Subscribe to the News Briefing Now

Stay up to speed with the latest news, developments and industry trends in the world of all things AI with our curated daily or weekly briefing.

March 24, 2017

Issue 181: CognitionX Data Science, AI and Machine Learning

Today, the Forbes Technology Council published ‘9 Ways Your Business Can Plan For Artificial Intelligence’. They discuss the need for companies to do their due diligence in deploying AI and the fact that AI is a must, and not an option, for startups. Also, we couldn’t agree more that “not everyone needs a plan, but everyone needs to research.”

How is your company planning for AI?

AI deployed      AI in PoC     Figuring things out 
Not started


Tabitha UntiltheBotsTakeOver Goldstaub

Business Impact of AI

Intel creates the Artificial Intelligence Products Group

Intel has just announced their newly minted single cross-Intel organization: the Artificial Intelligence Products Group, which will be led by Naveen Rao. The new organisation will align resources from across the company to include engineering, labs, software and more as they build on their current AI portfolio.

In addition, they will be creating an applied AI research lab dedicated to pushing the forefronts of computing. They will be exploring novel architectural and algorithmic approaches to inform future generations of AI. This includes a range of solutions from the data centre to edge devices, and from training to inference – all designed to enable Intel and its customers to innovate faster.

Exciting Opportunities

Octopus Ventures raises £120m to invest in UK AI startups

Prominent VC Octopus Ventures has raised a £120m tech investment fund to continue supporting AI startups in the UK. Octopus Ventures has so far invested in 50 companies across many different verticals.

Speaking to The Financial TimesAlex Macpherson, CEO at Octopus Ventures, said: “We have expertise
in the machine-learning field, but the challenge today is pretty much every business that comes through to us is machine learning or artificial intelligence.” Octopus is currently looking to participate in the Seed rounds of two AI startups in the UK over the coming months.

Open Source

Open sourcing TensorFlowOnSpark: distributed deep learning on big-data clusters

Last month, Yahoo released TensorFlowOnSpark to the community, their latest open source framework for distributed deep learning on big-data clusters.

Bottom line: using TFoS you can run TensorFlow free of Google’s cloud and have it share servers already running other big-data apps and processes rather than dedicated clusters.

Economic Impact of AI

Canada funds $125 million Pan-Canadian Artificial Intelligence Strategy

The Government of Canada announced on Wednesday that it is funding a Pan-Canadian Artificial Intelligence Strategy for research and talent that will cement Canada’s position as a world leader in AI.

The $125 million strategy will attract and retain top academic talent in Canada, increase the number of post-graduate trainees and researchers studying artificial intelligence, and promote collaboration between Canada’s main centres of expertise
in Montreal, Toronto-Waterloo and Edmonton. The programme will be administered through CIFAR, the Canadian Institute for Advanced Research.


$1 today or $2 tomorrow? The answer is in your Facebook likes

Delay discounting, a behavioural measure of impulsivity, is often used to quantify the human tendency to choose a smaller, sooner reward (e.g., $1 today) over a larger, later reward ($2 tomorrow). Delay discounting and its relation to human decision making is a hot topic in economics and behaviour science since pitting the demands of long-term goals against short term desires is among the most difficult tasks in human decision making [Hirsh et al., 2008].

Previously, small-scale studies based on
questionnaires were used to analyse an individual’s delay discounting rate (DDR) and his/her real-world behaviour (e.g., substance abuse) [Kirby et al., 1999]. In this research, they employ large-scale social media analytics to study DDR and its relation to people’s social media behaviour (e.g., Facebook Likes). They also build computational models to automatically infer DDR from Social Media Likes.

Education and Advice We Rate

Harry Potter Sorting Hat and classification algorithms

When you first start learning about data science, one of the first things you learn about are classification algorithms. The concept behind these algorithms is pretty simple: take some information about a data point and place the data point in the correct group or class.

Bryan Berend, Lead Data Scientist at Nielsen, teaches about this notion with a more fun example than spam filters and takes a stab at using classification algorithms on Harry Potter, trying to build a classifier to sort characters into the different houses. Although the classifier is not particularly successful, you will learn a lot about APIs and more in the process.


Floyd is a startup betting the world needs Heroku for deep learning

A startup called Floyd has developed a cloud service for deep learning. Yesterday Floyd’s founders talked about their product onstage at Silicon Valley accelerator Y Combinator’s demo day.

Floyd is seeking to build out a marketplace rich with data sets and algorithms. But at its core, it’s a
managed service for training neural networks and then running machine-learned models on an ongoing basis. In that sense, it competes with existing machine learning services from public clouds like Microsoft Azure, Google Cloud Platform, and of course the market-leading Amazon Web Services (AWS), on which Floyd is itself hosted. Check out this article and ProductHunt to see what makes Heroku unique.

Future of Art

Deep photo style transfer

This paper introduces a deep-learning approach to photographic style transfer that handles a large variety of image content while faithfully transferring the reference style. Their approach builds upon recent work on painterly transfer that separates style from the content of an image by considering different layers of a neural network.

Even when both the input and reference images are photographs, the output still exhibits distortions reminiscent of a painting. Their contribution is to constrain the transformation from the input to the output to be locally affine in colorspace, and to express this constraint as a custom CNN layer through which they can backpropagate.

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!

March 23, 2017

Issue 180: CognitionX Data Science, AI and Machine Learning

Robots can be programmed to perform all sorts of tasks. They can do legal researchdrive themselves, and sell you insurance. But can you program a robot to compute ethics? Mike Loukides from O’Reilly Media is less optimistic about this possibility.

What do you think: is it possible to imagine an AI that can compute ethics? 

Yes           No
It’s Complicated


Tabitha UntiltheBotsTakeOver Goldstaub


Thanks to all of you who answered our survey about ‘the future of work’. The audience was divided, with 1/3 being optimistic, 1/3 being pessimistic, and 1/3 saying that it was complicated.

Ethics Question for the Day

On computational ethics

Mike Loukides, Vice President of Content Strategy for O’Reilly Media, asks ‘is it possible to imagine an AI that can compute ethics?’ He says that if ‘living a good life’ isn’t a difficult optimisation problem, he doesn’t know what is. He doesn’t see how ethics based on a priori considerations, like Aristotle’s ideas of duty or virtue, the Ten Commandments, or Asimov’s laws,
could be computed.

These are all external inputs to the system, whether handed down on stone tablets or learned and handed down through many generations of human experience. He doubts that an AI could derive the idea that it must not harm a human—if we want AIs to behave according to anyone’s laws, they’ll have to be built into the system, including systems that have the ability to write their own code.

Future of Health

Machine learning lets scientists reverse-engineer cellular control networks

Researchers have used machine learning on the Stampede supercomputer to model the cellular control network that determines how tadpoles develop. Using that model, they reverse-engineered a drug intervention that created tadpoles with a form of mixed pigmentation never before seen in nature.

The utility of these methods is their ability to find novel regulatory interactions and even novel necessary regulatory genes. These methods are indeed becoming indispensable for understanding the complex coordination of signals necessary to develop and maintain correct body shapes and organs. Moreover, such methods are required in order to develop interventions to make rational changes to complex anatomy and physiology, in the context of regenerative medicine and systems-level diseases such as cancer.

