’s CEO, told Digital Trends.


IBM turns to AI to solve poverty, hunger, and illiteracy

This week, IBM announced the launch of Science for Social Good, a new program that partners IBM researchers with postdoctoral academic fellows and nonprofits to take on societal issues through data.

With the new initiative, IBM announced 12 projects planned for 2017. Each Science for Social Good project aligns with one or more of the 17 Sustainable Development Goals, the United Nations’ blueprint to address some of the globe’s biggest inequalities and threats by the year 2030. Six pilot projects were conducted in 2016 in order to develop the Science for Social Good initiative. These projects covered a broad range of topics, such as health care, humanitarian relief, and global innovation.

Future of Transportation

Transdev and Delphi announce partnership for autonomous transportation

The first open road, driverless on-demand mobility service in EU is set to be tested in Paris-Saclay and in Rouen.

Yann Leriche, chief performance officer and lead of the B2C business line at Transdev group said, “With Delphi’s expertise in driverless technology, we will accelerate our capabilities to develop a global autonomous transportation system, from client interface to vehicle intelligence. Combining the strengths of our two companies, leaders in their field, will enable us to introduce innovative driverless services in our current and future operations, confirming the position of Transdev as a pioneer in integrating autonomous transport systems into global mobility networks.”

Business Impact of AI

How Boeing plans to test the technologies for future self-flying passenger airplanes

The Boeing Co. says it’s on a path that could ultimately lead to self-flying commercial passenger airplanes, starting with simulations and ground-based experiments this year and progressing to flight tests by 2019.

“In commercial air transportation, we are not yet in this business,” said Mike Sinnett, vice president of product development at Boeing Commercial Airplanes. Boeing has long experience with autonomous systems – ranging from underwaterrobots and robotic surfboards to the military-grade drones built by its Insitu subsidiary. There’s even a robotic co-pilot system called ALIAS that made a simulated Boeing 737 landing all by itself.

Open Source

Facebook’s image-recognition tech
Is teaching 40,000 images a second to understand context

Now, Facebook’s AI and machine-learning teams have developed methods for training data sets—the procedure that’s used to teach visual recognition models to distinguish between large numbers of images—that are faster than anything else available today. The company said it has come up with a system that’s capable of training 40,000 images per second, making it possible to train on a 1k data set—the industry standard training set—in less than sixty minutes, and with no loss of quality. Until now, that was something that could take days, or even months to do.

Keeping with its philosophy of sharing its work freely and widely, Facebook says it’s open-sourcing the hardware stack it developed to achieve the improved training times. And that could benefit many companies that are taking in massive amounts of visual data and need to process it quickly.

Video Killed the Radio Star

How data is driving golf forward

WIRED explores golf’s data revolution, using Microsoft Cloud-enabled sensors to improve the swing, putt and grip of amateurs and pros alike.

Dates for Your Diary

I’ve been making some changes based on Feedback. Would love to hear from more of you. Please do click to share your thoughts!


June 12, 2017

Issue 231: CognitionX Data Science, AI and Machine Learning Briefing

Therapist chat bot. AI + creativity. AI-powered marketing for the win. http://cognitionx.com/news-briefing/

Can AI be creative? Sure, we’ve seen it create things, from email subject lines to music, but can it reach that next level of providing a ‘nonlinear solution to a problem’ or ‘thinking outside the box’?

 Recently, at The 2017 World Science Festival, they tackled this very question. Peter Ulric Tse, a neuroscientist at Dartmouth College, recently remarked that we should be focused on using AI to understand creativity, as opposed to exhibiting it. Some, such as Lav Varshney, another member of the onstage panel, are even working on a mathematical theory of creativity.

If you’re thirsting to learn more about The Impact of AI on creativity, then you should come along to CogX on Tuesday June 20th for a session featuring Luba Elliott (AI Advocate, Producer and Researcher), Siavash Mahdavi (CEO & Founder of AI Music), Ed Newton-Rex (Chief Executive Officer & Founder of Jukedeck), Matthew Postgate (Chief Technology & Product Officer at the BBC), and more.


Tabitha UntilTheBotsTakeOver Goldstaub

P.S. 12 days till CogX.  Be there or be intelligence artificial.

