Innovation and Technology Weekly – No. 26, 2016

This is the online version of UNU-MERIT’s I&T Weekly which is sent out by email every Friday. If you wish to subscribe to this free service, please submit your email address in the box to the right.

Issue 26, 2016

This week's headlines:

The natural selection of bad science
September 21, 2016

There's no shortage of warnings from the scientific community that science as we know it is being drastically affected by the commercial and institutional pressure to publish papers in high-profile journals - and now a new simulation shows that deteroriation actually happening.

To draw attention to the way good scientists are pressured into publishing bad science, researchers from the University of California developed a computer model to simulate what happens when scientists compete for academic prestige and jobs.

In the model all the simulated lab groups they put in these scenarios were honest - they didn't intentionally cheat or fudge results. But they received greater rewards if they published 'novel' findings as happens in the real world. They also had to expend greater effort to be rigorous in their methods - which would improve the quality of their research, but lower their academic output.

Over time, effort decreased to its minimum value, and the rate of false discoveries skyrocketed, the researchers found. And what's more, the model suggests that the 'bad' scientists who take shortcuts in relation to the incentives on offer will end up passing on their methods to the next generation of scientists who work in their lab, creating in effect an evolutionary conundrum that the study authors call 'the natural selection of bad science'.

Full story: Science Alert / The Guardian / Royal Society Open Science Back to top

Microsoft says it will 'solve' cancer in the next 10 years
September 22, 2016

Microsoft has announced an ambitious plan to use computer science to 'solve' cancer within the next decade. While that plan involves many ambitious projects, one of the most interesting of proposals involves creating ultra-small DNA computers that can live inside a person’s body, monitoring for cancer cells and reprogramming them into healthy cells as soon as they pop up.

To make their goal a reality, Microsoft has gathered a team of biologists and computer scientists from around the world to work on various aspects of cancer research. The details are still thin on the ground, but one team plans on using machine learning and computer vision – where computers glean information from images or videos – to give radiologists a better understanding of how a specific patient’s tumour is progressing. This could open up a more nuanced type of personalised medicine. Another team is working on algorithms to predict the best plan of attack for each specific tumour type.

Then there's the group working on that 'moonshot' idea to make computers out of DNA that will monitor and reprogram cancer cells inside the body. The thinking is that, every time cancerous cells arise in the body, the computer would know, and 'reboot the system and clear out the diseased cells', according to Microsoft.

Despite these varied approaches, Microsoft says that all of the projects – no matter how different – follow two similar computer science approaches: information processing and machine learning. The team says that machine learning will allow researchers to better analyse millions and millions of files of biological data in search of new treatment approaches, a process that – until recently – has been done by hand. Machine learning could have the power to finally speed this up faster than ever thought possible.

Full story: Science Alert Back to top

New high-tech satellite system spots illegal fishing
September 16, 2016

With around 71% of the Earth's surface covered in water, trying to police what happens out on the oceans is no easy task. But that's where a new Google-powered satellite surveillance system called Global Fishing Watch (GFW) can help. Designed to act as an always-watching eye in the sky, the system can spot illegal fishing activities from space and you can even explore the GFW map online.

Alongside Google, conservation and activism groups Oceana and SkyTruth have contributed to the GFW project. It has the capability to track more than 35,000 commercial fishing boats across the oceans, and the system is adding 22 million data points each day. All of that data requires some serious number-crunching, which is where the cloud computing and machine learning expertise of Google comes in.

Advanced algorithms are used to match up the inputs captured in real-time and convert it into maps: the data stretches back to 2012, and has a three-day lag. The backers of GFW want governments, journalists, and citizens alike to have access to the data through any web browser, using it to identify potential problems and wrongdoing - zoom in far enough and you can actually identify individual vessels.

Part of the system uses public data from the Automatic Identification System (AIS), which combines information pulled from receivers on land and in orbit to track the movements of ships. Vessels can disable this tracking technology if they wish, but that itself attracts suspicion, and leads to problems when a ship wants to dock at port or refuel.