Future of News

Facebook rolls out new alert to combat fake news

Facebook has started rolling out its third-party fact-checking tool in the fight against fake news, alerting users to “disputed content”. The site announced in December it would be partnering with independent fact-checkers to crack down on the spread of misinformation on its platform.

The tool was first observed by Facebook users attempting to link to a story that falsely claimed hundreds of thousands of Irish people were brought to the US as slaves.

Deal of the Day

AI-based legal research firm Casetext closes on $12M in funding

In one of the largest investments in the legal tech industry to date, Casetext, which provides AI-based legal research technology for lawyers, has closed on a $12 million Series B funding round.

Casetext’s CARA uses AI and natural-language technologies to automate key legal research tasks, arming lawyers with the highest quality research possible, ultimately allowing firms to spend time on higher-value, billable work—and not miss key precedents or decisions. Because CARA is powered by the Casetext research database, users have access to a full library of federal and state law, annotated by expert analysis from leading attorneys and law firms.

Business Impact of AI

AI, machine learning blossom in agriculture and pest control

Seed retailers are using AI products to churn through terabytes of precision agricultural data to create the best corn crops, while pest control companies are using AI-based image-recognition technology to identify and treat various types of bugs and vermin. Such markedly different scenarios underscore how AI has evolved from science fiction to practical solutions that can potentially help companies get a leg up on their competition.

Beck’s Hybrids, for example, is using an AI product to analyse large amounts of data to determine which corn breeds and which conditions will produce the highest yields. The company’s geneticists need to know how sun light, rain, location, terrain and could affect growth and profits for the more than 30,000 different types of seeds it offers.

Future of Sports

Football data-driven ghosting using deep learning

Current state-of-the-art sports statistics compare players and teams to league average performance. For example, metrics such as “Wins-above-Replacement” (WAR) in baseball, “Expected Point Value” (EPV) in basketball and “Expected Goal Value” (EGV) in football and hockey are now commonplace in performance analysis. Such measures allow us to answer the question “how does this player or team compare to the league average?” Even “personalised metrics” which can answer how a “player’s or
team’s current performance compares to its expected performance” have been used to better analyse and improve prediction of future outcomes.

Motivated by the original “ghosting” work, Disney researchers showcase an automatic “data-driven ghosting” method using advanced machine learning methodologies applied to a season’s worth of tracking data from a recent professional league in football.

Future of Transportation

21 industries other than auto that driverless cars could turn upside down

CBInsights recently put out a blogpost which discusses the many effects which driver-less cars will have, beyond the automotive industry itself. The author argues that the impact will be far-reaching and lists the industries which will be effected, including: insurance, hotels, airlines, and real estate.

Regarding real estate, for example, they said: “Bloomberg’s Noah Smith says faster and easier commutes will shift residential property value from properties in urban centers to those in surburban areas. In commercial real estate, spaces currently predicated on human drivers will be converted to other uses.”

Chat Bots, yadda yadda yadda

Next Insurance makes it easy for small businesses to buy insurance via a chat bot

Palo Alto-based startup Next Insurance has launched an insurance chat bot to enable small businesses to quote and buy insurance via Facebook Messenger. It partnered with enterprise-focused chat bot developer SmallTalk to provide tailored insurance policies for small businesses via a social channel.

Guy Goldstein, co-founder and CEO of Next, said: “70 percent of our customers are buying insurance on their phones. Enabling customers to buy insurance through a chatbot on Facebook Messenger brings simplicity, transparency, and easy access. We’re making sure that insurance is working for the small business owner and not the reverse.”

Education and Advice We Rate

Appreciating art with algorithms

Dean L., on HackerNoon, being inspired by a mosaic at the MoMA in San Francisco, decided to take a crack at making his own art using algorithms. He walks the reader through his attempts to do so, first by using deep neural networks (which took ages!), then by concocting some home-grown algorithms. He produced some really cool art and walks you through the steps so that you can too.

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!

March 22, 2017

Issue 179: CognitionX Data Science, AI and Machine Learning

The story is typically the same: “as automation increases, the amount of available jobs will decrease”. Ryan Avent, senior editor and economics columnist at The Economist, has recently highlighted another fear: ‘given the structure of our social safety net, automation tends to increase poverty and inequality rather than unemployment.’

Michael Karnjanaprakorn (CEO of Skillshare), on the other hand, is optimistic, providing ‘7 reasons to be excited about the future of work’. One thing he is excited about is the fact that with the rise of AI and automation comes the rise of creative (as opposed to routine) work.


Tabitha UntiltheBotsTakeOver Goldstaub

Another Example of “Robots Taking Humans’ Jobs”

The productivity paradox

Ryan Avent, senior editor and economics columnist at The Economist, recently wrote an article on Medium in which he tackles the perennial ‘robots are taking our jobs’ problem. He argues that the robot threat is totally overblown: the fantasy, perhaps, of a bubble-mad Silicon Valley — or an effort to distract from workers’ real problems, trade and excessive corporate power. Generally
speaking, the problem is not that we’ve got too much amazing new technology but too little.

In short, he says that continued high levels of employment with weak growth in wages and productivity is not evidence of disappointing technological progress; it is what you’d expect to see if technological progress were occurring rapidly in a world where thin safety nets mean that dropping out of the labour force leads to a life of poverty.

Deal of the Day

Andrew Ng, head of AI, is leaving Baidu

The chief scientist helping drive Baidu Inc.’s push into AI is quitting the Chinese search giant, putting at risk its efforts to put AI at the centre of a business revival.

Andrew Ng, who joined Baidu in 2014, said he’s leaving the business next month. Ng doesn’t plan to join another technology company and will seek to bring AI into sectors such as health care and education around the world.

Products We Love

Pottery Barn introduces augmented reality app

Pottery Barn worked with Google to create 3D Room View. The catch is that the app will work only on phones that run Google’s Project Tango technology.


ARM unveils Dynamiq multicore chip designs for faster AI and cloud computing

Chip design firm ARM has unveiled its Dynamiq technology to make better multicore processors that can handle artificial intelligence, cloud computing, and new kinds of devices. Cambridge, England-based ARM said in a press call that the new designs (ARM designs chips and its partners incorporate them into their own manufactured chips) will be available in ARM Cortex-A processors coming to market later this year in automotive, networking, server, and “primary compute devices.”

The new processors will enable flexible multicore processing — in which a computing device has to juggle many different tasks of varying sizes at once. It will also emphasise “heterogeneous compute,” or using different kinds of cores or processors in the same machine, said Nandan Nayampally, general manager of the ARM Compute Products Group, in a press briefing.

Future of Transportation

DeepScale raises $3 million for perception AI to make self-driving cars safe

DeepScale, a startup out of Mountain View, has raised $3 million in seed funding to help automakers use industry-standard low-wattage processors to power more accurate perception. Alongside sensors, mapping, planning and control systems, perception, (sometimes referenced as “computer vision”) enables vehicles to make sense of what’s going on around them in real time.

DeepScale is competing for a share of this burgeoning market versus some 800-lb. gorillas in automotive tech, like Mobileye, now owned by Intel, or Bosch, but also other funded startups like, Argo and, which are trying another approach of building their own, fully autonomous vehicles or retrofit systems.

