Hay Festival 2017: Stephen Fry

At the Hay Festival last week, he gave a powerful, impassioned talk on the future of the internet and technological advancement.

Fry criticised the “technophobes”, including politicians, who he said had been too slow to react to developments like AI. More importantly, he said that we must prepare
for this future: “Whether it is winter that is coming, or a new spring, it is entirely in our hands so long as we prepare”.

+So….how can you prepare? You can come to CogX London 2017 (June 20-21) to learn about and discuss the impact of AI on society, industry, and government from world thought leaders. CogX will move the conversation forward on the future of insurance, retail, government, health, transportation, and more. Trust me- you will not be disappointed.

Apple is working on a dedicated chip to power AI on devices

Apple is working on a processor devoted specifically to AI-related tasks, according to a person familiar with the matter. The chip, known internally as the Apple Neural Engine, would improve the way the company’s devices handle tasks that would otherwise require human intelligence — such as facial recognition and speech recognition, said the person, who requested anonymity discussing a product that hasn’t been made public. Apple declined to comment.

“Two of the areas that Apple is betting its
future on require AI,” said Gene Munster, former Apple analyst and co-founder of venture capital firm Loup Ventures. “At the core of augmented reality and self-driving cars is artificial intelligence.” ARM also just announced their latest AI chip.

+Want to know more about Apple and their latest innovations? Then you do not want to miss
Shiva Rajamaran (Product Lead, Apple) at CogX. He was previously ‘the product guy’ at Youtube, Spotify, Twitter, Google.

Data science and deep learning in retail

In this episode of the Data Show, Ben Lorica spoke with
Jeremy Stanley, VP of data science at Instacart, a popular grocery delivery service that is expanding rapidly. As Stanley describes it, Instacart operates a four-sided marketplace comprised of retail stores, products within the stores, shoppers assigned to the stores, and customers who order from Instacart. The objective is to get fresh groceries from popular retailers delivered to customers in a timely fashion. Instacart’s goals land them in the center of the many opportunities and challenges involved in building high-impact data products.

+Fascinated by how retail and the online customer experience? Come to CogX to hear the likes of Kenneth Cukier (The Economist) and Mike Hyde​ (Data Science Director at Facebook) discuss the changing landscape.

The $1700 great deep learning box: assembly, setup
and benchmarks

Slav Ivanov has written a great how-to guide on how to get started with deep learning and what software and hardware to use for it. He covers:

1. Choosing components

2. Putting it together

3. Software Setup

4. Benchmarks

This app uses AI to turn design mockups into source code

Copenhagen-based startup UIzard Technologies has leveraged the latest developments in the field of machine learning to build a neural network that, once fed with raw screenshots of graphical user interface, proceeds to automatically generate code.

What is particularly intriguing is that the so-called Pix2Code model has the capacity to produce code for three different platforms, including Android and iOS as well as other web-based technologies.

As UIzard founder Tony Beltramelli explains in his research, the novel approach could potentially “end the need for manually-programmed” user interfaces altogether. At present, the method generates code from screenshots with an impressive accuracy of over 77 percent, but the consistency of the algorithm is likely to improve in the future.

Tesla’s former Autopilot head is launching a self-driving-car company

Aurora Innovation is a new startup led by Chris Urmson, the former head of Google’s autonomous car team, and Sterling Anderson, the former director of Tesla Autopilot. Drew Bagnell, Uber’s former autonomy and perception lead, has also joined Aurora as CTO.

Aurora plans to develop the hardware, software, and data services necessary to build an autonomous driving platform — a similar approach to Uber and Waymo, the self-driving venture spun out of Google’s parent company.

Aurora plans to work directly with Tier 1 suppliers to design the best sensors to feed the algorithms that will create the car brain. That involves taking on tech giants when it comes to designing complex hardware like lidar, a key sensor that helps vehicles detect objects, which has become the focal point of a lawsuit between Uber and Waymo.

June 20-21: CogX: The AI Innovation Exchange

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

IBM’s introduces new PowerAI features

Recently, IBM announced a significant revamp of PowerAI (their AI software toolkit), seeking to address some of the bigger challenges facing developers and data scientists—cutting down the time required for AI system training significantly, and simplifying the development experience, including:

•  new software tool known as AI Vision, which seeks to improve ease of use for AI developers using images and
  videos.
• introduced a new cluster virtualization software called Spectrum Conductor
• new distributed computing edition of the popular open-source machine learning framework, TensorFlow
•  software tool called DL Insight, which IBM says will make model development easier and more accurate

+IBM fan? They’ll be at CogX ….obviously…

Football: A deep dive into the tech and data behind the best players in the world

S.L. Benfica—Portugal’s top football team and one of the best teams in the world—makes as much money from carefully nurturing, training, and selling players as actually playing football. They are buying young talent; using advanced technology, data science, and training to improve their health and performance; and then selling them for tens of millions of pounds—sometimes as much as 10 or 20 times the original fee.

