Privacy fears after woman says Alexa recorded a private conversation and sent it to a random contact

• Portland woman discovered her Alexa recorded and sent a private conversation
• It was only until a contact let her know that she became aware of the flaw
• Amazon says it’s aware of the issue and in the process of releasing a fix
• Issue raised the ire of privacy advocates who worry Alexa is spying on its users

Be careful of what you say around your Echo devices.

A Portland woman was shocked to discover that Echo recorded and sent audio of a private conversation to one of their contacts without their knowledge, according to KIRO 7.

The woman, who is only identified as Danielle, said her family had installed the popular voice-activated speakers throughout their home.

It wasn’t until a random contact called to let them know that he’d received a call from Alexa that they realized their device had mistakenly tra

As unlikely as this string of events is, we are evaluating options to make this case even less likely’, they added.

Amazon offered to ‘de-provision’ the communication features of the woman’s Echo device so that she could continue using the smart home features.

But she said she’d much rather get a refund and says she’s dumped all her Alexa devices from her home.

‘A husband and wife in the privacy of their home have conversations that they’re not expecting to be sent to someone [in] their address book,’ she told KIRO.

A number of privacy concerns have been raised in recent months around Amazon’s Alexa-enabled speakers.

The contact, who was one of her husband’s work employees, told the woman to ‘unplug your Alexa devices right now. You’re being hacked’, KIRO noted.

‘We unplugged all of them and he proceeded to tell us that he had received audio files of recordings from inside our house,’ Danielle said.

‘…I felt invaded. Immediately I said, “I’m never plugging that device in again, because I can’t trust it”‘, she added.

Thankfully, the recorded conversation was only about hardware floors, but the incident has still managed to spark fears of Alexa spying on its users.

A step towards limitless energy: ‘Ghostly’ lightning waves found inside a nuclear reactor could control electrons and make fusion power a reality

• Department of Energy’s Oak Ridge National Laboratory made the findings
• Whistlers waves, created by lightning, are found whipping across the ionosphere
• They travel across the globe at heights of 620 miles (1,000km) above the planet
• Scientists say they have detected the mysterious waves inside their reactor 
• They could be used to help prevent runaway electrons in the fusion process
• A breakthrough in fusion reaction technology and limitless energy could result

‘Ghostly’ whistler waves created by lightning may help to protect experimental nuclear fusion reactors and make limitless energy a reality.

That’s the claim made by a group of researchers who detected the ghostly presence of the waves inside their reactor, in what they say is a world first.

Whistlers are normally found whipping across the ionospheric layer of the Earth’s atmosphere at heights of up to 620 miles (1,000km) above the planet.

Now, scientists say they have detected the mysterious electromagnetic waves inside their reactor.

They were surprised to find that whistlers appeared to help prevent runaway electrons, produced in the nuclear fusion process.

These particles, travelling at ever faster speeds, risk burning holes in the equipment if they accelerate too fast.

Experts hope to tame this by developing methods of controls based on their findings, which could lead to a breakthrough in fusion reaction.

Scroll down for video 

+4

Ethereal waves known as whistlers created by lightning may help to protect experimental nuclear fusion reactors from damage. That’s the claim made by a group of researchers who detected the ghostly presence of whistler waves inside their reactor (pictured)

+4

Whistlers are normally found whipping across the ionospheric layer of the Earth’s atmosphere at heights of up to 620 miles above the planet. Now, scientists say they have detected the mysterious electromagnetic waves inside their tokamak reactor (graphical representation)

The findings were made by a research team led by the US Department of Energy’s (DOE) Oak Ridge National Laboratory (ORNL).

For decades, experts have been building the scientific basis for nuclear fusion as an energy source to generate limitless electricity.

Fusion happens when the nuclei of lighter atoms join under extraordinarily high temperatures to create a heavier nucleus, releasing energy.

The hot, fused atoms form a plasma that is confined by high magnetic fields in an experimental vessel called a tokamak – designed to withstand conditions hotter than the sun.