Chat Bots, yadda yadda yadda

Indian YES BANK partners with Payjo to make a banking chat bot

YES BANK is launching its wallet services through a chat-based financial assistant in partnership with Silicon Valley-based Payjo on Facebook Messenger.

“The experience offered through Payjo will make banking services for customers more accessible and fun, leading to more financial literacy in India,” said Payjo Founder and Chief Executive Officer, Srinivas Njay.

Education and Advice We Rate

Index of best AI/machine learning resources

Arun Agrahri, founder and CEO of Nucleus Labs, has put together a very useful collection of resources related to AI and machine learning.

His list includes articles, courses, videos, and newsletter which you can check out to educate yourself. If you have any resources which you feel like he should add, comment on his post or tweet at him.

Ethics Question for the Day

Google responds to outrage regarding inappropriately placed Youtube ads

There has been a lot of talk recently (and outrage) regarding Youtube ads being played alongside inappropriate material. Due to this concern, the UK government, among others, has pulled their ads from Youtube.

Google responded yesterday with a
blogpost. They said that they will be hiring significant numbers of people and developing new tools powered by our latest advancements in AI and machine learning to increase their capacity to review questionable content for advertising. In cases where advertisers find their ads were served where they shouldn’t have been, they plan to offer a new escalation path to make it easier for them to raise issues. In addition, they will soon be able to resolve these cases in less than a few hours.

Business Impact of AI

Introducing Foursquare Analytics: a dynamic foot traffic dashboard for brands

Foursquare is back. On Monday, they unveiled Foursquare Analytics, a clean, simple dashboard that puts the power of our proven location intelligence in the hands of brands. Foursquare Analytics allows brick-and-mortar retailers and restaurant chains to understand how their own company and an entire category are performing based on actual, measurable, real-world visits — and much more quickly than any option out there.

It’s a dashboard for insights on chain-level foot-traffic performance that can be easily compared to a competitive set and to the broader industry. It provides unprecedented metrics that measure loyalty, reveal demographic insights, and uncover sources of acquisition and loss. It allows analysts to have a precise understanding of changing store visit patterns and share of visits from a competitive set.

Future of Health

Introducing ZocDoc’s new machine learning search engine

Zocdoc’s new Patient-Powered Search provides an intuitive search experience, built specifically to bridge the gap between healthcare industry and human speak. With Patient-Powered Search, each person can use his or her own language – from “gyno” to “hurt wrist” to “post-election stress disorder”– to confidently
find the right provider for their needs.

In building Patient-Powered Search, their goal was to use data-driven technology to power a search engine that will evolve over time. The team started by building a machine learning algorithm using existing medical literature gathered online. This algorithm is a natural language processing model, which means it processes and understands the ways humans communicate, and then maps those colloquial terms to the appropriate specialty, visit reason or procedure type.

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!

March 22, 2017

Issue 178: CognitionX Data Science, AI and Machine Learning

Many leaders are advocating the very best way for AI to flourish is for people to invest in startups. So it was good to see Y Combinator announce they are creating a vertical group dedicated to AI. See details below to apply and watch this space as no doubt others are preparing to launch their AI focused funds also.

With this said, on Monday 27th you can ask our panel of investors, questions about the good bad and the ugly of Investing in AI. Send us questions now and we’ll be sure to field them on the night.

If you haven’t got tickets yet please do so asap as I believe we are almost at capacity.


Tabitha UntiltheBotsTakeOver Goldstaub

Future of Health

Cancer treatments boosted by AI genomic sequencing

“The technique that we use for this is genomic sequencing,” explained Dr. Jurgi Camblong, cofounder of Sophia Genetics, a provider of AI that pinpoints the genomic code mutations behind cancers and rare disorders to assist physicians and healthcare institutions in prescribing optimal drug treatments for their patients. Today, 240 hospitals in 39 countries use the Sophia platform.

“What the technology does is spot variations in different genetic codes so we can use historical data that aids in prescribing the best combination of drugs to treat a particular cancer or condition in an individual patient,” said Camblong. “This is next-generation genomic sequencing, and it is used in two different areas: chronic hereditary disorders and oncology. By using the algorithms that are part of our artificial intelligence, we can spot the origin of a genetic mutation causing a cancer or a particular condition and then give an idea of what the best drug treatment would be to the attending physician.”

Exciting Opportunities

Y Combinator “our goal is to democratize AI.”

Apply here “We can’t afford to ignore what might be the biggest technological leap since the Internet.”

They want to level the playing field for startups to ensure that innovation doesn’t get locked up in large companies like Google or Facebook.

I like to see they are also looking to fund companies focused on job re-training. It’s clear that this is because they believe the increased efficiency from AI will net out positive for the world, and so they are mindful of fears of job loss.

Business Impact of AI

AI and the legal sector

“There is this popular view that if you can automate one piece of the work, the rest of the job is toast,” said Frank Levy, a labor economist at the Massachusetts Institute of Technology. “That’s just not true, or only rarely the case.”

“Where the technology is going to be in three to five years is the really interesting question,” said Ben Allgrove, a partner at Baker McKenzie, a firm with 4,600 lawyers. “And the honest answer is we don’t know.”

Ethics Question for the Day

Is Facebook a structual threat to free society?

As Facebook grows, so will its ownership of the social graph and our digital selves. Systemic risk is highest in centralised systems. Extrapolating trends, the author on Truthhawk considers it possible, if not probable, that Facebook will become a systemic risk centre for free society. The argument goes:

  • Facebook engages in comprehensive and growing data collection on its billions of
  • This data allows for exponentially greater manipulation of human beings and their realities than ever seen before
  • Facebook is building towards human simulation and ownership
  • Extrapolating trends, Facebook will create unprecedented centralization of power and influence in the hands of an individual
  • Without a change of course, we are enabling a structural threat to free society, and potentially worse

Education and Advice We Rate

Creating your own data science MA

David Venturi decided that it didn’t make sense to spend upwards of $30K for another university degree. Therefore, he put together his own course from online resources, including Udacity and Coursera. He provides a thorough discussion of the courses he chose and why he chose them. The range of courses is quite broad and include instruction on both R and Python, standards data science tools.


Distill, an interactive and visual machine learning journal

The journal Distill launches today. In a nutshell, Distill is an interactive, visual journal for machine learning research.

It is a unique journal with articles that will integrate explanation, code, data, and interactive visualisations into a single environment. In such an environment, users can explore in ways impossible with traditional static media. They can change models, try out different hypotheses, and immediately see what happens. That will let them rapidly build their understanding in ways impossible in traditional static media.

Products We Love

Chinese park installs facial recogniton sotware to stop toilet paper thieves

Apparently toilet paper theft has gotten so bad at the Temple of Heaven Park in Beijing, authorities have installed facial recognition software in the stalls.

In order to get a few sheets, visitors now need to make eye contact with a computer before the dispenser spits out a serving of TP. The facial recognition program keeps the dispenser from offering another round to the same person; if you need more, you reportedly need to sit tight — literally — for nine whole minutes.