Chat Bots 4Eva

Meet Woebot, the chat bot therapist

Created by a team of Stanford psychologists and AI experts, Woebot uses brief daily chat conversations, mood tracking, curated videos, and word games to help people manage mental health. After spending the last year building a beta and collecting clinical data, Woebot Labs Inc. just launched the full commercial product—a cheeky, personalized chatbot that checks on you once a day for the price of $39 a month.

Finding the time and money to pay for talk therapy sessions is out of reach for many, so a chatbot could be a helpful stopgap for psychiatry. But Woebot’s creators believe it has the potential to actually improve on human therapists. “It’s almost borderline illegal to say this in my profession, but there’s a lot of noise in human relationships,” says Alison Darcy, one of the psychologists behind Woebot, and the company’s CEO.

+CogX will feature a session on The Impact of AI on Mental Health on Tuesday June 20th.

Future of Transportation

Here’s a sneak peek of Fisker’s all-new, ultra-luxury electric car

Legendary car designer Henrik Fisker released some new images today of the forthcoming EMotion, an ultra-luxury electric vehicle with a range of 400 miles and a top speed of 161 mph. The designer’s company also revealed some new details about the car that could help revitalize Fisker’s reputation in the EV market.

Fisker EMotion will come with fully autonomous capabilities, the company says. “The front end’s bright aluminum centerpiece houses a LIDAR system behind a tinted screen. Side mirrors conceal two cameras, which enable panoramic, 360-degree views to the driver.” Other details, such as whether the vehicle’s autonomous features are best used on highways only or in urban environments as well, are still unknown.

+Don’t miss the CogX session on self-driving cars (Wednesday, June 21).


How Harley-Davidson used AI to increase sales leads by 2930%

Asaf Jacobi’s use of ​Adgorithms’s AI-driven marketing platform at his New-York based Harley-Davidson dealership is a good use case of the major boost AI can give to a company’s sales.

His dealership went from getting one qualified lead per day to 40. In the first month, 15% of those new leads were “lookalikes,” meaning that the people calling the dealership to set up a visit resembled previous high-value customers and therefore were more likely to make a purchase. By the third month, the dealership’s leads had increased 2930%, 50% of them lookalikes, leaving Jacobi scrambling to set up a new call center with six new employees to handle all the new business.

Future of Art

Making music with AI: an introduction

In this first post in Sync Project’s latest series, Alex D. starts by discussing a cool project called Folk-RNN. Researchers entered thousands of transcribed examples of Celtic Folk music into a deep-learning system that learned from the MIDI song information to create new melodies. The researchers were surprised to find that the system could cook up “authentic” sounding melodies once every 5 times it tried.

Researcher Dr Oded Ben Tal, who is also a musician and used some of those melodies to make new original music, pointed out that technique could also help musicians make more compelling material.


Stealthy startup Alitheon raises $2.8M for digital fingerprinting technology

A startup called Alitheon wants to teach computers to recognize objects, and it’s attracted $2.8 million in funding to do it, according to a report filed with the Securities and Exchange Commission.

According to the company’s website, the patented technology identifies points on an item’s surface and creates a set of metadata. Then it turns those metadata into a “Featureprint” and stores it to the cloud, where the recognition process resides. As the item moves through the supply chain or a delivery process – for example, airline baggage handling – a standard-issue smartphone can snap images and compare them to the Featureprint for authentication.

Future of Health

UC Berkeley partners with blockchain startup for health data research 

The University of California, Berkeley is working with blockchain startup Bitmark on a pair of research studies focused on secure data sharing. Centred on public health and medical information, the two research projects will be funded by Bitmark. The project comes months after Bitmark raised $1.7m in funding, drawn from a group that included Cherubic Ventures and Digital Currency Group, among others.

In one, study participants will contribute data about their diabetes remission using Bitmark’s tech. A second study will focus on whether the availability of solutions based on the technology would encourage more involvement from test subjects that are contributing personal health data.

+DeepMind has also been applying blockchain tech to healthcare. Get your session pass (£199) to hear Dr Dominic King (Clinical Lead, DeepMind) speak at CogX .

Business Impact of AI

Intel teams up with DARPA to make HIVE Big Data platform a reality

Intel and the US Defense Advanced Research Projects Agency (DARPA) have announced a new collaboration designed to create a new, powerful data-handling and computing platform which leverages machine learning (ML) and artificial intelligence (AI) technologies.