Strong magnetic fields are used to keep the plasma away from the reactor’s walls, so that it doesn’t cool down and lose its energy potential.

Instabilities can occur in such an extreme environment, causing a dramatic quenching, or cool down, of the plasma and producing runaway electrons that could veer off and burn holes in the tokamak’s interior wall.

Using sophisticated technologies, the team made direct measurements of whistler waves produced when a laboratory plasma becomes unstable and generates runaway electrons.

By better understanding this process, they hope they can reverse engineer it, and create whistler generating antennas to prevent electrons from getting too fast.

Runaway electrons also occur in nature. They are energised when lightning strikes or solar substorms disrupt the plasma environment of the Earth’s ionosphere, which is the atmosphere’s ionised upper layer.

Scientists have theorised that whistler waves regulate space weather and may help mitigate the damaging effects of energetic electrons on satellites orbiting the Earth.

‘We have known that runaway electrons drive whistler waves in the ionosphere during natural events, which led to theories that runaways would also drive similar electromagnetic waves in a tokamak plasma, said ORNL’s Don Spong who led the study.

+4

For decades, scientists have been building the scientific basis for nuclear fusion as an energy source to generate electricity. This artist’s impression shows what fusion energy generation could look like inside a tokamak reactor

‘Observing whistler waves helps us better understand their underlying physical mechanisms, which could open an avenue to develop new techniques to control runaway electrons and keep them from potentially damaging fusion reactors.

‘As we learn more about the characteristics and excitation of whistler waves in tokamaks, we may be able to mimic similar behaviour to protect plasma-facing components.’

The team’s experiments took place at the DIII-D National Fusion Facility, a DOE user facility, which is operated by General Atomics in San Diego, California.

The full findings were published in the journal Physical Review Letters.

HOW DOES A NUCLEAR FUSION REACTOR WORK?

Fusion is the process by which a gas is heated up and separated into its constituent ions and electrons.

It involves light elements, such as hydrogen, smashing together to form heavier elements, such as helium.

For fusion to occur, hydrogen atoms are placed under high heat and pressure until they fuse together.

When deuterium and tritium nuclei – which can be found in hydrogen – fuse, they form a helium nucleus, a neutron and a lot of energy.

+4

In a fusion reactor, strong magnetic fields are used to keep plasma – a gaseous soup of subatomic particles – away from the reactor’s walls, so that it doesn’t cool down and lose its energy potential. This graphic shows Tokamak Energy’s experimental design

This is done by heating the fuel to temperatures in excess of 150 million°C and forming a hot plasma, a gaseous soup of subatomic particles.

Strong magnetic fields are used to keep the plasma away from the reactor’s walls, so that it doesn’t cool down and lose its energy potential.

These fields are produced by superconducting coils surrounding the vessel and by an electrical current driven through the plasma.

For energy production, plasma has to be confined for a sufficiently long period for fusion to occur.

When ions get hot enough, they can overcome their mutual repulsion and collide, fusing together.

When this happens, they release around one million times more energy than a chemical reaction and three to four times more than a conventional nuclear fission reactor.

Incredible video from inside a fusion reactor

http://imasdk.googleapis.com/js/core/bridge3.211.3_en.html#goog_698424228

Real-time Google lens features, redesign begin rolling out

Update includes new real-world copy and paste features

Google started rolling out the real-time Google Lens features it showed off at Google I/O earlier this month. Now, users can just point their phone cameras at the world. The app proactively searches for information and places small coloured dots on recognized objects. Tapping the dots load up information about the object.

A combination of on-device machine learning and the use of Google’s new cloud TPUs to identify things quickly. The smart text selection feature is also live. Users can now highlight text on objects in the real world as well as copy it and paste it. Additionally, the similar objects feature is live as well. Google Lens is capable of displaying objects that look similar to ones you’ve scanned with your phone. Along with the new features, Google Lens is sporting a new look. A round, white card now resides at the bottom of the screen.