With machine learning and predictive analytics running on Microsoft Azure, combined with Benfica’s expert data scientists and the learned experience of the trainers, each player receives a personalised training regime where weaknesses are ironed out, strengths enhanced, and the chance of injury significantly reduced.

+Machine learning and data science are having a tremendous impact on all forms of entertainment. Come to CogX to find out more.

DataRobot acquires Nutonian

DataRobot, a leader in automated machine learning, today announced it has acquired Nutonian, Inc., a data science software company specializing in time series analytical modeling. Terms of the acquisition were not disclosed, and the deal is officially closed.

Developed in Cornell’s Artificial Intelligence Lab by two of the “World’s Most Powerful Data Scientists,” Nutonian’s AI-powered modelling engine, Eureqa, powers predictive and prescriptive analytics at global companies, including Audi, Beck’s Hybrids, NASA, and RealPage. Eureqa is renowned for its success in time series analytics and for creating easy-to-interpret predictive models in minutes rather than weeks or months.

Tomorrow’s medical breakthrough? You’re already wearing it

It turns out the Apple Watch is a pretty mediocre heart rate sensor by medical standards, but researchers at the University of California, San Francisco used an app called Cardiogram with some machine learning to make the Watch a 97 percent accurate detector of one of the leading causes of stroke.

That remarkable bit of bootstrapping points to the broader trend of everyday tech becoming medical gear: Connected thermostats, light switches, voice assistants and cars all have data about our daily patterns that can be turned into wellness information when enough of it is gathered — what’s called “big data” — and sorted through by machines to spot patterns human observers would likely miss — AI.

Check out Blippar, the Shazam for cars

If you walk into a store, only to hear the latter half of an interesting song playing over the store’s speakers, you can identify what it is using the app Shazam. This same kind of functionality has now been launched for cars, though not by the company behind Shazam. Called Blippar, this app can tell you the make and model of any car using a smartphone’s camera. That may sound a bit useless, since most cars have this information on the truck, but Blippar adds some extras into the mix.

Blippar comes from the company by the same name, and it aims to make it easier for consumers to identify the cars that strike their fancy. The app supports all cars on the US market that were built after the year 2000, and it is said to have a 97.7-percent accuracy. The app taps artificial intelligence for its abilities, promising performance above what a human could offer.

Machine learning set to shake up equity hedge funds

Jeff Tarrant, the founder of Protégé Partners, says, “Jeff Bezos picked off the bookstore business. Apple totally picked off the music business and Netflix totally changed television. Now [machine learning] is going to pick off the hedge funds.”

To back up that theory, he is launching a business that will invest solely in start-up investment funds that employ artificial intelligence. Protégé’s new business, dubbed Mov37, will invest in as many as 10 managers through either seeding them or investing directly, and is targeting total investments of up to $1bn. While many of the biggest systematic hedge funds, which use computer algorithms to make trading decisions, have already started employing machine learning techniques, Mr Tarrant believes it is the young emerging managers who will succeed in disruption.

+Want to know more about how machine learning is disrupting financial as a whole? Good news.
Michael Harte, Chief Innovation Officer at Barclays will be at CogX. You don’t wanna miss you.

June 20-21: CogX: The AI Innovation Exchange

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

When will AI exceed human performance? Evidence from AI experts

Advances in AI will transform modern life by reshaping transportation, health, science, finance, and the military. To adapt public policy, we need to better anticipate these advances. In this paper, the authors report the results from a large survey of machine learning researchers on their beliefs about progress in AI.

Researchers predict AI will outperform humans in many activities in the next ten years, such as translating languages (by 2024), writing high-school essays (by 2026), driving a truck (by 2027), working in retail (by 2031), writing a bestselling book (by 2049), and working as a surgeon (by 2053). Researchers believe there is a 50% chance of AI outperforming humans in all tasks in 45 years and of automating all human jobs in 120 years, with Asian respondents expecting these dates much sooner than North Americans.

Researchers have created an AI that is naturally curious

Researchers at UC Berkeley have produced an AI that is naturally curious. They tested it successfully by having it play Super Mario and VizDoom (a rudimentary 3-D shooter), as the video below shows.

While the AI that was not equipped with the curiosity ‘upgrade’ banged into walls repeatedly, the curious AI explored its environment in order to learn more. Pulkit Agrawal, a Ph.D. student at UC Berkeley, a member of the team, compared it to babies, who “do all these random experiments, and you can think of that as a kind of curiosity.” Brenden Lake, a Data Science Fellow at New York University said in an email to MIT that this work is encouraging. “Developing machines with similar qualities is an important step toward building machines that learn and think like people.”

How is AI impacting creativity

Check out this thought-provoking article from
Julian Harris, Ex-Googler, Product Manager, and Innovation Consultant, in which he explores how AI is impacting creativity.

Creativity is one of the most human qualities. So how is the rapid evolution of AI and machine learning affecting our creativity? Here’s a simple template:

• Clarify: what do we mean by “creative”?
• Illustrate: show examples of both products and project
  into the future based on research today around AI creativity.
• Act: Where does this leave this human skill? What should we do concretely?