The tech giant and US agency said DARPA’s program, dubbed the Hierarchical Identify Verify & Exploit (HIVE), has the potential to surpass current hardware used to handle Big Data by up to 1,000 times in performance-per-watt, and Intel’s technological expertise has been chosen to push HIVE research and development further. The joint research program, which is estimated to take 4 and a half years, is worth over $100 million.

Podcasts We Love

Futuretech interview on genomics and AI

Richard Jacobs from the FutureTech podcast talks with
Eleonore Pauwels of the Wilson Institute about genomic technologies with a focus on their convergence with AI, as well as a host of other topics such as self-driving tech.

Eleonore Pauwels is an international science policy expert, who specializes in the governance of emerging technologies, including genomics and digital and bio-engineering.  At the Wilson Center, she is the Director of Biology Collectives, and Senior Program Associate within the Science and Technology Innovation Program.

Dates for Your Diary

I’ve been making some changes based on Feedback. Would love to hear from more of you. Please do click to share your thoughts!


June 7, 2017

Issue 230: CognitionX Data Science, AI and Machine Learning Briefing

DeepMind’s at it again. Sesame Street + IBM Watson. Turn your smartphone into a medical device. http://cognitionx.com/news-briefing/

Lyft claims that their latest partnership with nuTonomy will lead to self-driving taxis on the roads in Boston by the end of the year. Their approach differs from Uber’s who is developing self-driving tech in-house.

We’ll be discussing how self-driving technology will completely change the way we conceive of vehicle ownership, travel, and transportation at CogX’s session on the Impact of AI on Transportation & Smart Cities with experts including MC Srivas (Chief Data Architect, Uber), Stan Bolan (CEO of FiveAI) and Catherine Faiers (COO, Addison Lee).

Make sure you get your tickets..you can grab your session pass here.


The News Team

P.S. 13 days till CogX.  Be there or be intelligence artificial.


A neural approach to relational reasoning (DeepMind)

Modern deep learning methods have made tremendous progress solving problems from unstructured data, but they tend to do so without explicitly considering the relations between objects.

In two new papers (available here and here), DeepMind explored the ability for deep neural networks to perform complicated relational reasoning with unstructured data. In the first paper – A simple neural network module for relational reasoning – they describe a Relation Network (RN) and show that it can perform at superhuman levels on a challenging task. While in the second paper –  Visual Interaction Networks  – they describe a general purpose model that can predict the future state of a physical object based purely on visual observations.

+Fascinated by DeepMind? Come to
CogX to meet DeepMind’s Dr Domic King and hear his thoughts on .


Sesame Workshop and IBM Watson partner on platform to help kids learn

Sesame Workshop and IBM Watson announced yesterday that they are creating a vocabulary app and the Sesame Workshop Intelligent Play and Learning Platform. The new platform will be used by Sesame Workshop and IBM to create a series of cognitive apps, games, and toys to help kids learn. This is the first public action announced by the partnership, which was formed more than a year ago.

The platform will run on IBM Cloud and the partners invite the “ecosystem of software developers, researchers, educational toy companies, and educators to tap IBM Watson cognitive capabilities and Sesame Workshop’s early childhood expertise to build engaging experiences to help advance children’s education and learning,” according to a statement from IBM Watson.

+IBM will be at CogX, but Sesame Street couldn’t make it…sos.

Products We Love

Generative adversarial networks strike again

Just a few months ago we saw pix2pix with cats. Draw something in a little box and an algorithm will try to interpret it as a cat and then fill in the colors and textures according to a machine learning model training on thousands of cat images. This was the first iteration of the pix2pix project. Now, we have human faces and things have become fully uncanny.

The pix2pix project demonstrates something pretty profound about machine learning circa 2017: It’s awful at generating new images, or at least meaningful new images. Machine learning is better at classifying existing images, but, even then, things drop off dramatically as we move beyond a handful of really robust object-recognition models. Pix2pix is based on a promising technique called generative adversarial networks (GANs). GANs work by training generative models that seek to minimize a particular “loss function” according to a prediction that the generated image is fake or real.

Education and Advice We Rate

The evolution of Gradient Descent

Which optimiser should you use to train our neural network? Tensorflow gives you lots of options, and there are way too many acronyms. In this video, Siraj Raval goes over how the most popular ones work and in the process shows how gradient descent has evolved over the years.