Users can drag it up to see a list of features. When you select a recognized object, the card also pops up with the information. The redesign also moved the microphone icon to the right side of the screen. It also removed the ‘Remember this,’ ‘Important to keep’ and ‘Share’ suggestion bubbles.

Unfortunately the new features appear to be pushed out server-side. There is no app update or APK file to download that brings the new changes. As of yet, no one at MobileSyrup has received the update. It doesn’t appear to be device specific. Some users on Reddit confirmed they received the update on the Galaxy S8 and the OnePlus 5T. However, my Pixel 2 XL running the Android P beta has not received the update.

MRI scans suggest transgender people’s brains resemble their identified gender: study

The brains of transgender individuals share characteristics with those of the gender they identify with, according to new research.

Researchers used MRI scans to identify how adolescents’ brains responded to a pheromone that men and women are known to react to differently.

The brains of transgender people who identified as women reacted more like female brains, and transgender people who identified as men had brains that responded more like males than their biological sex.

The researchers, who presented their findings Tuesday at the European Society of Endocrinology’s annual symposium, focused on the brains of adolescents who had gender dysphoria. People with gender dysphoria have a conflict between their physical or assigned gender and the gender they identify with, according to the American Psychiatric Association. This can cause them considerable distress.

READ MORE: Is the world more accepting of transgender people? Yes, but many people still aren’t, says Ipsos poll

There are sex differences in the brain at the structural level and also how male and female brains perform certain tasks, said neuroscientist Julie Bakker of the University of Liege in Belgium via email. Bakker’s research “found that adolescents with gender dysphoria had brain activity patterns very similar to their desired/experienced gender,” she wrote.

“At the moment, most available evidence suggests that it is a developmental effect, taking place before birth, but of course, we cannot rule out any effects of sex hormones later in life.”

Bakker’s study was small: looking at only about 150 individuals. As such, its findings should be interpreted with caution. Doug VanderLaan, an assistant professor of psychology at the University of Toronto, is currently working on a similar study at a larger scale, which has yet to publish results.

“This research area is still very much in its early days,” he said. “There have been relatively few studies and the methods have not been consistent. Consequently, there are few findings regarding specific brain areas that have been shown to be reliable and more research is needed.”

READ MORE: Transgender man gives birth to baby boy, says he is ‘overwhelmed with love’

However, he said, across studies so far, it has generally been the case that the brains of transgender people share certain resemblances to those of their identified gender.

“Overall, it seems there is merit to the idea that certain aspects of the brains of transgender people align with their experienced gender identity.”

Bakker suggests that her research could be used to inform how young people with gender dysphoria are treated. “Although more research is needed, we now have evidence that sexual differentiation of the brain differs in young people with GD (gender dysphoria), as they show functional brain characteristics that are typical of their desired gender,” she said.

With more research, “We will then be better equipped to support these young people, instead of just sending them to a psychiatrist and hoping that their distress will disappear spontaneously.”

VanderLaan thinks it’s a little too soon to jump to that conclusion.

“At present, MRI is not a reliable tool for determining whether a person is transgender or cisgender, and there is debate about its utility for distinguishing cisgender males and females.”

Because studies so far have focused on group averages, he said, individuals could vary considerably.

“Even if one day it were possible to use MRI to determine gender identity, I have a difficult time imagining that a care provider would give more weight to the result of a brain scan than to a person’s stated identity as transgender or cisgender.”

Alexa recorded one family’s conversations and sent them to a friend, without them knowing

May 24, 2018 — Carlo Giacometti, Kernel Developer, Algorithms R&D

Introduction

Recognizing words is one of the simplest tasks a human can do, yet it has proven extremely difficult for machines to achieve similar levels of performance. Things have changed dramatically with the ubiquity of machine learning and neural networks, though: the performance achieved by modern techniques is dramatically higher compared with the results from just a few years ago. In this post, I’m excited to show a reduced but practical and educational version of the speech recognition problem—the assumption is that we’ll consider only a limited set of words. This has two main advantages: first of all, we have easy access to a dataset through the Wolfram Data Repository (the Spoken Digit Commands dataset), and, maybe most importantly, all of the classifiers/networks I’ll present can be trained in a reasonable time on a laptop.