Pinterest’s Lens now lets foodies find similar dishes by snapping a photo

Pinterest has just launched a handful of new ways to find recipes using its Lens and Search tools.

While Lens already recommended recipes for certain ingredients, such as coconut and squash, it’s now going the extra mile by unearthing recipes related to an entire dish. Examples provided by Pinterest include waffles, quesadillas, bruschetta, and fried chicken, with computer vision smarts not only recognizing what’s on the plate but also identifying similar existing recipes and showing the user how to make the dish.

Google Attribution is a free and easy way to evaluate marketing efforts

Up until recently, most marketers would credit any sale to the last touchpoint their company had with a customer. This flawed strategy allows marketers to quantitatively evaluate marketing campaigns, but its more of a heuristic than a ground truth.

New machine learning-powered methods enable marketers to model the relative contribution of disparate advertising efforts. This is a much more informative method of evaluation. In the real world, video ads, banner ads, emails and other materials all work
in consonance to drive conversions — it doesn’t make sense for the final social media ad to get all the credit.

“We capture the clicks, as long as there was a click, we can, for example, say how many of the conversations were from the social channel,” explained Babak Pahlavan, Google’s senior director of product management for analytics and measurement, in an interview.

Roborace’s Robocar, the first autonomous car tested in Paris

Robocar may not be the first car to race through Paris, but it is the first to do it autonomously. At Formula E’s Paris ePrix, the Robocar negotiated the 1.9-kilometer circuit’s 14 turns without human intervention, and the company says this is the first time an autonomous car has been tested on Parisian streets.

On the computer side, the car “sees” with five lidar emitters, two radar emitters and 18 ultrasonic sensors, six cameras and two optical speed sensors. The data crunching is done by a Nvidia Drive PX2 computer that can perform 24 trillion A.I. functions a second.

Nasdaq invests in AI

Following a string of market-disrupting financial scandals, stock exchange Nasdaq needed a more effective way to detect financial fraud that didn’t depend on older lexicon search tools. After months of searching, it found a way — via technology based on AI.

With the help of a cognitive computing platform from Digital Reasoning, Nasdaq is now capable of detecting divulging language in trader communications based on communication patterns. It can analyze data in order to better understand a trader’s intentions and actions, thereby exposing potential fraud.

Twitter announces new ways to engage with chat bots

This week, Twitter launched a new, customizable Direct Message Card that businesses can use to promote and share bots and other customer experiences built in Direct Messages. This new card will help businesses drive discovery of such experiences both through Promoted Tweets and organic sharing.

Several brands around the world are launching Promoted Tweet campaigns using the Direct Message Card this week to drive discovery of the experiences they have built. Patrón Tequila, a leading spirits brands, is using the Direct Message Card to entice people to engage with their messaging bot — Bot-Tender — which creates personalized cocktail recommendations based on responses to questions around occasion, flavor, and even emoji.

Thursday May 25: Will robots replace HR? 

June 20-21: CogX: The AI Innovation Exchange

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

Google’s AlphaGo defeats Chinese Go master in win for AI

The world’s best player of what might be humankind’s most complicated board game was defeated on Tuesday by a Google computer program. Adding insult to potentially deep existential injury, he was defeated at Go — a game that claims centuries of play by humans — in China, where the game was invented. The human contender, a 19-year-old Chinese national named Ke Jie, and the computer are only a third of the way through their three-game match this week.

London-based Cera wants to solve the social care crisis with AI

Cera is a medical AI startup looking to change the way we approach social care. Currently, the company offers a marketplace-platform to match carers with patients and a newly released virtual assistant called Martha. Martha can be accessed online or by text message.

While it is relatively simple now, in the future the plan is for Martha to parse a patient’s medical records. The hope is this would allow Martha to send warnings if a patient was behaving strangely or if behaviour could be used to give early warnings of illness.

When AI botches your medical diagnosis, who’s to blame?

Robert Hart, from Quartz, argues that when 1 in 6 patients in the NHS receives an incorrect diagnosis, it is unsurprising that AI could have an important role to play. There have already been examples of AI having more success than human doctors within lung cancer detection and identification of rare eye diseases.

However, what happens when something goes wrong? Hart states that even computer-literate doctors will be unable to find out why an AI has made a specific decision given their ‘blackbox’ nature. This leads to a broader philosophical debate around who is to blame. The AI itself? The developers? The organisation responsible for the AI? As Hart suggests, “Not knowing undermines patient trust, places doctors in difficult positions, and potentially deters investment in the field”.

Researchers develop machine learning-based rehabilitation system

Researchers at the University of Waterloo, in partnership with Cardon Rehabilitation and Medical Equipment, have created an Automated Rehabilitation System to allow people recovering from hip and knee replacements to see how well they are performing rehabilitative exercises.

The system, which combines motion sensors
with software programs, attaches sensors to a person’s limbs — such as above the knee and ankle in knee replacement cases — which send data to a computer while patients do exercises prescribed by physiotherapists. Using human body modelling and machine learning, the system then analyzes, organizes, and stores the data and generates visual representation of the motion.