The Global Fishing Watch initiative has been in development for two years, and has already been used to highlight illegal tuna fishing off the coast of the Pacific Islands.

Full story: Science Alert Back to top

Germany to create world’s first highway code for driverless cars
September 21, 2016

This month, Germany’s transport minister, Alexander Dobrindt, proposed a bill to provide the first legal framework for autonomous vehicles. It would govern how such cars perform in collisions where lives might be lost. The laws attempt to deal with what some call the 'death valley' of autonomous vehicles: the grey area between semi-autonomous and fully driverless cars that could delay the driverless future.

Dobrindt wants three things: that a car always opts for property damage over personal injury; that it never distinguishes between humans based on categories such as age or race; and that if a human removes his or her hands from the steering wheel – to check email, say – the car’s manufacturer is liable if there is a collision.

Lack of clarity about who is responsible for the operation of such vehicles is a major point of confusion among manufacturers, consumers and lawyers. In the US, guidelines for companies testing driverless cars state that a human must keep their attention on the road at all times. This is also an assumption behind UK insurance for driverless cars, introduced earlier this year, which stipulates that a human 'be alert and monitoring the road' at every moment. But that is clearly not what many people have in mind when thinking of driverless cars.

Such confusion may be partly to blame for two fatal crashes involving Tesla cars this year. In the US, Joshua Brown was allegedly watching a DVD when his vehicle crashed in autopilot mode, killing him. And in China, Gao Yaning died when his car hit a road-sweeping vehicle. His family believe the vehicle was in autopilot mode at the time and is suing Tesla.

Dobrindt and others favour a 10-second rule, which requires a human to be sufficiently alert to take control of the vehicle within 10 seconds.

Full story: New Scientist Back to top

Teleportation step toward quantum internet
September 20, 2016

Physicists have set a new bar for quantum teleportation: moving information from one place to another without physically sending anything between the locations. Two separate teams managed to teleport information across several kilometres of optical fibre network in two cities.

Teleportation over long distances and across optical fibre networks is an important step towards the ultra-secure communications promised by quantum cryptography. And the set-ups could be seen as building blocks for a future 'quantum internet'.

In one of the papers, researchers from the University of Calgary describe how they teleported the quantum state of a photon, or light particle, over 8.2km in the Canadian city of Calgary. This teleportation process occurs via a phenomenon known as entanglement, which describes how sub-atomic particles can be linked even if they are separated by a large distance. In the other study researchers from the University of Science and Technology of China used a different set-up to achieve teleportation over a 30km optical fibre network in the city of Hefei.

In future, theoretical devices known as repeaters could help amplify signals, enabling communications - a quantum internet, if you like - over much bigger distances. This could enable the roll-out of much more secure communications offered by quantum cryptography. That would involve separate parties producing a shared and random secret key known only to them that could be used to encrypt and decrypt messages.

Full story: BBC News / Nature Photonics journal Back to top

Device detects human emotions using wireless signals
September 21, 2016

Researchers at MIT’s Computer Science and Artificial Intelligence Laboratory have created a device that can read human emotions using wireless signals. The EQ-Radio reads subtle changes in breathing and heart rhythms to figure out if a person is happy, excited, angry or sad.

The device measures heartbeats like an ECG monitor with a margin of error of 0.3%. It then analyses the waveforms within each heartbeat to determine the person’s emotion. The researchers believe that the device could be used in entertainment, studying consumer behaviour and in healthcare.

The EQ-Radio sends wireless signals that bounce off the person’s body. The device’s beat-extraction algorithms then break the signals into individual heartbeats. It then analyses the subtle changes in the heartbeat intervals to determine the person’s level of arousal and positive affect.

The device boasts of an accuracy rate of 87% in detecting human feelings. Since the device’s algorithm can capture the heartbeat in waveform, it shows promise in helping with non-invasive health monitoring and in diagnostic settings.

Full story: International Business Times Back to top