Amazon struggles in the grocery business

 “Online grocery is failing,” said Kurt Jetta, chief executive officer of TABS Analytics, a consumer products research firm. Only 4.5 percent of shoppers made frequent online grocery purchases in 2016, up just slightly from 4.2 percent four years earlier despite big investments from companies such as Amazon, according to the firm’s annual surveys. “There’s just not a lot of demand there. The whole premise is that you’re saving people a trip to the store, but people actually like going to the store to buy groceries.”

Amazon CEO Jeff Bezos now seems to understand that he can’t win the grocery game with websites, warehouses and trucks alone. The world’s biggest online retailer sees brick-and-mortar stores playing a key role in a renewed grocery push, documents reviewed by Bloomberg show. And like it did with Amazon Fresh, the company is launching its newest projects in Seattle, its home town. One of these initiatives is Amazon Go, which uses facial recognition and machine learning
to facilitate a checkout-less and cashier-less grocery store.

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

March 22, 2017

Issue 177: CognitionX Data Science, AI and Machine Learning

Rob McCargow, Dad (and Programme Lead – AI at PwC) UK spent 10mins discussing Ethics and AI with his 5 & 7 year old kids and they came up with 4 rules for Robots:

  • Bad people shouldn’t build Robots
  • There has to be an off switch
  • There shouldn’t be bombs in Robots
  • Robots
    shouldn’t look like humans

What do you think the most important rules for “robots” are? Add your rule by clicking here


Tabitha UntilTheGoodRobotsTakeover

Dates for Your Diary

Impact on Everything

Is an AI Arms Race Inevitable?

As the race to create increasingly powerful AI accelerates, and as governments increasingly test capabilities in weapons, many AI experts worry that an equally terrifying AI arms race may already be under way.

In fact, at the end of 2015, the Pentagon requested $12-$15 billion for AI and autonomous weaponry for the 2017 budget, and the Deputy Defense Secretary at the time, Robert Work, admitted that he wanted “our competitors to wonder what’s behind the black curtain.”

This article is part of a weekly series on the 23 Asilomar AI Principles. The Principles offer a framework to help artificial intelligence benefit as many people as possible. 

Impact on the Economy

Computer says NO Did Artificial Intelligence Deny You Credit?

Traditionally, a loan officer who rejected an application could tell a would-be borrower there was a problem but computerized systems that use complex machine learning models are difficult to explain, even for experts.

In this Forbes article Anupam Datta Associate Professor of Computer
Science and Electrical and Computer Engineering at Carnegie Mellon University explains the risks and rewards of this method.

Ethical Question of the Day

Researchers are using Darwin’s theories to evolve AI, so only the strongest algorithms survive

Google’s hybrid approach combines classic neuroevolution with the techniques, like backpropagation, that have made deep learning so powerful today: Teach an algorithm how to act in the world, let it evolve, and that algorithm’s child will have most of the accrued knowledge. OpenAI’s approach was more true to how evolution works in biology. The team only let the randomised mutations in every generation govern how the networks improved or failed, meaning improvement was only created through random evolution. But both attempts had very clear goals—recognize an image, or get a high score in a game (or make a horse run faster). How the algorithms got there was up to nature.

Education and Advice We Rate

Cutting through the hype (when you’re not an ML researcher) 

Here’s the latest installment of Rachel Thomas’s Ask-A-Data-Scientist advice column. She makes clear that although Machine learning hasn’t been fully commoditized yet, that doesn’t mean you need a PhD.

Future of Music

They steal your whole sound, that’s a Real World Challenge Kayne West or Bot?

A West Virginia teenager has created a rapping A.I. bot that self-generates bars using Kanye West lyrics. Robbie Barrat, 17, uses open-source code and programmed the A.I. with 6,000 Kanye lines and finished the project in a week to show his peers at the next club meeting. “Originally it just rearranged existing rap lyrics, but now it can actually write word-by-word,” Barrat said. The bot can also incorporate pauses for rhythm and effect. Barrat is now working on more neural networks that can potentially write melodies and produce “abstract art.”

Advice We Rate

Nathan Benaich (Early Stage Investor and London.AI Organiser ) gave a great talk at Oxford AI Society about starting an AI business. You can check his slides here.

Impact on Art

Art for art’s sake money for god’s sake

Arthena, which is part of the current batch of startups at Y Combinator, says it can help investors double the art market’s standard annual return of 10 percent reliably from art. Founder and CEO Madelaine D’Angelo said Arthena first launched as an equity crowdfunding platform for purchasing art. More recently, it’s added financial tools to create “accommodate that quantitative strategy for the art market.”

Deal of the day

Amazon applies it’s AI tools to Cyber Security.

Amazon’s acquisition of startup is a natural fit for the company. uses AI-based algorithms to identify the most important documents and intellectual property of a business, then combines user behavior analytics with data loss prevention techniques to protect them from cyber attacks.

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

March 17, 2017

Issue 176: CognitionX Data Science, AI and Machine Learning

Charlie and I are honoured to be attending the All Party Parliamentary Group on Artificial Intelligence (APPG AI) on 20th March.  Since the session may influence decision-making in government, we’d appreciate your thoughts on ‘what sector the government should focus its AI-related spending on‘. Answer the survey below and please leave a comment as well.

Healthcare          Cybersecurity          Automobiles

We look forward to seeing some of you there and reporting back on the event.


Tabitha UntiltheBotsTakeOver Goldstaub

Dates for Your Diary

Ethics Question for the Day

Pros and cons of the Algorithm Age

As with everything else there are pros and cons when it comes to algorithms. Are the glass half-full or half empty? Pew Research Center and Elon University’s Imagining the Internet Center conducted a large-scale canvassing of technology experts, scholars, corporate practitioners and government leaders. They asked:  Will the net overall effect of algorithms be positive for individuals and society or negative for individuals and society in the next decade? Some 1,302 responded.

 Some believe they will mainly be of benefit to humans and society, others  worry that it will be the opposite. The non-scientific canvassing found that 38% of these particular respondents predicted that the positive impacts of algorithms will outweigh negatives for individuals and society in general, while 37% said negatives will outweigh positives; 25% said the overall impact of algorithms will be about 50-50, positive-negative.

Video Killed the Radio Star

Probabilistic machine learning: foundations and frontiers (Strachey Lecture, Oxford)

Professor Zoubin Ghahramani, professor of machine learning and recently appointed chief scientist at Uber AI Lab, delivered the Strachey lecture about probabilistic machine learning and its applications, as well as implications.

Probabilistic modelling provides a mathematical framework for understanding what learning is, and has therefore emerged as one of the principal approaches for designing computer algorithms that learn from data acquired through experience. Professor Ghahramani reviews the foundations of this field, from basics to Bayesian nonparametric models and scalable inference. He will then highlight some current areas of research at the frontiers of machine learning, leading up to topics such as probabilistic programming, Bayesian optimisation, the rational allocation of computational resources, and the Automatic Statistician.

Stats that Impress

Two out of three consumers don’t realize they’re using AI

Inbound marketing company HubSpot has been looking at consumer sentiment towards AI and chatbots. It has released its Global AI survey for Q4 2016.

In its survey of more than 1,400 people from Ireland, Germany, Mexico, Colombia, UK, and the US, HubSpot found that many respondents did not know that they were already using

37 percent of respondents said they have used an AI tool. However, of the respondents who said they have not used AI, 63 percent were actually using it. They just were not aware that they were.