Code from this video (with coding challenge) can be found here.

Podcasts We Love

Senosis Health
is turning smartphones into the ultimate medical device

Check out GeekWire’s new Health Tech podcast, exploring the frontiers of digital health. In this first episode, Clare McGrane speaks to  Shwetak Patel. Through his work at the UW and his new startup, Senosis Health, Patel is developing apps for smartphones that can measure health almost as well as expensive medical equipment found in hospitals.

“Our idea has been: How do we repurpose the sensors that are already on a mobile phone to do similar things that you would find in a clinician’s office or at a hospital? And so, we’re looking at how to use microphones, the camera, the flash, the accelerometer, the gyro in new ways that people never used them before,” he said. To that end, they have created apps to measure blood pressure, hemoglobin in the blood, the capacity in one’s lungs to diagnose asthma, and more.

Business Impact of AI

Apple announces new machine learning API to make mobile AI faster

It’s hard to catch up with all the WWDC updates from Apple. I’ve found this video to be a good summary.

Like the rest of the tech world, Apple wants to make AI on your mobile device as fast and powerful as possible. That’s why the company unveiled a new machine learning framework API for developers named Core ML.

The key benefit of Core ML will be speeding up how quickly AI tasks execute on the iPhone, iPad, and Apple Watch. This could cover everything from text analysis to face recognition, and should have an effect on a broad category of apps. It means, says Apple, that image recognition on the iPhone will be six times faster than on Google’s Pixel. Core ML will support a number
of essential machine learning tools, including all sorts of neural networks (deep, recurrent, and convolutional), as well as linear models and tree ensembles.


Fujitsu partners with 1QBit to apply quantum-inspired tech to AI

Fujitsu has partnered with Vancouver-based 1QBit to apply quantum-inspired technology to artificial intelligence.

1QBit, which develops hardware-agnostic quantum and quantum-inspired software to solve computational challenges, will work with Fujitsu to develop applications using AI developed for use with quantum computers.The partnership will focus on areas of combinatorial optimization, the process of finding an optimal object in an extremely large dataset, and machine learning. Eventually, the two companies plan to provide a variety of services that combine 1QBit’s software and expertise in building applications which benefit from the capabilities of quantum computers.

Dates for Your Diary

I’ve been making some changes based on Feedback. Would love to hear from more of you. Please do click to share your thoughts!


June 6, 2017

Issue 229: CognitionX Data Science, AI and Machine Learning Briefing

Apple’s new smart speaker. Microsoft’s open source deep learning toolkit. Andrew Ng interview. http://cognitionx.com/news-briefing/

As the rate of automation increases, the rate of mentions of Universal Basic Income increases. Even the Green Party called for UBI in election manifestos.

Yuval Noah Harari, author of Sapiens, wrote in an article this week that the term itself is highly ambigious and the concept is problematic, with both the words ‘universal’ and ‘basic’ being poorly defined or determined.

He asks: who should be paying this income (national governments? municipalities?) and to whom? He argues that it is particularly difficult to determine what sort of basic needs it should provide: education? healthcare?

Do you think that Universal Basic Income (however it may be defined) is a viable solution?

Yes          No


Tabitha UntilTheBotsTakeOver Goldstaub

P.S. 14 days till CogX. We are announcing the nominees for the CogX AI Awards…don’t forget to book your ticket before they sell out. Be there or be intelligence artificial.


Apple is upping its AI game big time

Jordan Novet from CNBC discusses how Apple is catching up to Google in AI. During the Monday keynote at the company’s 2017 Worldwide Developers Conference, the company announced a bunch of innovations in AI, with perhaps the biggest one being their Google Home/Amazon Alexa competitor, HomePod:

  • A HomePod speaker that plays music and lets you talk to Siri. The Google Home does similar things but comes with the Google Assistant instead of Siri.
  • Siri now has the ability to translate languages, something Google has provided through Google Translate since 2006.
  • Siri will soon be front and center on the Apple Watch, just like the Google Assistant has started showing up on Android Wear smartwatches.

+Shiva Rajaraman, Product Lead at Apple will be at CogX.

