Anna Clark

Google Overhauls Search Algorithm in Bid to Fight Fake News

The Telegraph, 4/25/2017

Google is changing their search algorithm to better distinguish websites that contain fake news, conspiracy theories, or extreme views from more reputable sources or websites. The search engine’s Autosuggest and Direct Answer features will also be modified, to allow users to report offensive suggestions or false information (respectively). This comes after months of growing public pressure for Google to make changes to combat the spread of misinformation on the Internet.

Machine Learning Shows Exactly When to Zap Brain to Boost Memory

New Scientist, 4/20/2017

Michael Kahana and other researchers at the University of Pennsylvania have used machine learning to analyze data on brain function and brain wave patterns. Electrodes implanted in the subjects’ brains measured brain activity while the subjects attempted to memorize and recall information. The electrodes could also transmit electric shocks to the brain. Results showed that carefully timed shocks made people 13 percent more likely to recall material.

Google Uses Neural Networks to Translate Without Transcribing

New Scientist, 4/4/2017

Researchers at Google Brain are developing a method to translate speech in one language to text in a different language, using neural networks. They hope to improve over more conventional automatic methods, where the speech is transcribed into written text, and then the written text is translated. The older method can be cumbersome, and initial experiments show that direct speech-to-text translation seems less subject to error. The new method could especially help speakers of rare languages communicate with others around the globe.

As Congress Repeals Internet Privacy Rules, Putting Your Options in Perspective

National Public Radio, 3/28/2017

Rules made in October by the Federal Commercial Commission (FCC) are expected to be overturned by Congress and President Trump. These rules would have regulated how Internet Service Providers (ISPs) collected and used data from users, by giving users more control over what information ISPs collect. This rule set would not apply to websites and app providers, like Facebook or Google. However, critics of the overturn argue that it is much more difficult for users to choose to avoid ISPs if they do not wish to be tracked.

Skin Powered by the Sun? Energy-Saving Prosthetic Limbs Get Better Feeling

Reuters, 3/22/2017

Researchers from the University of Glasgow, UK, have developed a prototype for solar-powered prosthetic skin. The prosthesis is wrapped in a thin layer of carbon, which allows light to pass through the skin and be collected as energy. Thus far, the prototype uses this energy to power additional sensors in the prosthetic, giving the skin a heightened sensitivity to pressure and texture as well as temperature.

Machine Learning Reveals Lack of Female Screen Time in Top Films

New Scientist, 3/8/2017

Shri Narayanan, from the University of Southern California, has developed a new program that runs facial recognition software and voice recognition software simultaneously to analyze gender bias in high-grossing box-office films. This program has analyzed 300 films, and the study shows that women are underrepresented on the big screen. Narayanan is also using machine learning and natural language processing to evaluate film scripts, to explore further biases.

Bridging a Digital Divide That Leaves Schoolchildren Behind

The New York Times, 2/22/2017

The federal government is attempting to bridge the growing digital divide in low-income families by expanding a subsidy program, Lifeline, to include broadband Internet access. The Federal Communications Commission’s main goal in revising Lifeline is to address the increasing number of students without online access needed to complete schoolwork and homework. Other methods are being used to in an effort to patch this divide, such as wifi-equipped buses and school-provided wireless hotspots.

AI Predicts Autism From Infant Brain Scans

IEEE Spectrum, 2/15/2017

Researchers at the University of North Carolina, Chapel Hill have applied deep-learning algorithms to brain scans of children with a high risk of autism. Algorithms used three indicators, brain surface area, volume, and the gender of the child, to determine if 6- to 12-month-old infants were likely to develop autism. The results were 81% accurate at predicting later diagnosis. This improves over a 50% prediction rate from behavioral questionnaires.

DACC and Virgin Galactic Team Up to Explore Virtual Reality

Las Cruces Sun-News, 2/10/2017

Doña Ana Community College (DACC) and Virgin Galactic are teaming up to create a new educational outreach program. The partnership focuses on new virtual reality (VR) technologies and explores how they might be used in various settings and applications. It will begin with students using a VR simulation to learn concepts in aerospace technology.

Toyota Funds AI Research to Build Autonomous Cars

Network World, 2/7/2017

Toyota is investing 50 million dollars and partnering with Stanford University and the Massachusetts Institute of Technology (MIT) to develop artificial intelligence and robotics technology needed for advanced driver-assistance systems. While Toyota is not trying to develop a fully automated car, they note that incremental advances in such driver-aid systems could eventually lead to driverless smart cars.

New Algorithms by University of Toronto Researchers May Revolutionize Drug Discoveries

University of Toronto News, 2/6/2017

University of Toronto researchers have developed algorithms which can efficiently generate an accurate 3D image of a protein molecule from several 2D images in just a few minutes. This advance has far-reaching implications for the medical field. For example, drug researchers will be able to use these 3D protein models to analyze the structure of disease-specific proteins and predict the way experimental medications will bind to those proteins inside the body.

Private Medical Data Is For Sale – and It’s Driving a Business Worth Billions

The Guardian, 1/10/017

Private medical data is a multi-million dollar industry that is growing rapidly, according to Adam Tanner at Harvard’s Institute for Quantitative Social Science. When medical data is initially sold to big data miners, it may be referred to only by unidentifiable numbers. However, data miners can re-identify patients by cross-referencing the medical data with data collected from other sources.