Effective injury prediction in professional soccer with GPS data and machine learning

Injuries have a great impact on professional soccer, due to their large influence on team performance and the considerable costs of rehabilitation for players. Existing studies in the literature provide just a preliminary understanding of which factors mostly affect injury risk, while an evaluation of the potential of statistical models in forecasting injuries is still missing.

In this paper, the authors propose a multidimensional approach to injury prediction in professional soccer
which is based on GPS measurements and machine learning. By using GPS tracking technology, they collect data describing the training workload of players in a professional soccer club during a season. They show that their injury predictors are both accurate and interpretable by providing a set of case studies of interest to soccer practitioners.

Women in AI – a visual study of leadership across industries

Women in artificial intelligence (AI) and machine learning (ML), or the lack thereof, is not a new topic in media, just as gender equality and disparity in the workplace is not a new subject of research for academics and think tanks. But discussing these issues openly is no less important. While they address the potential reasons and implications of these issues toward the end of this article, their initial interest in this subject came from our desire to know the following:

1. How many women are in C-level and other leadership roles in the AI and ML industries compared to males?
2. How might these numbers compare to other industries and the workforce at large?
3. What are the potential implications of female presence (or lack thereof) in leadership roles within AI and ML companies?

Stanley Robotics is building robots that can park your car for you

French startup Stanley Robotics just raised $4 million (€3.6 million). The company is building giant robots that pick up your car at the entrance of a parking lot and park it for you.

Stanley Robotics plans to take advantage of that with a robot called Stan. It is going to make airport parking lots more efficient through its creation robot valet-parking. All of this sounds great on paper, but the most reassuring thing is that Stanley Robotics is already operating at Charles-de-Gaulle airport in Paris. It’s been years in the making, but a parking lot is now operated by robots.

IKEA’s affordable smart lights will dim with your voice

Last month, IKEA launched its own line of low-cost smart lighting, called TRÅDFRI, and up until now, users have had to rely on a remote control or a proprietary app to use the product. But no longer.

Yesterday, the Swedish retailer announced that their IKEA Home Smart products will respond to voice commands from Alexa, Siri, and Google Assistant starting this summer. Additionally, the product line will integrate with Apple’s HomeKit. “With IKEA Home Smart we challenge everything that is complicated and expensive with the connected home. Making our products work with others on the market takes us one step closer to meet people’s needs, making it easier to interact with your smart home products,” said IKEA Home Smart’s business leader Björn Block.

Thursday May 25: Will robots replace HR? 

June 20-21: CogX: The AI Innovation Exchange

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

Uber Starts Charging What It Thinks You’re Willing to Pay

Uber started using data science to predict how much we are willing to pay for a ride. They call this new system “route-based pricing” and it is a significant shift from the current model which is based on mileage, time and demand.

Uber drivers have been complaining that the gap between the fare a rider pays and what the driver receives is getting wider. After months of unsatisfying answers, Uber is providing an explanation: It’s charging some passengers more because it needs the extra
cash. How will customers and driver react to these changes in business model?

Google’s AlphaGo AI is about to face off against the world’s best Go player

Last year, DeepMind’s AI AlphaGo scored a momentous victory over Lee Se-dol, one of the world’s best Go players, achieving something many thought impossible. Is AlphaGo truly better than the best humanity has to offer? The debate will be settled this week when AlphaGo takes on the reigning world Go champion Ke Jie, who laid down the gauntlet last year after claiming he could beat the AI.

Your exclusive discount to The Europas, June 15, London

Interested in AI and startups? Hear more from experts who will be debating the future of AI’s impact on jobs and the economy at The Europas Conference & Awards 2017 on June 13 in London. The Europas is a one-day conference with awesome speakers and breakouts. This must-attend event will feature speakers, panels and breakout sessions drawn from Europe’s hottest tech startup founders and investors. More details in the
agenda here. We are excited to announce that briefing subscribers of CognitionX can get a 15% discount on the ticket price using this link. Grab your ticket now!

Introducing Seldon Deploy

Seldon deploy is a new enterprise product that will be entering closed beta this summer. Seldon Deploy helps data science teams put machine learning models into production. It’s designed to streamline the data science workflow, with audit trails, advanced experiments, continuous integration and deployment, rolling updates, scaling, model explanations, and more — all served through a delicious UI. Do you want to gain early access to Seldon Deploy? Sign up to their closed beta to provide feedback to help shape the product before launch.

Under the Hood of Google’s TPU2 Machine Learning Clusters

In this (slightly speculative!) article, The Next Platform takes a deep dive into what they believe Google’s second-generation TensorFlow Processing Unit (TPU2) might be capable of. Google’s original announcement was rather light in detail, so in this article the author digs into the photos and provides his thoughts based on the few bits of detail Google did provide.

5 Ways Artificial Intelligence May Help Us Live At Home Longer

Although journalists tend to enjoy writing about how AI is going to take over the world and rule us, there is some good news from Forbes this week on the impact Artificial Intelligence will have on older generations.