Applying data and machine learning to scale education

Daphne Koller, Chief Computing Officer and Co-founder of Coursera, gave a lecture about the founding of Coursera and how they are using data science to teach data science. When she (along with some other students at Stanford) saw that the courses that they put online were attracting over 100K viewers, they realized that there was a serious market for online learning.

She discusses how Coursera understands their learners (by recommending courses based on country, economic standing, etc.). She also explains the differences between building an education website versus other types of sites and how Coursera uses a two-level hierarchy and decision trees to cluster the courses.

Future of Health

AliveCor CEO sees role for machine learning in medicine

Vic Gundotra, AliveCor’s chief executive officer, discusses the company’s medical technology and the role of machine learning in medicine with Bloomberg’s Caroline Hyde on “Bloomberg Technology.”

He discusses the company’s new product, aimed at physicians which allows the doctor to monitor all of their patients who are using AliveCor technology. They can monitor weight, view EKGs, and more. He believes that doctors will be skeptical until they see “that the data that they are seeing is

Future of Transportation

BMW says self-driving car to be level 5 capable by 2021

German carmaker BMW is on track to deliver a self-driving car by 2021, the company’s senior vice president for Autonomous Driving, Elmar Frickenstein, said on Thursday. “We are on the way to deliver a car in 2021 with level 3, 4 and 5,” Frickenstein told a panel discussion in Berlin, explaining the vehicle will have different levels of autonomy, depending on how and where it is used.

A level 5 vehicle is capable of navigating roads without any driver input, while a level 3 car still needs a steering wheel and a driver who can take over if the car encounters a problem.

Future of News

The Washington Post preps its AR push

The Washington Post plans to use AR to enhance its reporting and storytelling in 2017.

The Post first used AR last year to explain the events that led up to Freddie Gray’s arrest death in Baltimore in 2015. But people had to download an app to access the experience. Since then, the Post has been building an AR framework into its two existing apps — its magazine-style “Rainbow” app and its more traditional, newspaper-style app — to take friction out of the process.

It plans to launch its first AR story this spring, and then one per quarter. The first will be for a series by its art and architecture critic Philip Kennicott that will examine new, innovative buildings. AR will be used to let people look around the interiors and listen to narration using their smartphones.


AI agents are learning to communicate (OpenAI)

The authors’ hypothesis is that true language understanding will come from agents that learn words in combination with how they affect the world, rather than spotting patterns in a huge corpus of text. As a first step, they wanted to see if cooperative agents could develop a simple language amongst themselves.

Their approach yields agents that invent a (simple!) language which is grounded and compositional.

They hope that this research into growing a language will let us develop machines that have their own language tied to their own lived experience. They think that if we slowly increase the complexity of their environment, and the range of actions the agents themselves are allowed to take, it’s possible they’ll create an expressive language which contains concepts beyond the basic verbs and nouns that evolved here.

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

March 16, 2017

Issue 175: CognitionX Data Science, AI and Machine Learning

Amazon Alexa developers are jumping for joy. Yesterday, Amazon announced that they would be providing AWS credit to developers who have published Alexa skills. This removes significant cost roadblocks which developers face when building new Alexa skills. With the recently announced Alexa Fund Fellowship, they are also trying to tackle another issue – the lack of proper training.

What do you think the most significant roadblocks and challenges are facing developers working in the home assistant space? Training? Cost? Competition? Customer Adoption? Something else? Drop us a comment please.


Tabitha UntiltheBotsTakeOver Goldstaub

Dates for Your Diary

Education and Advice We Rate

Learning AI if you suck at math (part 6)

If you’ve followed parts 1234 and 5 of this series you know that you really don’t need a lot of math to get started with AI. You can dive right in with practical tutorials and books on the subject.

In this part, Daniel Jeffries walks you through set theory, algorithms, and matrices. He recommends you pick up Mathematical Notation: A Guide for Engineers and Scientists by Edward R. Scheinerman to get familiarised with the notation.


UK FinTech market matures as focus turns to AI and putting customers first (Startupbootcamp/PwC report)

FinTech startups are increasingly focusing on building smarter, faster machines as they seek to gain a better understanding of artificial intelligence and its potential to solve customer problems, a new report by the Startupbootcamp FinTech London programme and PwC can reveal.

Working with hundreds of startups and financial services companies, the report’s authors have seen a significant change in culture over the past year. Almost 1 in 7 (16%) of applications to the incubator programme in 2016 looked to build new prototypes and many were focused on AI and machine learning.

Future of Health

Pharma is using deep learning to cure aging

One company that has been using deep learning to take on aging is InSilico Medicine. The company uses deep neural networks (DNNs) to sort through huge amounts of biological data. The DNNs look for biomarkers (measurable indicators of your biological state such as those included in blood tests) that correlate with aging. For humans, this would be an impossibly complicated and time-consuming task.

Last week, InSilico Medicine announced a product called Ageless Cell. The supplement contains four natural compounds that DNNs have shown can rejuvenate older cells. LEF has access to blood tests from its customers who take the product. That means data should be available in less than a year. If it works, we can expect
other DNN-developed geroprotectors.

Business Impact of AI

Recall Masters uses machine learning to find owners of recalled cars

When a car company issues a recall, it’s typically on dealerships to reach out to affected customers. But since vehicles can change hands, leaving records out of date, dealers aren’t always able to provide drivers with this at times vital information. One company that addresses this issue is Recall Masters, founded by programmer Chris Miller.

Recall Masters, which employs 20 people and even a lobbyist in Washington, D.C., collects data from more than 50 different sources, then utilizes machine learning to analyze it. The startup can then determine if a vehicle qualifies for a recall and who its current owner is — even if it has been resold multiple times — by poring over billions of transactions, according to Miller. He dubs the process “digital forensics.”

Something to Get Involved In

Diversity in the workplace survey

Alzbeta Dlha, a student at UCL, is currently writing a paper on bridging the gender gap in tech startups, aiming to help growing startups attract more talented female engineers and boost team diversity. It would be great if you could help her make the research more relevant by filling in this short survey about your pain points in HR management (it takes under 3 minutes).

In return, she will be happy to share the results of her work and best practices for diversity management from industry leaders (Google, Facebook, Amazon, McKinsey). Also it would be great if you could forward this to your friends who might be interested in this.

Stats that Impress

AI to have dramatic impact on business by 2020

Tata Consultancy Services, a leading global IT services, consulting and business solutions organization, unveiled its Global Trend Study titled, “Getting Smarter by the Day: How AI is Elevating the Performance of Global Companies.” Focused on the current and future impact of AI, the study polled 835 executives across 13 global industry sectors in four regions of the world,
finding that 84% of companies see the use of AI as “essential” to competitiveness, with a further 50% seeing the technology as “transformative.”

Exploring the views and actions of decision makers from global companies with average revenues of $20 billion, the study revealed AI is spreading across almost all areas of a company. The biggest adopters of AI today are, not surprisingly, IT departments, with two-thirds (67%) of survey respondents using AI to detect security intrusions, user issues and deliver automation. However, by 2020, almost a third (32%) of companies believe AI’s greatest impact will be in sales, marketing or customer service, while one in five (20%) see AI’s impact being largest in non-customer facing corporate functions, including finance, strategic planning, corporate development,
and HR.

