Humans have travelled the sea in some form for tens of thousands of years—with the earliest crossings occurring around 53,000 to 65,000 years ago (when Australo-Melanesian populations migrated into the Sahul landmass – known today as Australia and New Guinea – from what used to be the Sundaland peninsula.)

It’s often said how we know more about the moon than our oceans, with around 95 percent still unexplored. Arguably, the last major ocean research expedition was between 1872 and 1876 when a converted Royal Navy gunship known as the Challenger travelled close to 70,000 nautical miles and catalogued over 4,000 previously unknown species.

Inspired by the Challenger’s story, IBM has teamed up with non-profit ProMare to make a similarly large impact on ocean research.

The autonomous ship, Mayflower, is named after the ship which carried pilgrim settlers from Plymouth, England to Plymouth, Massachusetts in 1620. On its 400th anniversary, it was decided that a Mayflower for the 21st century should be built.

Brett Phaneuf, a Founding Board Member of ProMare and Co-Director of the Mayflower Autonomous Ship project, said:

“Putting a research ship to sea can cost tens of thousands of dollars or pounds a day and is limited by how much time people can spend onboard – a prohibitive factor for many of today’s marine scientific missions.

With this project, we are pioneering a cost-effective and flexible platform for gathering data that will help safeguard the health of the ocean and the industries it supports.”

Naturally, there are more than a few differences between the original ship and the Mayflower Autonomous Ship (MAS).

Mayflower 2.0 no longer relies solely on wind power and will use a wind/solar hybrid propulsion system with a backup diesel generator. The new ship also trades in a compass and nautical charts for navigation in favour of a state-of-the-art GNSS positioning system with SATCOM, RADAR, and LIDAR.

IBM’s deep learning technology is on-board to help the ship traverse the harsh and rapidly-changing environment of the ocean.

Donald Scott, Director of Engineering at Marine AI (which partnered with ProMare on the project), explained:

“In the middle of the ocean, communications are severely limited. Conditions can change very suddenly, and you don’t have the option to stop and power down.

With MAS, we needed to go beyond the existing technology for unmanned ships, creating a vessel that isn’t just operated remotely and doesn’t simply react to the environment, but learns and adapts independently.

To do this, we had to develop state-of-the-art capabilities around navigation, collision avoidance, communications and more.”

The training of AI models for the MAS began in October 2019. The actual hull for the ship arrived in Plymouth in March and sea trials began. Over the next few months, the ship was fitted with its advanced navigation and research equipment.

Andy Stanford-Clark, CTO of IBM UK & Ireland, added:

“IBM helped put man on the moon and is excited by the challenge of using advanced technologies to cross and research our deepest oceans.

By providing the brains for the Mayflower Autonomous Ship, we are pushing the boundaries of science and autonomous technologies to address critical environmental issues.”

MAS’ voyage couldn’t arrive at a more needed time with humans causing huge amounts of damage to the health of our oceans. A UN report found our oceans are now warmer, more polluted, more depleted, and more acidic than ever before.

Rising sea levels are among the key concerns about the impact on humans, but another is the increasing number of plastics in the sea which is simultaneously causing harm to sealife and ending up in the food we eat.

Professor Richard Thompson, OBE, Director of the Marine Institute, University of Plymouth, commented:

“Microplastics present a substantial challenge to our oceans. Over 700 species come into contact with marine litter which is found from the poles to the equator, and estimates are that the quantity of plastic in the oceans will triple in the decade to 2025.”

However, armed with the right data, it’s not too late to change course and heal our oceans.

MAS is fitted with a range of sensors including acoustic, nutrient, temperature, and water and air samplers. Edge devices will store and analyse all data locally until connectivity is available. When a link has been established, the data will be uploaded to edge nodes onshore.

Unless there are any last-minute delays, MAS is set to depart on its voyage this month. The ship is due to arrive in Plymouth, Massachusetts around two weeks later. Where required, updated deep learning models can be pushed out to the ship.

MAS’ virtual crew will be based in Plymouth, UK but IBM says millions of virtual “pilgrims” will be able to experience the voyage online.

The post A(I)hoy, mateys: IBM’s crewless ocean research ship to launch ‘very soon’ appeared first on AI News.