Randy Rieland, in interview with Richard Adler, Research Fellow at the Institute for the Future in Palo Alto, California, discusses the thought that despite the initial reticence of older generations to adopt new technology, AI will provide more independence and freedom in old age and allow us to stay in our homes for longer. Not only by having a handy robot to remind you to take your medicine or one that can predict problems, machines are now learning to diagnose diseases and may even learn to back up your memory. Although perhaps still a way off, the potential benefits of a human memory ‘backup’ for those suffering with dementia or schizophrenia could be monumental.

Curiosity May Be Vital for Truly Smart AI

The relatively new field of AI curiosity (a field pioneered by Pierre-Yves Oudeyer) has been incorporated by researchers in UC Berkeley to successfully demonstrate the benefits of this approach and show how it may be crucial to develop better AI. In contrast to reinforcement learning, which was the technique behind AlphaGo, curiosity driven AI generates positive signals from behaviour relative to an internal score. You can read their paper here.

The practical applications of AI: 6 videos

A great, simple and quick resource to give you an insight into 6 ways AI is being used right now. From cryptocurrencies to the link between gaming and the AI there’s something for everyone. Well worth a look.

Thursday May 25: Will robots replace HR? 

June 20-21: CogX: The AI Innovation Exchange

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

Bill Gates has a message for every college grad who wants to change the world

The founder of Microsoft and Co-Chair of the Bill and Melinda Gates Foundation has an important message for those ambitious students who are graduating from university this year and want to change the world: Artificial Intelligence.

In an amusing open letter written to students, Gates celebrates their success. “Congratulations! You’ve just accomplished something I never managed to do – earn a college degree”. While reflecting on his own career, Gates points to Energy, Biosciences, and Artificial Intelligence as the industries that will fundamentally alter the way we live our lives.

Drones and AI help stop poaching in Africa

Drones have been used to protect herds of elephants and rhinos for several years now but a recent partnership between Lindbergh Foundation and Neurala has taken this to the next level. The drone footage is poured over by an AI that can recognize the pachyderms and poachers, spotting them before they get close to a herd. The Air Shepherd program is a great example of the real good AI can do. Watch below as the drone identifies targets in real-time.

A16Z AI Playbook

“How do I get started with artificial intelligence?” “What can I do with AI in my own product or company?” If you find yourself asking these question, you will find this AI Playbook, put together by Andressen Horozwitz incredibly useful. This microsite was written to help newcomers (both non-technical and technical) begin exploring what’s possible with AI.

An AI invented a bunch of new paint colors that are hilariously wrong

At some point, we’ve all wondered about the incredibly strange names for paint colors. Research scientist and neural network goofball Janelle Shane took the wondering a step further. Shane decided to train a neural network to generate new paint colors, complete with appropriate names.

Writes Shane on her Tumblr, “For this experiment, I gave the neural network a list of about 7,700 Sherwin-Williams paint colors along with their RGB values. (RGB = red, green, and blue color values.) Could the neural network learn to invent new paint colors and give them attractive names?”

Machine learning algorithms re-create city in 3D using only image data

The city of Zurich, Switzerland has been reconstructed in 3D using millions of images and videos. Developed by researchers at ETH Zurich, the Varcity platform pulls from huge volumes of image data and uses algorithms to automatically stitch it all together. Urban city planning, architectural design, traffic modeling, autonomous navigation, and tourist guidance, as well as catastrophe response planning are some of the applications for this new technology.

If You Want to Glimpse the Power of AI, Play These Games

“What we’re trying to do with these AI experiments and demos is to show that it’s beyond academics at this point,” says Suzanne Chambers, the executive producer of the Google Creative Labs. These AI experiments have been featured here before but in light of the recent Google Developer conference it might be worth looking at the Google AI Experiments again. Some are just simple, fun and, engaging, while others offer a glimpse into the systems increasingly affecting our lives. As Google, an “AI-first company,” plunges headlong into this world we can expect more fun experiments to wow and hopefully make us think.

P: A programming language designed for asynchrony, fault-tolerance and uncertainty

Have you heard of P: programming language? This project is a collaborative effort between Microsoft researchers and engineers, and academic researchers at the University of California, Berkeley
and Imperial College in London. Azure, similar to other cloud providers, faces the challenge of Heisenbugs caused by unexpected race conditions and software or hardware faults. These bugs result in disruption of live services — a huge problem for both customers and providers of cloud services. Read here some of the characteristics of P that are transforming the development of cloud infrastructure in Azure.

Thursday May 25: Will robots replace HR? 

June 20-21: CogX: The AI Innovation Exchange

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

Intel-Turing programme helps to detect cancer cells more accurately in hospital with artificial intelligence

Scientists at Warwick University are creating a large, digital repository of a variety of tumour and immune cell images from thousands of human tissue samples. They are developing algorithms to develop an AI to recognise these cells automatically, with an initial focus on lung cancer. In collaboration with Intel, they are improving these models to recognise cellular distinctions associated with various grades and types of lung cancer using artificial intelligence frameworks
such as TensorFlow.