Future of Transportation

Alibaba invests in WayRay, a maker of augmented-reality dashboards for smart cars

After launching its first car last year, Alibaba is digging deeper into the automobile industry. The Chinese Internet and e-commerce behemoth is the lead investor in smart car tech developer WayRay’s $18 million Series B round, the startup announced.

Founded in 2012, WayRay makes holographic navigation systems. According to its funding announcement, WayRay has already spent $10 million of its own capital, as well as previous venture funding, on the technology that underpins Navion, an augmented-reality dashboard that overlays directions and other information onto a driver’s view of the road. The company plans to launch a consumer version of Navion in 2017.

Podcasts We Love

Max Ogden on data preservation, distributed trust, and bringing cutting-edge technology to journalism

In this special episode of the Data Show, O’Reilly’s Jenn Webb speaks with Maxwell Ogden, director of Code for Science and Society. Recently, Ogden and Code for Science have been working on the ongoing rescue of and assisting with other data rescue projects, such as Data Refuge; they’re also the nonprofit developers supporting Dat, a data versioning and distribution manager, which came out of Ogden’s work making government and scientific data
open and accessible.

He said, “I would say in general our three focus areas are access to research data for scientists, and then access to public data for journalists, and then access to government data for civic hackers or governments that want to publish data themselves. Those are government, science, and journalism, are the three most exciting public data areas that I think need a lot better tools for a lot of this stuff.”

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

March 15, 2017

Issue 174: CognitionX Data Science, AI and Machine Learning

We love bringing the community together and as such, we have some great events coming up within the next few months. We’ve added an events section in the newsletter so you can keep track and ensure you don’t miss out on tickets. Please do send us suggestions of other events, speakers and panelists, or put yourself forward 😀.

We are also excited to announce that Siavash Mahdavi will be joining our “Investing in AI” panel. He’s got the full spectrum of experience, having started Within– a 3D printing company with AI at the heart (which was sold to Autodesk for $90m) and most recently started a new venture A.I.Music *AND* he’s an angel investor. Check out our latest blogpost to learn more about him.


Tabitha UntiltheBotsTakeOver Goldstaub

Dates for Your Diary

You don’t want to miss CognitionX’s upcoming events

Monday March 27: Investing in AI (if you want to know how many real investment opportunities you can expect to see this year, how to craft a sensible investment strategy, or how your startup can access investment, then come to our event to network and meet the people you need to know)

Thursday April 13How AI Is Transforming The Online Customer Experience (featuring Jedidiah Francis: Head of Data Science, ASOS)

Tuesday May 2Why Women in AI (empathy, nurturing, listening, multi-tasking, intuition, teaching and mothering are skills and qualities we need to be involved in training AI machines. The question is how to ensure they are at the table)

Business Impact of AI

Adobe is building an AI to automate web design

Adobe, one of the world’s largest and most powerful software companies, is trying something new: It’s applying machine learning and image recognition to graphic and web design. In an unnamed project, the company has created tools that automate designers’ tasks, like cropping photos and designing web pages. Should designers be worried?

The new project, which uses Adobe’s AI and machine learning program Sensei and integrates into the Adobe Experience Manager CMS, will debut at the company’s Sneaks competition later in March. While Adobe hasn’t committed to integrating it into any of its products, it’s one of the most ambitious attempts to marry machine learning and graphic design to date. There have been efforts to use AI in the design world before—for instance, Wix’s Advance Design Intelligence and automated projects like Mark Maker, but Adobe’s is notable because of the company’s sheer reach in the design world. Although it’s just a prototype, it’s one to watch closely.

Future of Health

Improving health through AI and machine learning

The Cochrane Transform Project is now applying AI and machine learning to analyze thousands of reports to automatically select ones to include in systematic reviews. Systematic reviews bring together the best available research evidence from individual clinical trials and study data from around the world to inform the development of guidelines, individual practice decisions and national-scale health policymaking.

This new approach is successfully saving weeks of monotonous work, freeing up the expert health care reviewers to spend their time and energy on high-level analysis. This will speed up bringing the latest health interventions to the UK National Health Service, and beyond.

Professor James Thomas at University College London is using Cortana Intelligence to quickly develop and deploy AI solutions in the cloud. He explains, “What makes this particularly efficient is the fact that I can build a classifier using the studio and then just deploy it as a web service with the click of a button, without deploying a server.”


Everypixel trains neural network to measure aesthetic beauty in stock images

A new stock image search engine, Everypixel, is beta testing its unique algorithm to measure the aesthetics of stock images through neural networks.

Everypixel’s team trained a neural network to see the beauty in photos the same way humans do. While the company specifically built the algorithm to identify and weed out the most aesthetically-pleasing stock images from the ugly ones, it also works with basically any type of image.

To develop its algorithm, the team at Everypixel asked designers, editors and experienced stock photographers to help generate a training dataset.They tested 956,794 positive and negative patterns, and their “‘Heartless algorithm’ learned to see the beauty of shots in the same way as you do.”


DeepMind is enabling continual learning in neural networks

Computer programs that learn to perform tasks also typically forget them very quickly. Demis Hassabis et. al. show that the learning rule can be modified so that a program can remember old tasks when learning a new one. This is an important step towards more intelligent programs that are able to learn progressively and adaptively.

The ability to learn tasks in succession without forgetting is a core component of biological and artificial intelligence. In this work they show that an algorithm that supports continual learning—which takes inspiration from neurobiological models of synaptic consolidation—can be combined with deep neural networks to achieve successful performance in a range of challenging domains. In doing so, they demonstrate that current neurobiological theories concerning synaptic consolidation do indeed scale to large-scale learning systems. This provides prima facie evidence that these principles may be fundamental aspects of learning and memory in the brain.

Products We Love

Japan takes a big step toward widespread drone delivery service

Japan, perhaps more than any nation on Earth, has a deep history with autonomous drones. Its companies have been using them for decades to assist with agriculture, infrastructure inspection, and construction. Yesterday one of Japan’s biggest tech companies, Rakuten, announced it was forming a joint venture with the American startup AirMap. The goal is to develop a robust traffic management system for unmanned aerial vehicles, allowing large numbers of drones to operate autonomously in the same airspace.

Rakuten, which is best known as an e-commerce company, has been experimenting with drone delivery since June of last year. Like Amazon, it wants to enable customers to order something online and have it delivered to their doorstep, or windowsill, in under an hour. The company sees low-altitude airspace as a wide-open market, where the only competition comes from birds and radio waves.

Chat Bots, yadda yadda yadda

This Indian AI startup is helping huge brands to get on the chat bot bandwagon

“AI is changing the way we interact with technologies across multiple industries. In a fast-growing market such as India, AI helps making technology-based companies more efficient,” says Sachin Jaiswal, cofounder of This AI startup has recently launched the chatbot SDK to help brands deliver the “conversational” experience consumers are demanding on mobile and web apps.

HDFC Bank, Oxigen Wallet, Intex Smartphones and booking service Ticketgoose are some of the early adopters of its chatbot service. Two years ago, started out by developing chatbot programmed to respond to the chat requests of users for services such as cab ordering, food delivery and phone credit top-ups, and has now served over 50 million interactions.