Scientists around the world are doing incredible work to increase our understanding of COVID-19. Thanks to their findings, existing medications have been discovered to increase the likelihood of surviving the virus.

Unfortunately, there are still fatalities. People with weakened immune systems or underlying conditions are most at risk, but it’s a dangerous myth that the young and otherwise healthy can’t die from this virus.

According to a paper published in science journal Nature, around 6.5 percent of COVID-19 cases have a “worrying trend of sudden progression to critical illness”. Of those cases, there’s a mortality rate of 49 percent.

In the aforementioned paper, the researchers wrote: “Since early intervention is associated with improved prognosis, the ability to identify patients that are most at risk of developing severe disease upon admission will ensure that these patients receive appropriate care as soon as possible.”

While most countries appear to be reaching the end of the first wave of COVID-19, the possibility of a second threatens. Many experts forecast another wave will hit during the winter months; when hospitals already struggle from seasonal viruses.

One of the biggest challenges with COVID-19 is triaging patients to decide who are most at risk and require more resources allocated to their care. During the peak of the outbreak in Italy, doctors reported reaching a point of having to make heartbreaking decisions over whether it was a waste of limited resources even trying to save someone.

A team led by China’s senior medical advisor on COVID-19, Zhong Nanshan, was established in February. The team consisted of researchers from Tencent AI Lab in addition to Chinese public health scientists.

Nanshan’s team set out to build a deep learning-based system which can predict whether a patient is likely to become a critical case. Such information would be invaluable to ensuring the patient gets early intervention to improve their chances of surviving the virus in addition to supporting medical staff with their triaging decisions.

The deep learning model was trained on data from 1590 patients from 575 medical centers across China, with further validation from 1393 patients.

Tencent has made the COVID-19 tool for predicting critical COVID-19 cases available online here (Please note the small print which currently says “this tool is for research purpose and not approved for clinical use.”)

(Photo by Ashkan Forouzani on Unsplash)

The post Deep learning is being used to predict critical COVID-19 cases appeared first on AI News.

Last month, Facebook refused to remove a deepfake video of House Speaker Nancy Pelosi. Rather than making Pelosi appear to say things she never did, the video aimed to portray her as being intoxicated.

Deepfakes have the potential to spread misinformation and damage the reputation of individuals. Particularly in the world of politics and increasingly sophisticated state disinformation campaigns, it’s easy to imagine the danger they pose.

Pelosi later told California’s KQED: “I think they have proven — by not taking down something they know is false — that they were willing enablers of the Russian interference in our election.”

Governments and activists have called on social networks to assist in the detection and deletion of such videos. Now it seems that campaigners are targeting social network executives with deepfakes in a bid to make them take such videos more seriously.

In the deepfake of Zuckerberg, the Facebook CEO is portrayed to say: “Imagine this for a second: One man, with total control of billions of people’s stolen data, all their secrets, their lives, their futures.”

Fortunately for Zuckerberg, the deepfake is not malicious and most people will know it’s not real. The video both helps to make social networks assist more in tackling deepfakes while also raising public awareness.

Society is going through a period of serious change when it comes to knowing what’s real and fake. Many now question what they read because anyone could have written it and the issue of fake news is well-publicised. Most people, however, are still accustomed to believing that someone is saying what they are when they can see them doing it.

Zuckerberg’s deepfake was created Israeli startup Canny AI. The firm has also debuted fake videos with the likes of President Trump and Kim Kardashian as part of a commissioned art installation called Spectre that was on display at the Sheffield Doc/Fest in the UK.

deepgeniusai.com/">AI & Big Data Expo events with upcoming shows in Silicon Valley, London, and Amsterdam to learn more. Co-located with the IoT Tech Expo, , & .

The post Zuckerberg is deepfaked a month after Facebook refused to remove others appeared first on AI News.

Some pioneering students from Stanford University exposed a neural network to over 400,000 memes in a bid to teach it the nuances of modern humour.

Armed with this information, the AI attempted the artistic endeavour of creating its very own.

In the ‘Dank Learning’ paper, the students detail how they applied deep learning methods to create memes. Ideas are passed between different layers of the system until images similar to what it learned from are created.

Now, I’m not the best judge of memes, but I had a chuckle at a few of its creations:

In order to see how people respond, a group of volunteers were given random memes created by the AI and humans.