How Salesforce CEO Marc Benioff uses artificial intelligence to end internal politics at meetings

Salesforce CEO Marc Benioff isn’t just predicting that artificial intelligence will one day help run everyone’s companies, he’s already using it at Salesforce today. He’s got a special, not-yet-released version of Einstein, the company’s artificial intelligence tech baked into its products, helping him run his company, he told Wall Street analysts on Thursday.

“This is a capability that I use with my staff meeting, when I do my forecast and I do my analysis of the
quarter, which happens every Monday at my staff meeting … sometimes it will point out a specific executive, which it’s done in the last three quarters, and said, this executive is someone who needs specific attention during the quarter.”

Future of work: 21 major AI / machine learning reports on future of work, labour, and skills

Product Manager and Ex-Googler, Julian Harris has compiled a list of reports focusing on The Future of Work. Given the rise of automation and artificial intelligence in all aspects of the workplace, more and more organisations are starting to decipher what this means for our future.

The reports come from a variety of academic institutions, government bodies and consultancies across the UK. While the focus is on the impact
of automation on employment, some of the reports question how the change will affect the economy, society and industry.

Find out more at CogX we will be discussing Singularity, Happiness and the Future of Work. The conversation will centre around the impact that automation will have on society and how we can stay ahead. The panel is being chaired by Calum Chace, the author of “The Economic Singularity“ & “Surviving AI“ and is certainly one not to miss.

Spotify Has Acquired Machine-Learning Startup Niland

Streaming music powerhouse Spotify has acquired French machine learning startup Niland, who describe themselves as “a music technology company that provides music search & discovery engines based on deep learning & machine listening algorithms.” Niland is the fourth company Spotify has snapped up in 2017 alone, following the acquisitions of Sonalytic, MightyTV and most recently, Mediachain.

Tesla factory workers reveal pain, injury and stress

When Tesla bought a decommissioned car factory in Fremont, California, Elon Musk transformed the old-fashioned, unionized plant into a much-vaunted “factory of the future”, where giant robots named after X-Men shape and fold sheets of metal inside a gleaming white mecca of advanced manufacturing. However, some of the human workers who share the factory with their robotic counterparts complain of grueling pressure – which they attribute to Musk’s aggressive production goals – and sometimes life-changing injuries.

Cinematography on the fly

In recent years, a host of Hollywood blockbusters – including The Fast and the Furious 7, Jurassic World, and The Wolf of Wall Street have included aerial tracking shots provided by drone helicopters outfitted with cameras. A team of researchers from MIT’s Computer Science and Artificial Intelligence Laboratory
and ETH Zurich hope to make drone cinematography more accessible, simple, and reliable, by creating an algorithm that will help drones fly and avoid collision.

Hear Me Out: Let’s Elect an AI as President

Is this really such a mad idea? It seems many today would rather anyone in certain seats of power, but an AI? On paper it seems to offer many benefits; an AI could really try to enact certain policies over its term, align its outlook along the split of the vote using a weighted sum of ideologies. An optimizer that seeks to make the maximum number of people happy… I think we can all see potential problems along that path,however, we increasingly allow machine learning algorithms more and more influence (knowingly and unknowingly) in our day to day lives. Is the idea of explicitly turning over the highest office in the land so far fetched? We for one welcome our future AI overlords.

FlyAI – Flies lives dependent of AI

Flies crawl around inside a transparent sphere, spied on by a camera, as a disembodied voice reads out image recognition data that determines whether or not they will be fed. This is the unsettling theme of FlyAI, a Raspberry Pi-powered art installation by David Bowen, who was inspired to create it after reading philosopher Nick Bostrom’s Superintelligence: Paths, Dangers, Strategies.

If the AI successfully classifies the fly, then the fly is fed. This certainly raises some ethical questions about the use of AI, although, according to Bowen “they live for about 30-40 days, which I’m told is longer than they live in the wild… they were getting kind of geriatric!”

Thursday May 25: Will robots replace HR? 

June 20-21: CogX: The AI Innovation Exchange

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

Artificial intelligence could be as sexist as humans

The newly appointed UK Artificial Intelligence leader at PwC, Euan Cameron, has warned of the risks of tech being designed by overwhelmingly male teams. The obvious issue of the lack of gender diversity in the AI environment particularly on the technical side is that it can lead to potentially hidden or unconscious biases being coded in. Cameron says he’ll be working with companies to highlight these biases and recommend ways to protect against them.

New products and features announced at Google I/O

Google I/O, Google’s annual developer conference and showcase for new products opened yesterday with a keynote from their CEO Sundar Pichai. He reiterated Google’s commitment to an AI-first approach across their entire product range, and announced a slew of products and features. These include Google Lens, a reverse image search allowing users to search the world around them using their phone’s camera; Cloud TPU, a high performance chip specifically designed for training AI algorithms; and standalone headsets for
Daydream, the VR platform announced at last year’s I/O, which can run without a phone and track the user’s position in space.