Ethics Question for the Day

Machine learning can also aid the cyber enemy: NSA research head

“One might find that an adversary is able to control, in a big-data environment, enough of that data that they can feed you in misdirection,” said Dr Deborah Frincke, head of the Research Directorate (RD) of the US National Security Agency/Central Security Service (NSA/CSS).

As one example, an organisation may decide to use machine learning to develop a so-called “sense of self” of its own networks, and build a self-healing capability on top of that. But what if an attacker gets inside the network or perhaps was even inside the network before the machine learning process started?

“Their behaviour now becomes part of the norm. So in a sense, then, what I’m doing is that I’m protecting the insider. That’s a problem,” Frincke said.

Education and Advice We Rate

How to install a Python 3 environment on Mac OSX for machine learning and deep learning

It can be difficult to install a Python machine learning environment on Mac OS X. Python itself must be installed first, and then there are many packages to install, and it can be confusing for beginners.

In this tutorial, you will discover how to setup a Python 3 machine learning and deep learning development environment using macports. After completing this tutorial, you will have a working Python 3 environment to begin learning, practicing, and developing machine learning and deep learning software.

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

March 14, 2017

Issue 173: CognitionX Data Science, AI and Machine Learning

Just yesterday Intel agreed to buy Israeli autonomous vehicle technology firm Mobileye for over $15B. DeepMind was only 4 years old when it was sold to Google for over $600M. We know that AI companies are hot and are seeing excellent exits, but what is the best way for you to invest?

We are excited to announce our next event on the 27th of March, “Investing in AI: Is this for real or should I wait?” Tickets are free, but to guarantee a seat you only need to pay £10.

We will have an awesome panel of expert investors, including Christoph Auer-WelsbachMichael Axelgaard, and Richard Muirhead. We will soon be announcing the addition of another VC partner and a founder of an AI company.  The discussion will be chaired by Charlie our CEO and Founder.

If you are want to know how many real investment opportunities you can expect to see this year, how to craft a sensible investment strategy, or how your startup can access investment, then come to our event to network and meet the people you need to know!


Tabitha UntiltheBotsTakeOver Goldstaub

Prediction of the Day

Mark Cuban: The world’s first trillionaire will be an AI entrepreneur

Bill Gates is the richest man in the world right now, with more than $85 billion to his name, and, according to one estimate, if he makes it to his mid-80’s, he will likely be the world’s first trillionaire. But self-made billionaire Mark Cuban predicts that the world’s first trillionaires will actually be entrepreneurs working with artificial intelligence.

“I am telling you, the world’s first trillionaires are going to come from somebody who masters AI and all its derivatives and applies it in ways we never thought of,” says the star investor of ABC’s “Shark Tank,” speaking to a packed house in Austin at the SXSW Conference and Festivals Sunday night.

Chat Bots, yadda yadda yadda

Forksy is the chat bot that easily tracks users’ food intake

Manually logging every thing we eat throughout the day can be time consuming. And while we have every intention to use that calorie counting app, it isn’t before long before we stop using it consistently. But there now is a more simple way for users to keep track of a food diary. In fact, it’s as easy as chatting to a friend—or in this case, to a bot.

Meet Forksy, the chat bot for tracking calories and eating healthy that available for Facebook Messenger and other messaging apps like Viber, Kik and Telegram.

The best part is that unlike nutrition and food tracking apps already out there, Forksy allows users to speak to it in their natural language to log their food. For example, the user can tell Forksy they had a cup of coffee and bagel with cream cheese for breakfast, and the bot will bring the nutritional info for those items. Just type the food items, or as how you would speak it like ,”I had a burger with fries and a soda.”

Deal of the Day

Lantern Credit buys machine learning library to enhance offers

Lantern Credit, a financial technology company working to solve systematic inefficiencies in the consumer credit industry, is enhancing its proprietary machine learning engine, Beam AI, with the acquisition of the Abstract Regression-Classification (ARC) Machine Learning Library.

The machine learning library enables Lantern Credit to use a human-machine hybrid learning approach that incorporates human guidance in the machine learning training process to produce more reliable outputs.

Lantern Credit’s Beam AI will use the symbolic regression technology to ensure that credit offers presented to consumers are actionable and timely.


Neurala provides a brain for drones to enable autonomy, AI and more

 Hoping to understand how Neurala can help the drone industry with AI, Commercial UAV News reached out to Neurala’s CEO Massimiliano “Max” Versace to ask about what Neurala can offer in terms of AI, how their offering is different and plenty more.

When asked what Neurala is currently working on he said, “We are working with major drone companies for both consumer applications and for commercial inspection applications. As an example of consumer application is the Teal Drone that will be bundled with The Neurala Brain. We are also working with Parrot.

On the commercial side, we are working with several players for inspections of power lines, cell towers and solar panel arrays. By highlighting the areas of interest, The Neurala Brain can save much more than it costs in inspection effort. Additionally, we are working with special-purpose drone manufacturers in the search & rescue space.”


How AI became a key technology in finding missing and exploited children

In an effort to find missing children, the National Center for Missing & Exploited Children and Intel Corp. recently formed a new program, called Intel Inside, Safer Children Outside, to apply AI to the problem.

During the South by Southwest event held in Austin TX, an “AI for Good: Harnessing Power to Solve Problems” panel convened. “We have assisted in the recovery of 220,000 missing children in that time, and we are also involved in recovering exploited children,” said Mark Gianturco, CTO of the National Center for Missing & Exploited Children.

The session, moderated by Lisa Thee, AI and analytics solution owner at Intel, covered how the technology industry is coming together using AI to uncover patterns by analyzing vast amounts of data to resolve missing children cases.

Podcasts We Love

AI adoption at the atomic level of jobs and work

Last week, Jenn Webb sat down with David Beyer, an investor with Amplify Partners. They talk about machine learning and artificial intelligence, the challenges he’s seeing in AI adoption, and what he thinks is missing from the AI conversation.

He thinks that AI adoption is actually a multifaceted question. It’s something that touches on policy at the government level. It touches on labor markets and questions around equity and fairness. It touches on broad commercial questions around industries and how they evolve over time. There’s many, many ways to address this. He thinks that a good way to think about AI adoption at the broader, more abstract level of sectors or categories is to
actually zoom down a bit and look at what it is actually replacing.

Products We Love

Baidu launches SwiftScribe, an app that transcribes audio with AI

Baidu, the Chinese company operating a search engine, a mobile browser, and other web services, is announcing today the launch of SwiftScribe, a web app that’s meant to help people transcribe audio recordings more quickly, using AI.

Baidu in the past few years has been honing its DeepSpeech software for speech recognition. Last year, the company introduced TalkType, an Android keyboard that, using DeepSpeech, puts speech input first and typing second, based on the idea that you can enter information more quickly when you say it than when you peck. Now Baidu is coming out with another app enhanced with DeepSpeech, one that could arguably find better footing in a professional setting.

Today, Baidu is providing SwiftScribe as a free service — unlike Nuance’s Dragon software. “But in the future, we hope to turn it into a business,” Baidu project manager Tian Wu said.