Something You Might Have Missed From Last Week

Microsoft releases open-source toolkit to accelerate deep learning

The full release of Microsoft Cognitive Toolkit 2.0 for use in production-grade and enterprise-grade deep learning workloads includes hundreds of new features incorporated since the beta to streamline the process of deep learning and to ensure the toolkit’s seamless integration throughout the wider AI ecosystem.

New with the full release is support for Keras, a user-friendly open-source neural network library that is popular with developers working on deep learning applications.


Robots will be more useful if they are made to lack confidence

Overconfident AI can cause all kinds of problems, says Dylan Hadfield-Menell at UC Berkeley. Take Facebook’s newsfeed algorithms, for example. These are designed to suggest articles and posts that people want to see and share. Yet by following this remit unquestioningly, they have ended up filling some people’s feeds with fake news.

For Hadfield-Menell and his colleagues, the answer is to make AIs that seek and accept human oversight. “If Facebook had this thinking, we might not have had such a problem with fake news,” he says. Check out the research paper here.

+Like robots? Come to CogX to meet Sophia, Hanson Robotics’ latest creation and her creator, David Hanson (CEO and Founder).

Education and Advice We Rate

An overview of multi-task learning in deep neural networks

Sebastian Ruder
(PhD student in NLP and research scientist at AYLIEN) has written about the next frontier in machine learning, multi-task learning.

He writes that “generally, as soon as you find yourself optimising more than one loss function, you are effectively doing multi-task learning (in contrast to single-task learning). In those scenarios, it helps to think about what you are trying to do explicitly in terms of MTL and to draw insights from it.”

Podcasts We Love

Andrew Ng plans the next stage in his career

In this interview, Andrew Ng discusses World Class IT Principle One, People, and World Class IT Principle Five, External Partnerships.

Speaking about what he wants to do next, he said, “I am looking into quite a few ideas in parallel and exploring new AI businesses that I can build. One thing that excites me is finding ways to support the global AI community so that people everywhere can access the knowledge and tools that they need to make AI transformations.”

Other topics include:

  • Recommendations for CIOs and CEOs starting their AI journey
  • Advantages of implementing a centralized AI org structure
  • How lifelong learning can minimize jobs displaced through AI

Chat Bots 4Eva

Inside the bizarre human job of being a face for AI

Some people are famous only among fans of a particular sport, a specific age group, or their hometown locals. Lauren Hayes, a 27-year-old model and entrepreneur, is famous at the automation software company IPsoft for being the inspiration behind their chat bot, Amelia.

Hayes says she didn’t realise she had become the face of an AI until about a year after the first version of Amelia launched in 2014. She knew something was different about the modelling job when she showed up to the photo shoot and found what she calls a “death star,” a sphere-shaped structure that supports many cameras (it looks something like this). “At that moment, I was like, this is not like anything I’ve ever done before,” Hayes says.

Something to Get Involved In

Mental health AI hackathon

Acorn Aspirations is running a Mental Health Artificial Intelligence Hackathon to inspire teens to get involved in AI and is looking for mentors (tech, designers, marketeers, entrepreneurs) and sponsors who
would be interested in exposure during London Tech week or could “sponsor a child” as the majority of attendees are in receipt of Free School Meals and will attend for free.

Come hack with teens, learn to create chat bots, program Alexa, and help us nurture the next generation of A.I. entrepreneurs…with the help of Sophia the robot, who has just confirmed her attendance as Judge! For more info please contact Elena Sinel at elena.sinel@acornaspirations.co.uk.

+It is estimated that 1 in 6 people in the past week experienced a common mental health problem (Mental Health Foundation). Want to learn about how
companies are using AI to address these problems? Come to CogX on Tues June 20 to learn more.

Video Killed the Radio Star

Algorithmic beautification of selfies

The paper “Perspective-aware Manipulation of Portrait Photos”, which this great video is based on and its demo is available here.

The paper introduces a method to modify the apparent relative pose and distance between camera and subject given a single portrait photo. Their approach fits a full perspective camera and a parametric 3D head model to the portrait, and then builds a 2D warp in the image plane to approximate the effect of a desired change in 3D.

Dates for Your Diary

I’ve been making some changes based on Feedback. Would love to hear from more of you. Please do click to share your thoughts!