How AI and machine-learning tools lighten the eDiscovery load

TAR (Technology Assisted Review) is a document classification process based on input from human reviewers, currently gaining traction in several industries. In the case study presented here nearly halfway through a review of 40,800 documents the firm switched to using TAR. Having results from the first half allowed them to jump start the algorithm by pre-training the TAR Engine. Only another 6,800 documents needed to be reviews manually before finishing the work, completing the work 70% faster than expected and saving an estimated $70,000 for the client.

A leaked letter questions DeepMind’s adherence to health data privacy laws

DeepMind was given permission to use 1.6 million patient records from the British National Health Service to test the Streams app, which alerts to right clinician that a patient requires their urgent assistance. The letter, written by Fiona Caldicott, the National Data Guardian, to Stephen Powis, the medical director at London’s Royal Free Hospital, suggests the method of transfer of data between the hospital and DeepMind did not constitute “implied consent.”

AI-powered Google for Jobs has work for everybody

In a slight twist on the usual narrative, the announcement today from Google will help find jobs for non-techies rather than “steal” them away. Google for Jobs is an AI-powered search engine developed in partnership with Linkedin, Glassdoor, Monster, and Facebook. Google’s CEO Sundar Pichai even demoed the system by finding a retail, rather than software role. Using AI to help the wider public in such a visible way is something that can’t be applauded enough and highlights the growing necessity for internet access in the modern
world.

Google wants to be everywhere with everyone

“In an A.I.-first world, we are rethinking all our products” said Mr. Pichai, Google’s chief executive, at Google I/O this week. He envisions a world where people interact seamlessly with their Google devices and are constantly connected to each other. With over a billion users in their Youtube, Google Maps, and Search services, how will Google’s decisions shape the future? Especially with services like Virtual Positioning System; “GPS gets you to the door, and then VPS can get you the exact item you’re looking for,”
said Clay Bavor, Google’s virtual reality team leader.

Bad bots do good: random artificial intelligence helps people coordinate

Unpredictable AI doesn’t sound like a good thing. But a new study shows that computers that behave randomly can push us to better coordinate our actions with others and accomplish tasks more quickly. The approach could ease traffic flow, improve corporate strategy, and possibly even tighten marriages.

Thursday May 25: Will robots replace HR? 

June 19-21: CogX: The AI Innovation Exchange

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

Insurance: Robots learn the business of covering risk

One of the most complex professions in the world is at risk of being replaced. By a selfie. Life insurers are looking not just for basic information, such as gender, but also for clues about how quickly people are ageing, their body mass index, and whether they smoke. Lapetus, a US based start-up, thinks their model is a far more accurate prediction of life expectancy than traditional methods, and the whole process takes only a few minutes. The bespoke risks in society will never go away as there is always innovation in the economy. Nonetheless, AI will not make risks disappear entirely and people will always want to protect themselves against the unexpected.

Google’s AI Invents Sounds Humans Have Never Heard Before

Google Magenta (subset of the Google Brain team) have released some new sounds from there AI, NSynth, and by new sounds we mean fundamentally new. By ingesting a massive database of sounds the neural nets in NSynth developed a representation of a range of instruments and can reproduce them. The innovation is that it can blend these, not by overlaying them, but rather by combining the mathematical representations to produce new sounds.

The full database of sounds and the algorithm used is freely available online.

The Lack of Intelligence About Artificial Intelligence

In conversations about AI much of the opinions expressed are based on lots of “information” that contains little in the way of knowledge. Politicians have a tenuous grasp if any on the systems that are rapidly moving to change our world and government more broadly seems unprepared for the changes that are already here. “What happens to the lawyers, accountants, medical diagnosticians, manufacturers, supply chain managers and customer service representatives when they’re displaced? Where will they go? What will they do?”

BMW just strengthened its alliance to take on Mercedes in the self-driving-car race

McKinsey & Co. predicts car data could become a $750 billion industry by 2030. Companies like Delphi
and startup Otonomo are in the market to collect and sell car data, building platforms that the big car makers desire. As self-driving platforms develop, automakers will partner with these companies because, as Intel CEO Brian Krzanich said, data is “the next oil”.

Partnership on AI strengthens its network of partners and announces first initiatives

The Partnership on AI to Benefit People and Society announced yesterday they are welcoming 22 new members, including tech giants, such as Intel, Sony, and Salesforce, and nonprofits the Electronic Frontier Foundation and UNICEF. The partnership aims to “study and formulate best practices on AI technologies, to advance the public’s understanding of AI, and to serve as an open platform for discussion and engagement about AI and its influences on people and society.”

In addition to increasing their membership, the partnership announced plans to launch a set of activities around the challenges and opportunities in their thematic pillars. These include: topic-specific and sector-specific Working Groups to research and formulate best practices; the creation of a Civil Society Fellowship program aimed at assisting people at non-profits and NGOs who wish to collaborate on topics in AI and society; the formation of a cross-conference “AI, People, and Society” Best-Paper Award; and the start of an AI Grand Challenges series to stimulate aspirational efforts in harnessing AI to address some of the most pressing long-term social and societal issues.