The researchers observed “the average meme produced is difficult to differentiate from a real meme.”

Volunteers participating in the study were also asked to rank the memes in order of hilarity. They found “variants scored close to the same hilarity rating as real memes, though this is a fairly subjective metric.”

Microsoft and Google may be making you no longer trust who you’re talking to on the phone, but Stanford is now making you question who — or rather, what — made the image you’re laughing at.

The full Dank Learning paper can be found here (PDF)

Do you think Stanford University’s project is dank or stank?

The post Stanford University students employ AI to make the dankest memes appeared first on AI News.

Earlier this year, the Department of Justice in the U.S. ordered the domain host for a site which organised an anti-Trump demonstration to hand over details of those who even visited the page.

This incident alone goes to show how valuable profiling technology such as this would be to some governments. The paper even notes the danger of such a system being used by governments for tracking down citizens.

Politics aside, the AI’s technology is groundbreaking. With due oversight, it could improve public safety. Everyone has the right to protest and should not be profiled for doing so peacefully, but those which incite or take part in violence could find themselves prosecuted and/or banned from any future demonstrations.

The AI is trained with over 4,000 images across two datasets. Half of the images have simple backgrounds while the rest are more complex such as featuring multiple people. The deep learning system then maps 14 points on a face and measures the distance between them. With this data, the AI can recognise people even if they’re wearing accessories such as a hat, scarf, or sunglasses.

Credit: University of Cambridge / National Institute of Technology / Indian Institute of Science

When disguised, the AI system is able to identify people with a 56 percent success rate. This drops to 43 percent when the subject is wearing glasses – perhaps helping to explain why no-one can recognise Clark Kent as Superman when he wears his.

For now, your best shot at remaining anonymous is still the nefarious get-up of a balaclava and shades. Perhaps not for much longer, however.

You can find the full paper here.

Are you concerned by AIs which profile citizens? Share your thoughts in the comments.

The post AI identifies protesters – even when they’re disguised appeared first on AI News.

Based on the comments made by Elon Musk and Vladimir Putin in our article yesterday, you’d be forgiven if this raises some concerns. AI like Prowler, however, could be what saves us from destruction.

If you missed it, Musk voiced his concern about Putin’s comment that the nation which leads in AI “will become the ruler of the world.”

Some are concerned about an AI arms race and that, without human input, an AI relying solely on logic could decide on things such as launching preemptive strikes as being the most likely to achieve victory or protect its host nation. This decision would be devoid of empathy for the human, environmental, political, and long-term destabilising catastrophes this would create.

Mark Cuban, a billionaire and possible U.S. president runner in 2020, shares a concern about killer AI:

Autonomous weaponry is the ultimate threat to humanity. https://t.co/C0EEfwOCDw

— Mark Cuban (@mcuban) September 4, 2017

Cuban, however, points out it’s AIs not taking on human-like qualities which pose the biggest threat:

I’m concerned about machines NOT taking on human like qualities https://t.co/1pG75agllk

— Mark Cuban (@mcuban) September 4, 2017

Most AI development today focuses on deep neural networks using vast amounts of data. This approach is vital for problem-solving as it is very effective, but for decision-making, Prowler argues it is limiting.

Building a decision-making AI

Prowler is building a decision-making AI which is based on probabilistic modelling, reinforcement learning, and game theory. For each of those areas, Prowler has experts in their field combining their knowledge to create an AI it hopes will make decisions as well as a human.

The company was co-founded by two alums from AI company VocalIQ, which was acquired 13 months after launch by Apple. Today, Cambridge Innovation Capital (CIC) announced it led a £10 million Series A funding round for Prowler.

“Prowler has assembled a world-class team of researchers to tackle some of the most intractable problems of our age,” comments Andrew Williamson, Investment Director at CIC, who is joining the Board of Prowler. “It is hugely exciting that the company is able to capitalise on the expertise in probabilistic modelling, principled machine learning and game theory available in Cambridge.”

“This investment allows us to expand our world-leading team of academics and developers, enhancing our research bandwidth and accelerating our technology into the market,” added Vishal Chatrath, CEO of Prowler. “As a team, we will use the funding to take the business to the next stage and we will continue to solve some of the world’s hardest machine learning problems.”