June 6, 2017

Issue 228: CognitionX Data Science, AI and Machine Learning Briefing

Storyzy + fake news. Goodyear jumps on self-driving bandwagon. Data is king at Airbnb. http://cognitionx.com/news-briefing/

In 2013, Carl Benedikt Frey and Michael A. Osborne published a report titled The Future of Employment: How susceptible are jobs to computerisation?. The authors examine how susceptible jobs are to computerisation, by implementing a novel methodology to estimate the probability of computerisation for 702 detailed occupations.

Recently, a very interesting site has been created which extracts the jobs and the probability of automation from the report and makes it easy to search for your job. They also added some additional information from the Bureau of Labor Statistics to provide some additional information about the jobs.

Rob McCargow, AI Programme Leader at PWC UK, reminded us last week during a BBC Five Live interview that it’s about the tasks being automated not the jobs, so maybe that will inspire an additional review of the engine.

The real question, however, is how do we retain, educate, and redeploy people to keep them in employment?  All of this and more will be debated at CogX and discussed at the breakout session. Looking forward to seeing you there.


Tabitha UntilTheBotsTakeOver Goldstaub

P.S. 15 days till CogX. Although the early bird pricing has ended, we have just released session passes which can be purchased for £199 per session. Check out the full schedule and agenda here. Be there or be intelligence artificial.

Future of News

Storyzy is a quote verifier that wants to skewer fake news

Storyzy wants to help out with the oft-discussed ‘fake news’ problem — with a just-launched quote verifier tool which, while it’s unlikely to be able to provide
copper-bottomed reassurance for more obscure quotations across the full gamut of online media, might at least be able to give a thumbs up or down on words (apparently) uttered by higher-profile individuals.

The tool, which uses natural language processing to harvest quotations from its trusted media sources, is the work of a French team. They actually started on their idea back in 2012 — under a different name and with the rather more ambitious aim of building an “automated fact checking add on/overlay on every website,” says CEO and co-founder Stan Motte.

+Want to learn more about how companies are using machine learning to combat fake news? Come to CogX to hear from David Benigson, CEO and founder of Signal Media, discuss how his company is fighting the fight.

Business Impact of AI

Morgan Stanley’s 16,000 human brokers get algorithmic makeover

Morgan Stanley is about to augment its 16,000 financial advisers with machine-learning algorithms that suggest trades, take over routine tasks and send reminders when your birthday is near.

The project, known internally as “next best action,” shows how one of the world’s biggest brokerages
aims to upgrade its workforce while a growing number of firms roll out fully automated platforms called robo-advisers. The thinking is that humans with algorithmic assistants will be a better solution for wealthy families than mere software allocating assets for the masses.

+Come to CogX to hear the panel on
The Impact of AI on Financial Services & The Economy, featuring

  •  Helia Ebrahimi (Business Correspondent at Channel 4)
  •  Michael Harte (Chief Innovation Officer at Barclays)
  • Doyne Farmer (Director of the Complexity Economics Program at the Institute for New Economic Thinking at the Oxford Martin School, Professor at Oxford)
  • and more…

Future of Transportation

Goodyear set to accelerate with self-driving cars

Jessica Dye, from FT (paywall), discusses how the Ohio-based company’s shares jumped 6.8 per cent – the most in 15 months – last week after Morgan Stanley analysts gave it a substantial upgrade and more than doubled its share-price target on expectations that driverless car and truck technology will pump up demand for tyres.

At first blush, it would seem that cars and trucks would require the same number of tyres to operate no matter who — if anyone — is behind the wheel. But
Adam Jonas
‘ (Morgan Stanley analyst) report pointed to several key upsides for tyre makers in the take-off of autonomous driving technology.

+Scared, excited, or nervous about the future of transportation? Come to CogX to hear from world experts like MC Srivas (Chief Data Architect, Uber),
Catherine Faiers (COO, Addison Lee), and Stan Boland (Co-founder and CEO, FIVEAI) discuss the future of transportation and smart cities.

Products We Love

Visualise data instantly with machine learning in Google Sheets

Sorting through rows and rows of data in a spreadsheet can be overwhelming. That’s why Google rolled out new features in Sheets that make it even easier for you to visualise and share your data, and find insights your teams can act on.