Tools of the Trade

Xero intros machine learning system for SMB invoicing

Cloud accounting software provider Xero announced a new product that it says will bring machine learning and AI to the process of small business invoicing.

The machine learning automation technology is currently without an official title, but Xero says it’s designed to simplify and improve the way invoices are filed. More specifically, the technology will automatically suggest the appropriate account codes for an invoice based on past invoicing behaviour of the business owner and their advisor.

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

March 13, 2017

Issue 172: CognitionX Data Science, AI and Machine Learning

Data, data, everywhere and not a drop to drink. Sir Tim Berners-Lee, founder and inventor of the World Wide Web, warned yesterday of the ‘lost control of our personal data,’ as we continue to give it away for free content. He sees this as one of 3 major challenges for the web outlined here.

How do you feel about the ‘loss of your personal data’? Click below to tell us and add a comment.

Necessary          Necessary Evil          Evil


Tabitha UntiltheBotsTakeOver Goldstaub

Ethics Question for the Day

Three challenges for the web, according to its inventor

On the 28th birthday of the World Wide Web, its founder and inventor Sir Tim Berners-Lee shared three challenges of the web: 1) We’ve lost control of our personal data, 2) it’s too easy for misinformation to spread on the web, and 3) political advertising online needs transparency and understanding.

He argues that “these are complex problems, and the solutions will not be simple. But a few broad paths to progress are already clear. We must work together with web companies to strike a balance that puts a fair level of data control back in the hands of people, including the development of new technology like personal ‘data pods’ if needed and exploring alternative revenue models like subscriptions and micropayments. We must fight against government over-reach in surveillance laws, including through the courts if necessary.”

Deal of the Day

Intel to buy Israeli Mobileye for $14-$15 billion

U.S. chip giant Intel has agreed to buy Israeli technology firm Mobileye for $14-$15 billion, according to TheMarker, an Israeli financial newspaper. The companies will announce the acquisition, the largest ever for an Israeli high-tech company, later on Monday, TheMarker reported on its website. Mobileye is a leading supplier of collision-avoidance car sensor systems.

The two companies are already collaborating with BMW on a project to put a fleet of around 40 self-driving test vehicles on the road in the second half of this year. BMW announced its partnership with the two firms in July, with the goal of developing the capability of introducing fully autonomous vehicles to the market by 2021.


Baxter the friendly robot functions using mind control

Communication with a robot using brain activity from a human collaborator could provide a direct and fast feedback loop that is easy and natural for the human, thereby enabling a wide variety of intuitive interaction tasks. This paper explores the application of EEG measured error-related potentials (ErrPs) to closed-loop robotic control. ErrP signals are particularly useful for robotics tasks because they are naturally occurring within the brain in response to an unexpected error.

The authors decode ErrP signals from a human operator in real time to control a Rethink Robotics Baxter robot during a binary object selection task. They also show that utilizing a secondary interactive error-related potential signal generated during this closed-loop robot task can greatly improve classification performance, suggesting new ways in which robots can acquire human feedback.


Meet AI-CD β, the Japanese AI creative director

In 2015, ad agency McCann Japan’s creative planner Shun Matsuzaka set himself a task he called the “creative genome project”: he wanted to create the world’s first AI creative director, capable of directing a TV commercial. And last week, Matsuzaka showed off his creation at the UK advertiser trade body ISBA’s annual conference in London.

The bot was tasked with creating a TV commercial for Clorets Mint Tab. The bot first crafted “a creative brief” and then drew up the various elements of the TV ad, including tone, manner, music, and context. The two spots would then be thrown to a nationwide poll, where consumers could vote for which ad they preferred.

Fortunately for McCann’s human creative director, his ad narrowly won 54% of the public vote. However, when the 200-or-so advertising executives at the ISBA Conference
were asked which they preferred, they voted for the crazy dog spot, directed by the robot.

Chat Bots, yadda yadda yadda

A New York startup shakes up the insurance business

Daniel Schreiber and Shai Wininger, tech entrepreneurs with no insurance background, spotted that the insurance industry is huge (worth $4.6trn in global premium income a year, reckons Swiss Re, a reinsurer), distrusted, antiquated and hopelessly unreformed.

In September they started Lemonade, a New York-based insurer for homeowners and renters. Some describe it as a peer-to-peer insurer (“Spiritually we’re a tech company,” says Schreiber). Instead of underwriters it uses algorithms; and instead of expensive brokers and salespeople it uses chatbots. It even uses AI and machine-learning to handle claims, a job typically seen as needing a human touch.


AI figured out the word people text when their suicide risk is high

If you were asked to guess the words people use when they’re most at risk for suicide, you’d be right to think of obvious nouns and verbs like die, overdose and, yes, the word suicide itself. So when Crisis Text Line, a free mental health support service, built an algorithm to flag high-priority texts, it included those among 50 words to indicate the person messaging desperately needed help.

But when Crisis Text Line started using AI to analyze the 22 million messages about emotional distress in its database last summer, its researchers made a surprising discovery: The word ibuprofen was 16 times more likely to predict the person texting would need emergency services than the word suicide. Another highly predictive type of content wasn’t even a word but a crying face emoji. When people included that sad character in their messages, Crisis Text Line supervisors were 11 times more likely to call 911 for assistance.

Future of Transportation

California could get truly driverless cars with new rules

The California Department of Motor Vehicles (DMV) last Friday released proposed regulations to establish a path for testing and future deployment of fully autonomous vehicles without drivers.

“California has more manufacturers testing autonomous vehicles than any other state and today’s rules continue our leadership with this emerging technology,” said California Transportation Agency Secretary Brian P. Kelly. “These rules protect public safety, promote innovation and lay out the path for future testing and deployment of driverless technology. This rulemaking is the next step in working with stakeholders to get this right.”

“These rules expand our existing autonomous vehicle testing program to include testing vehicles where no driver is present,” said DMV Director Jean Shiomoto. “This is the next step in eventually allowing driverless autonomous vehicles on California roadways.”


DeepWarp: roll your eyes (in any picture) with deep learning

See deep learning in action with DeepWarp by uploading any picture you want.

This tool allows you to input a picture of any face and it will output the face moving its eyes, either side-to-side, up-and-down, or in a circle.

In the creators’ words: “The proposed system takes an input eye region, feature points (anchors) as well as a correction angle and sends them to the multi-scale neural network predicting a flow field. The flow field is then applied to the input image to produce an image of a redirected eye. Finally, the output is enhanced by processing with the lightness correction neural network.”

Video Killed the Radio Star

AI expands into the artistic realm

Can machines make art and music that moves us? Engineers and artists are testing that notion with an array of new AI that is expanding the boundaries of how imagery, music and video games are created.

The Wall Street Journal recently came out with a video which presented the latest examples of AI’s impact on art from companies like Sony and Google. For example, they discussed Flow Machines which created a song based on Beatles’ sheet music it was fed. In addition, they discuss Adobe’s web brush which is a digital paintbrush which has the look and feel of an actual paintbrush.

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

Subscribe to the News Briefing Now

Stay up to speed with the latest news, developments and industry trends in the world of all things AI with our curated daily or weekly briefing.

Log in with your credentials


Forgot your details?

Create Account