Month: September 2017

Enlarge (credit: Ivan Etimov et al. / Thinkstock)

Progress in the field of machine vision is one of the most important factors in the rise of the self-driving car. An autonomous vehicle has to be able to sense its environment and react appropriately. Free space has to be calculated, solid objects avoided, and any and all of the instructions we helpfully leave everywhere—painted on the tarmac or posted on signs—have to be obeyed.

Deep neural networks turned out to be pretty good at classifying images, but it's still worth remembering that the process is quite unlike the way humans identify images, even if the end results are fairly similar. I was reminded of that once again this morning when reading about a method of spoofing road signs. It's a technique that just looks like street art to you or me, but it completely changes the meaning of a stop sign to the machine reading it.

No actual self-driving cars were harmed in this study

First off, it's important to note that the paper is a proof-of-concept; no actual automotive-grade machine vision systems were used in the test. Covering your local stop signs in strips of black and white tape is not going to lead to a sudden spate of car crashes today. Ivan Evtimov—a grad student at the University of Washington—and some colleagues first trained a deep neural network to recognize different US road signs. Then, they created an algorithm that generated changes to the signs that human eyes find innocuous, but which changed the meaning when a sign was read by the AI classifier they just trained.

IFA (the Internationale Funkausstellung Berlin) is happening in Germany right now, and the show is apparently the coming out party for third-party speakers with an integrated Google Assistant. Imagine a Google Home, but made by somebody else, without the all-white, air freshener design. The companies are in various stages of rolling out information, ship dates, and pricing, but we have a few details and lots of pretty pictures.

Sony

Let's meet our contestants:

Enlarge (credit: Liz West)

The Windows 10 Fall Creators Update now has a release date: October 17. Microsoft started finalizing the release last week, and we'd expect this release to follow the pattern seen in previous Windows updates: the final build will be done some time in September and roll out to members of the Windows Insider program's fast, slow, and release preview rings. Then it will hit Windows Update. From there, we'd expect a slow ramp up in availability.

The biggest feature of the update is that it opens up support for virtual reality headsets with what Microsoft calls its Mixed Reality platform. Headsets from Acer, Dell, Lenovo, and HP will start shipping on the same day (and will be joined by Asus' headset next year), with prices starting at $299. These headsets don't need any external tracking boxes; they'll just plug into your PC with an HDMI and USB cable, or in Dell's case, a single USB Type-C cable. In addition to software written for Microsoft's Mixed Reality API, they will also support the SteamVR platform, so a library of suitable games should materialize soon.

Other features of the new version are seamless OneDrive integration, the ability to pin contacts to your taskbar, a beta version of eye-tracking-based navigation, expanded integrated exploit mitigation systems, and a new searchable emoji keyboard.