Explore in Sheets, powered by machine learning, helps teams gain insights from data, instantly. Simply ask questions—in words, not formulas—to quickly analyze your data. For example, you can ask “what is the distribution of products sold?” or “what are average sales on Sundays?” and Explore will help you find the answers.


Convolutional neural networks for medical image analysis: full training or fine tuning?

In this paper, the authors seek to answer the following central question in the context of medical image analysis: Can the use of pre-trained deep CNNs with sufficient fine-tuning eliminate the need for training a deep CNN from scratch? To address this question, they considered 4 distinct medical imaging applications in 3 specialties (radiology, cardiology, and gastroenterology).

Their experiments consistently demonstrated that (1) the use of a pre-trained CNN with adequate fine-tuning outperformed or, in the worst case, performed as well as a CNN trained from scratch; (2) fine-tuned CNNs were more robust to the size of training sets than CNNs trained from scratch; (3) neither shallow tuning nor deep tuning was the optimal choice for a particular application; and (4) their layer-wise fine-tuning scheme could offer a practical way to reach the best performance for the application at hand based on the amount of available data.

+Healthcare is changing rapidly due to advances in machine learning. You don’t want to miss Dr Jack Kreindler (CHHP), Maxine Mackintosh (One HealthTech), Dr Dominic King (DeepMind), and more at
CogX…be there to learn more and to move the conversation forward.

Video Killed the Radio Star

Ray Kurzweil predicts when we’ll be able to program matter

In this video, Kurzweil  predicts when he thinks we’ll get programmable matter—or the ability to manipulate everyday objects at the atomic level—and what that means not just for the things around you, but for you as a person.

According to Kurzweil, beginning with virtual reality in the digital realm and moving into the physical with programmable matter, people will make themselves into avatars of other people, or even each other.


Check out Microsoft’s Face Swap

Face-swapping — taking your image and adjusting it to fit on to a different body, creature or object — has taken the photo app world by storm, as millions of people look for new and novel things to do with their selfies. Microsoft is the latest company to enter the fray. The company’s in-house app studio Microsoft Garage has quietly launched a new app — the aptly called Face Swap.

Part image search engine, and part photo editing toolkit, you take a pic of yourself, adjust it a little, and then use keywords to find an image where you can replace the face with yours either from your camera roll, from the catalog in the app, from the wider world of images as indexed by Bing. Microsoft automatically adjusts your skin tone to the surroundings and textures, and you can make other edits with lighting and tilting and resizing your face, before sharing it on other social channels.

Dates for Your Diary

I’ve been making some changes based on Feedback. Would love to hear from more of you. Please do click to share your thoughts!


June 6, 2017

Issue 227: CognitionX Data Science, AI and Machine Learning Briefing

Making AI transparent. Google I/O 2017. Amazon Alexa. http://cognitionx.com/news-briefing/

Mary Meeker is at it again. She has just released her report on the future of the internet
(which gets longer and longer every year). Although the report is not AI-focused, there are a number of key takeways related to AI and machine learning:

  • Text-based search is so yesterday. As image and voice recognition tech gets better and better, we are seeing a shift toward image and voice search. Twenty percent of mobile queries were made via voice in 2016 and accuracy is now about 95 percent.
  • The volume of data we are producing is growing at a rapid rate. GPU
    processing is growing alongside it so we can efficiently process it.
  • She’s got a lot to say about the rise and adoption of cloud computing. She discusses the new players who are emerging (like Rubrik and Looker) and the serious ethical issues which must be considered.


Tabitha UntilTheBotsTakeOver Goldstaub

P.S. 18 days till CogX. Be there or be intelligence artificial. Early bird ticket prices end on Sunday, so grab them while they’re hot. On Monday June 5th, conference passes will change from £899 to £1,799.

Ethics Question for the Day

Top UK firms’ websites violate key GDPR principle

Just over one-third of all the public web pages of leading UK companies that collect personal information violate a key principle of new European data protection regulations.

The study revealed that 34% of web pages of FT30 firms that collect PII are doing so insecurely, 29% are not using encryption, 3.5% are using vulnerable encryptions algorithms, and 1.5% have expired security certificates.

While the insecure collection of PII is a violation of the GDPR, the study said the loss of personal data, profit and reputation resulting from the use of insecure forms is a legitimate concern for consumers and shareholders.