Robots that learn

OpenAI has created a robotics system, trained entirely in simulation and deployed on a physical robot, which can learn a new task after seeing it done once. Using an algorithm called one-shot imitation learning, a human can communicate how to do a new task by performing it in VR. Given a single demonstration, the robot is able to solve the same task from an arbitrary starting configuration.

Ocado’s robo-factory

Wired’s Matthew Reynolds visits Ocado’s warehouse near Birmingham to discover how they are using AI and Automation to reduce human involvement as much as possible. In a 90,000 square foot space in Dordon, 35 kilometres of conveyor belts transfer 50,000 products to ensure that their 190,000 deliveries find their way smoothly to customers every week.

Ocado’s CTO, Paul Clarke explains that the crate’s journey through the warehouse is calibrated entirely by algorithms, including ensuring that heavy items aren’t placed on top of delicate produce, “from the moment an item arrives in the warehouse, Clarke says, a human never touches it until it’s placed into a shopping bag just minutes before it goes out for delivery.”

Thursday May 25: Will robots replace HR? 

June 19-21: CogX: The AI Innovation Exchange

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

Uber Engineer Barred From Work on Key Self-Driving Technology, Judge Says

Uber narrowly avoided having its self-driving car development halted on Monday after a federal judge stopped short of issuing a temporary injunction resulting from a lawsuit brought by Waymo, Google’s self-driving car subsidiary. Anthony Levandowski, the lead engineer on Uber’s programme and former Waymo employee, has been barred from working on a key component of the technology during the litigation due to the accusation he stole trade secrets from Waymo.

AI playbook

Written by Andreessen Horowitz (a16z), this microsite is intended to help newcomers (both non-technical and technical) begin exploring what’s possible with AI. a16z has met with hundreds of Fortune 500 / Global 2000 companies, startups, and government policy makers asking: “How do I get started with artificial intelligence?” and “What can I do with AI in my own product or company?” The survey has been written specifically as a starting place for business and technical people just beginning their own journeys with artificial intelligence and who are interested in teasing apart what’s
real from what’s hype.

Facebook to launch ParlAI, a testing ground for AI and bots

FAIR (Facebook Artificial Intelligence Research) have announced the opening of an open-source testing facility for AI and robotics platforms. One fascinating branch of the research will be the interaction between bots and the collective possibilities of many bots together. Interestingly, all researchers that will be accepted must know python and use open-source platforms to develop their AI. The purpose of ParlAI, said director of Facebook AI Research Yann LeCun, is to “push the state of the art further.”

From chatbots to self-driving cars: what worries people about machine learning?

Following six days of structured public dialogue with participants of mixed socio-economic backgrounds, The Royal Society has published a report entitled “Machine learning: the power and promise of computers that learn by example”. The aim of the report is to understand the public’s awareness of Artificial Intelligence and Machine Learning, and delve deeper into the concerns and considerations of those who are less data-savvy.

Claire Craig and Jessica Montgomery from The
Royal Society argue that while participants see the advantages of AIs in areas such as healthcare and medicine, they fear the ‘de-personalisation’ of these services. As expected, the theme of Artificial Intelligence replacing jobs was a common occurrence during the discussions and perhaps surprisingly, only 9% of those questioned had heard of Machine Learning.

Farming (r)evolution

Many advances in electric self‑driving car technology and robotics are transferring across to industrial and commercial vehicles, which account for some 60% of the value of the overall electric vehicle market. Agriculture is one of the fields where exciting innovation happens due to the need to have higher yields and the rising cost of manual labour. Will agribots revolutionise this industry?

CMU’s plug-In “Synthetic Sensor” transforms any room into smart environment

Researchers at Carnegie Mellon University’s Human-Computer Interaction Institute (HCII) have taken the concept of ubiquitous sensors further than before with this amazing demo. Ubiquitous sensing in the home could mean a single sensor in each room rather than a host of smart appliances. “It can tell you not only if a towel dispenser is working, but can keep track of how many towels have been dispensed and even order a replacement roll when necessary,” The full paper for you techies can be found here.

Accelerating change at Mishcon: MDR LAB announces tech startup lineup

Mishcon de Reya has chosen the six early-stage to growth startups participating in their new technology incubator MDR LAB for the next ten weeks, in which the top-40 firm may ultimately invest. The tech incubator follows Mishcon’s 10-year strategy unveiled in July 2016, as part of which it plans to become more tech enabled.

Two of the selected companies apply Machine Learning techniques to the legal world. Surukam helps corporate legal teams automate their workflows and the decision-making process. Orbital Witness uses machine learning techniques to analyse the imagery in conjunction with existing real estate datasets to solve clients’ queries and innovate in real estate practice

Thursday May 25: Will robots replace HR? 

June 19-21: CogX: The AI Innovation Exchange

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