Despite popular belief, the closer we get to AI taking on human-like qualities could save us all. At the least, it will be far more efficient.

What are your thoughts about AI making human-like decisions?

The post Prowler.io aspires to build AI which makes human-like decisions appeared first on AI News.

The course on Coursera is simply called ‘Deep Learning Specialization’ and promises to teach students vital skills for the future. Some of the advanced topics covered include neural networks, backpropagation, convolutional networks, recurrent networks, computer vision, natural language processing, and more.

“Just as electricity transformed every major industry starting about 100 years ago, AI is now poised to do the same,” observes Ng. “Several large tech companies have already built divisions dedicated to AI, and have made huge strides transforming themselves with AI.”

For AI to flourish and make a notable difference to everyday lives; it will take a workforce trained with entirely new skills. 1.8 million people enrolled in Ng’s prior Machine Learning course since 2011. Ng’s courses are putting students on the path to accomplishing exciting new breakthroughs.

“I hope we can build an AI-powered future that provides everyone affordable healthcare, accessible education, inexpensive and convenient transportation, and a chance for meaningful work for every man and woman,” says Ng. “An AI-powered future that improves every person’s life.”

Confusion about AI

The course is part of a bigger project from Ng which aims to disseminate AI knowledge and can be found here. He’s observed many people are confused about the various specialities and lingo within it.

Part of the confusion is due to comments by people considered leaders in technology. People like Elon Musk have voiced concerns about AI posing a danger to humanity. Ng, meanwhile, says there seems to be a negative correlation “between how much someone knows AI and how afraid they are of it.”

When you earn a Deep Learning Specialization certificate, Ng claims you will be able to confidently put ‘Deep Learning’ onto your resume. That is something which is sure to offer a lucrative in the coming years.

Picture credit: “Baidu Chief Scientist Andrew Ng”, by “NVIDIA”, used under CC BY-NC-ND 2.0

What are your thoughts on Ng’s launch of his deep learning course?

To learn more about AI for Developers register for your free pass to the AI Expo North America conference (29-30th November 2017) in Santa Clara, CA today!

The post AI guru Andrew Ng launches an online deep learning course appeared first on AI News.

IBM has already cemented its place as a leader in AI with Watson and its deep learning capabilities. Deep learning is a subset of AI which mimics the human brain and is therefore a focus of researchers.

Companies including Microsoft, Facebook, Amazon, and Google have made advancements in deep learning, but IBM’s could be the most groundbreaking. Today’s deep learning systems can take days to analyse data, but IBM has managed to reduce this to just hours.

IBM has achieved this by overcoming a core hurdle of deep learning systems. Most current projects are run on a single server due to how much data is involved and the difficulty of keeping it all synchronised. New software developed by IBM, however, claims to split deep learning tasks across 64 servers running up to 256 processors.

This is available to customers using IBM Power System servers or anyone interested in testing its capabilities. For its tests, IBM used 64 of their Power 8 servers which feature Intel microprocessors and NVIDIA GPUs linked with an NVLink connection. This configuration allows for rapid data flow between the chips.

Using this method, IBM claims performance can be boosted 50-60 times over a single server typically used today. However, it requires some complex work under-the-hood to take full advantage of the extra servers and processing power.

Managing traffic is the main priority to ensure all the processors are working and not sitting idle. Each processor has a dataset to work on but also requires data from other processors to understand the complete picture. Optimising this management of traffic is an ongoing process to achieve peak levels of performance.

A record breaker

In an ideal scenario, each processor added would deliver a 100 percent improvement to performance, but this is unrealistic. IBM claims its system achieved 95 percent across 256 processors, which is both impressive and sets a new record. The previous record was held by Facebook who achieved 89 percent.

This isn’t the only record IBM has claimed to have broken with their latest advancement. For image recognition, IBM says it observed a 33.8 percent accuracy rate analysing 7.5 million images over seven hours. The prior record was held by Microsoft with 29.8 percent accuracy, and that effort took 10 days to complete.

Are you impressed with IBM’s machine learning advancement? Share your thoughts in the comments.

To learn more about AI in the Enterprise register for your pass to the AI Exhibition and Conference this fall in Santa Clara, CA today!

The post IBM breaks records with deep learning advancement appeared first on AI News.