+Fascinated, confused, or concerned about GDPR? Then grab a ticket to hear Dr Sandra Wachter (Lawyer & Researcher in Data Ethics at Oxford Internet Institute at the University of Oxford), an expert on the topic.

Deal of the Day

May 2017 fundings, acquisitions, IPOs and failures

The Robot Report has put together a helpful list of noteworthy funding, M&As, IPOs, and failures related to AI and robotics.

May 2017 had two robotics-related companies get $9.5 billion in funding (Didi Chuxing (Uber of China) and Nvidia) and 22 others raised $249 million. Acquisitions also continued to be substantial with Toyota Motor’s $260 million acquisition of Bastian Solutions plus three others (where the amounts weren’t disclosed).

+Can’t keep track of all the investments in the AI space? At CogX, we will have a stellar panel on investing in AI, featuring Richard Muirhead (Co-Founder & Chairman of ​Firestartr, General Partner at Open Ocean) and many others, not to mention the many VCs and angel investors who will be attending the conference.

Podcasts We Love

Making AI transparent

In this podcast episode, Jon Bruner speaks with Andy Hickl, chief product officer at Intel Saffron Cognitive Solutions Group. Their subject: common sources of bias in artificial intelligence—and how to identify and address them by using transparent AI methods.

For more on the problem of interpretability in
AI and techniques for building transparent applications, register for the free webcast “Transparency in AI Decision-making,” hosted by Andy Hickl.

Video Killed the Radio Star

Past, present and future of AI / machine learning (Google I/O ’17)

You don’t want to miss this panel with Diane Greene, Fei-Fei Li, Daphne Koller, Françoise Beaufays, and Fernanda Viégas. Really inspiring to see this all female power-house panel.

+Want to hear more about women in AI or the lack thereof? Come to CogX to hear our ‘Women in AI’ panel.


Amazon rejects AI2’s Alexa skill voice-search
engine. Will it build one?

Amazon released its own “Skill Finder” skill last year, but it’s a bare-bones experience that can only read off the most popular apps in certain vague categories, or list the top or newest Alexa skills. You can’t ask it for a skill with a specific use case or functionality.

So the Allen Institute For Artificial Intelligence figured it’d build a full-fledged voice keyword search engine for Alexa skills. Funded by Microsoft co-founder Paul Allen, AI2 is one of the largest not-for-profit AI institutes in the world. But when AI2 submitted the Skill Search engine to the Alexa platform, Amazon rejected it, citing that “We don’t allow skills that recommend skills to customers at this time. We will contact you if this feature becomes available.”

+Wanna learn more about Alexa? The founder and Chief Product Officer of Evi (tech behind Alexa), William Tunstall-Pedoe will be speaking at CogX about the ‘Customer interface of the future’.

Products We Love

New smart wearable protects workers on the job

A building site can have its fair share of dangers for workers, who can be at risk daily of falls and other injuries. In 2015, almost 5,000 workers experienced a fatal incident on the jobsite, 38.8% of which were caused by falls according to the Bureau of Labor Statistics.

In an effort to make jobsites safer, Triax Technologies, Inc., a provider of wearable technology for construction site connectivity, has launched its new spot-r technology for wearables. The IoT-based connected lightweight sensor clips on to a worker’s belt, and in the event of an incident will alert safety personnel to slips, trips and falls in real time so workers can receive aid faster. Workers can also send safety alerts in a time of crisis. In the case of an emergency, authorized personnel can trigger site-wide evacuation alerts emitted by the sensor.

Education and Advice We Rate

Apple Swift Playgrounds can now program your robots and drones

On Thursday, Apple announced that, starting June 5, its increasingly popular iPad-based programming app, Swift Playgrounds, can be used to program robots built with LEGO Mindstorms EV3. Version 1.5 of the app will also handle programming tasks for Sphero’s SPRK robot ball, Parrot’s Mambo, Airborne and Rolling Spider drones, as well as robots from UBTECH, Dash, and Skoog. All devices will connect to the iPad and Swift Playgrounds via a Bluetooth connection.

Dates for Your Diary

I’ve been making some changes based on Feedback. Would love to hear from more of you. Please do click to share your thoughts!


Thank you! Your subscription has been confirmed. You'll hear from us soon.
Subscribe to the News Briefing

Log in with your credentials


Forgot your details?

Create Account