It’s been about two years since the initial introduction of the Audio object into the Wolfram Language, and we are thrilled to see so many interesting applications of it. One of the main additions to Version 11.3 of the Wolfram Language was tight integration of Audio objects into our machine learning and neural net framework, and this will be a cornerstone in all of the examples I’ll be showing today.

Without further ado, let’s squeeze out as much information as possible from the Spoken Digit Commands dataset!

The Data

Let’s get started by accessing and inspecting the dataset a bit:

&#10005 ro=ResourceObject["Spoken Digit Commands"]

The dataset is a subset of the Speech Commands dataset released by Google. We wanted to have a “spoken MNIST,” which would let us produce small, self-enclosed examples of machine learning on audio signals. Since the Spoken Digit Commands dataset is a ResourceObject, it’s easy to get all the training and testing data within the Wolfram Language:

&#10005 trainingData=ResourceData[ro,"TrainingData"]; testingData=ResourceData[ro,"TestData"]; RandomSample[trainingData,3]//Dataset

One important thing we made sure of is that the speakers in the training and testing sets are different. This means that in the testing phase, the trained classifier/network will encounter speakers that it has never heard before.

&#10005 Intersection[trainingData[[All,"SpeakerID"]],testingData[[All,"SpeakerID"]]]

The possible output values are the digits from 0 to 9:

&#10005 classes=Union[trainingData[[All,"Output"]]]

Conveniently, the length of all the input data is between .5 and 1 seconds, with the majority for the signals being one second long:

&#10005 Dataset[trainingData][Histogram[#,ScalingFunctions->"Log"]&@*Duration,"Input"]

Encoders

In Version 11.3, we built a collection of audio encoders in NetEncoder and properly integrated it into the rest of the machine learning and neural net framework. Now we can seamlessly extract features from a large collection of audio recordings; inject them into a net; and train, test and evaluate networks for a variety of applications.

Since there are multiple features that one might want to extract from an audio signal, we decided that it was a good idea to have one encoder per feature rather than a single generic "Audio" one. Here is the full list:

• "Audio"
• "AudioSTFT"
• "AudioSpectrogram"
• "AudioMelSpectrogram"
• "AudioMFCC"

The first step (which is common in all encoders) is the preprocessing: the signal is reduced to a single channel, resampled to a fixed sample rate and can be padded or trimmed to a specified duration.

The simplest one is NetEncoder["Audio"], which just returns the raw waveform:

&#10005 encoder=NetEncoder["Audio"]

&#10005 encoder[RandomChoice[trainingData]["Input"]]//Flatten//ListLinePlot

The starting point for all of the other audio encoders is the short-time Fourier transform, where the signal is partitioned in (potentially overlapping) chunks, and the Fourier transform is computed on each of them. This way we can get both time (since each chunk is at a very specific time) and frequency (thanks to the Fourier transform) information. We can visualize this process by using the Spectrogram function:

&#10005 a=AudioGenerator[{"Sin",TimeSeries[{{0,1000},{1,4000}}]},2]; Spectrogram[a]

The main parameters for this operation that are common to all of the frequency domain features are WindowSize and Offset, which control the sizes of the chunks and their offsets.

Each NetEncoder supports the "TargetLength" option. If this is set to a specific number, the input audio will be trimmed or padded to the correct duration; otherwise, the length of the output of the NetEncoder will depend on the length of the original signal.

For the scope of this blog post, I’ll be using the "AudioMFCC" NetEncoder, since it is a feature that packs a lot of information about the signal while keeping the dimensionality low:

&#10005 encoder=NetEncoder[{"AudioMFCC","TargetLength"->All,"SampleRate"->16000,"WindowSize" -> 1024,"Offset"-> 570,"NumberOfCoefficients"->28,"Normalization"->True}] encoder[RandomChoice[trainingData]["Input"]]//Transpose//MatrixPlot

SEE