The QWERTY typewriter was introduced in 1872, and since then tapping on a keyboard or screen has become the standard way to interact with digital technology. But this isn't always convenient or safe, so new "touchless" ways to control machines are being developed. Imagine being out for a jog, headphones on, and wanting to turn up the volume without breaking your stride. Or receiving a "new message" alert on your phone while driving and wanting to activate the text-to-speech function without taking your eye off the road. These are scenarios where touchless control would come in handy.

Recent advancements in radio frequency machine learning (RFML) have demonstrated the use of raw in-phase and quadrature (IQ) samples for multiple spectrum sensing tasks. Yet, deep learning techniques have been shown, in other applications, to be vulnerable to adversarial machine learning (ML) techniques, which seek to craft small perturbations that are added to the input to cause a misclassification. The current work differentiates the threats that adversarial ML poses to RFML systems based on where the attack is executed from: direct access to classifier input, synchronously transmitted over the air (OTA), or asynchronously transmitted from a separate device. Additionally, the current work develops a methodology for evaluating adversarial success in the context of wireless communications, where the primary metric of interest is bit error rate and not human perception, as is the case in image recognition. The methodology is demonstrated using the well known Fast Gradient Sign Method to evaluate the vulnerabilities of raw IQ based Automatic Modulation Classification and concludes RFML is vulnerable to adversarial examples, even in OTA attacks. However, RFML domain specific receiver effects, which would be encountered in an OTA attack, can present significant impairments to adversarial evasion.

Truth may or may not be stranger than fiction, but it certainly can be more dramatically compelling. A documentary on the mission that took astronauts Neil Armstrong, Buzz Aldrin and Michael Collins to the moon and back half a century ago, "Apollo 11" has no talking heads telling us what it all means or modern re-creations like the unemotional Ryan Gosling-starring "First Man." Instead, as directed by Todd Douglas Miller, "Apollo 11" relies on footage shot and audio recorded at the time, and the results, from liftoff to landing to the journey back home, are completely riveting. "Apollo 11" succeeds as well as it does for several reasons, but a key one was the discovery in the National Archives of hours of large-format 65-millimeter color footage covering all aspects of the mission, footage that had never been seen by the public and was subsequently digitized to the highest resolution possible. Also discovered were 11,000 hours of audio recordings from key personnel on the ground as well as the astronauts way out in space. But just having great material does not guarantee a superior film, which is where Miller's skill and experience Miller come in.

Say you have a 10x10 black and white image where every pixel specifies whether the space is traversable. A maze is solvable if you can travel from the centre to the edge. Can a feedforward convolutional network distinguish between solvable and insolvable mazes? How would this scale with maze size and number of layers? Do you know any research papers on this?

Diana Greene, former CEO of Google Cloud and co-founder of VMWare, joins CNBC's Becky Quick at the MIT Schwarzman College Of Computing …

Artificial Intelligence: what is it? How will it impact our lives? How will it be used? Will it actually take over? As the field emerges and becomes more widely discussed, social media's role in shaping public debate answering these questions and more has never been more apparent, with more of us turning to social media for news.

If Netflix's new Dating Around reality series is a touch too maudlin for your taste, then you'll probably dig its upcoming French original Osmosis. The first trailer for the technophobic show, which premieres March 29th, just landed and it's a literal head trip. Set in a near future, it follows several young Parisians who sign up to a beta program for an experimental dating technology designed to pair up soul mates. The clip shows the eager participants pop a pill like Neo in The Matrix (which turns 20 next month). But instead of taking back their life from a machine-dominated simulation, it puts their romantic fate in the hands of an AI named Martin.

"We are on the extreme promontory of the centuries! What is the use of looking behind at the moment when we must open the mysterious shutters of the impossible? Time and Space died yesterday. We are already living in the absolute, since we have already created eternal, omnipresent speed." In February 1909, those words from The Futurist Manifesto appeared in the Italian newspaper La Gazzetta dell'Emilia.

When you enter Star Wars: Galaxy's Edge, the 14-acre expansion coming to Disneyland early this summer, you are faced with a choice. Walk around a bend -- and under an archway crafted to look centuries old -- to discover the starship the Millennium Falcon, nestled comfortably under hand-sculpted mountains designed to evoke the petrified forests of New Mexico. Or wander into a marketplace, one inspired by Moroccan and Turkish bazaars. Intergalactic creatures are said to live in the ramshackle, factory-like apartments above the shops, here presented as stalls, creating a cacophony of life and noise. Consider this the "Star Wars" equivalent of Main Street, U.S.A, but instead of quaint stores there are mysterious cat-like creatures in cages and toys that feel patched together from found parts. If you bypass the town you'll enter a forest where the Resistance, the "good guys" in the "Star Wars" universe, have set up a camp, hiding ships among shrubbery and building a base inside alien ruins -- a twisting cave where digital schematics clash with remnants of a long-lost civilization.

In 1985, the TV film Max Headroom: 20 Minutes into the Future presented a science fictional cyberpunk world where an evil media company tried to create an artificial intelligence based on a reporter's brain to generate content to fill airtime. There were somewhat unintended results. Replace "reporter" with "redditors," "evil media company" with "well meaning artificial intelligence researchers," and "airtime" with "a very concerned blog post," and you've got what Ars reported about last week: Generative Pre-trained Transformer-2 (GPT-2), a Franken-creation from researchers at the non-profit research organization OpenAI. Unlike some earlier text-generation systems based on a statistical analysis of text (like those using Markov chains), GPT-2 is a text-generating bot based on a model with 1.5 billion parameters. With or without guidance, GPT-2 can create blocks of text that look like they were written by humans.