New Device Could Greatly Improve Speech and Image Recognition

jan_jes writes: Scientists have successfully demonstrated pattern recognition using a magnonic holographic memory device, a development that could greatly improve speech and image recognition hardware. The researchers built a prototype eight-terminal device consisting of a magnetic matrix with micro-antennas to excite and detect the spin waves. The micro-antennas allow the researchers to generate and recognize any input phase pattern, a big advantage over existing practices. It takes about 100 nanoseconds for recognition, which is the time required for spin waves to propagate and to create the interference pattern. The main challenge associated with magnonic holographic memory is the scaling of the operational wavelength, which requires the development of sub-micrometer scale elements for spin wave generation and detection.

Comments

Car analogy?

[sinij]

Could someone please explain this with a car analogy?

Re:Car analogy?

[gstoddart]

Spinners, a big antenna, and fuzzy dice make for one bitchin' ride.

Beyond that, I got nothing.

Re:Car analogy?

[Impy the Impiuos Imp]

It's a cannon.

Re:Car analogy?

[fyngyrz]

I guess you haven't tried to actually use a Google product from the inside. Fundamentally broken, obvious and repeatable bugs have gone unfixed for years, but as they tell us: "they're working on it." (cough[Shopping]cough)

If it's in a Google car, they'll claim it isn't evil, while being really underhanded (cough[IP rights]cough), but it won't work right (cough[Shopping]cough), and just as you you commit a significant amount of resources to it, they'll either discontinue it (cough[cough]cough) or sideline it. Or never, ever add the features that would make it something actually reasonable (cough[Gmail]cough) Or simply blow out the decent features (cough[Maps]cough) Or never bother to bring it to a level of performance that is even moderately reasonable (cough[Google+]cough)

Unless it never becomes popular. In that case, it might hang around forever. But still under-performing / broken / evil, etc.

No, I'm not bitter. I love when a company wastes my time as if it's worth nothing. Finally I realized that trying to work with Google was making my time worth nothing. So in a way, they had the right idea from the start.

The only car analogy I can come up with is the insufficiently Humvees the government gave our soldiers to drive over IEDs in.

I understood some of those words

[Anonymous Coward]

It seems like the kind of thing a random science-word generator would produce.

Re:I understood some of those words

[orasio]

This is news for nerds.
It's a pattern detection strategy that relies on generating waves with input data, interweaving them physically, and using arrays of antennas to detect patterns.
That's from the first couple of paragraphs.
I don't know a lot of physics, but I am a nerd, and I like this kind of thing, so I can learn about cool stuff.
If you don't care about it, you can look at other stories that talk about tesla and bill gates and whatever else. Posting is not mandatory.

Re:I understood some of those words

[MobyDisk]

I would come here more often if orasio wrote the summaries.

The problem with the summary is that it assumes the reader is already familiar with the device. Your summary does not suffer from that problem. For instance "prototype eight-terminal device consisting of a magnetic matrix with micro-antennas to excite and detect the spin waves." WHAT spin waves? What is a terminal in this context and why is the a key thing in the summary? The summary already presupposes too much, even for a technical news site.

On the flip side, it would be nice if you didn't also insult the person who asked for clarification. The summary is indeed confusing.

+1 for insightful explanation.
-1 for being an asshat about it.

Re:I understood some of those words

[Sardaukar86]

If you don't care about it, you can look at other stories that talk about tesla and bill gates and whatever else. Posting is not mandatory.

I'm not sure his point was that he didn't care about it.

I also found the summary fairly opaque, although that is likely a reflection of my minimal understanding of math and physics. The editors usually provide summaries of complex topics that are a little more approachable for the layman but I suppose I could always flout the law and RTFA.

Hmmm ...

[gstoddart]

The researchers built a prototype eight-terminal device consisting of a magnetic matrix with micro-antennas to excite and detect the spin waves. The micro-antennas allow the researchers to generate and recognize any input phase pattern

Resulting in improved leveraging of synergies and a minty fresh taste.

Honestly, I hope that makes sense to someone, because it sounds like computer generated gibberish to me. :-P

Re:Huh?

[pushing-robot]

Don't worry, scientists are working on a device that can recognize them.

We have a winner...

[jeffb (2.718)]

...for this year's "Science or Star Trek Technobabble?" championship!

Confusing article

[omnichad]

The only thing I can gather from the article is that this enables a massively parallel comparator. And then somehow that translates to someday faster pattern recognition for speech / image. Very scarce on details.

Miracles, yo

[wcrowe]

After reading that summary I feel like a juggalo. "...F*cking magnonics, how do they work?.."

Re:Huh?

[gstoddart]

Nope, stupid human can't type ... magnonic is apparently a real thing.

Well played, PR team

[jfengel]

Most of the time I gripe about press releases that massively overpromise, and don't contain enough real information to figure out what was actually done.

In this case they at least put the word "could" rather than "will" in the headline. And while the barrage of jargon that followed is incomprehensible to me, it's at least pretty clear that it's real work rather than science-by-press-release.

Now I'd appreciate it if somebody would come along and dumb it down to my level, which is still considerably higher than the fourth-grade education they usually target. So, in all seriousness: thanks to the PR team.

Re:Huh?

[SwingKing]

Looked at that page, which lead me to the SpinWaves page, which had the text: "The simplest way of understanding spin waves is to consider the Hamiltonian {H} for the Heisenberg ferromagnet:" at which point my head exploded and I went back to watching Star Trek reruns. At least there the technobabble make sense...

Back to the analog future

[jtara]

It's a pattern detection strategy that relies on generating waves with input data, interweaving them physically, and using arrays of antennas to detect patterns.

So, it's an analog computer.

Yah but

[JustNiz]

I tried this out in the lab but it melted my interocitor.

Re:Huh?

[ClickOnThis]

If I could explain it in 30 seconds, it wouldn't be worth a Nobel Prize.
-- Richard Feynman, responding to a reporter's question, after he won the Nobel Prize

Simplicity is good, but not always attainable.

Ok I'll try:

[tomxor]

From what i can quickly gather from the article:

This is all based on magnonics, which in short - is the use of magnetic spin for binary storage and or logic. This device focuses on the later...

It does this by constructing a matrix of magnetic nodes that are effectively interconnected to neighbours (moor?) via spatial magnetic-spin sensitivity, these interconnects form the dynamic logic processing ability of the matrix.

I think that this is somewhat like a (soft) convolutional artificial neural network for image recognition, these are constructed out of a 2d or 3d matrix of nodes with weighted interconnects in a moor-neighbourhood arrangement. The difference here i guess is that a) it's done with magnetic spin (i really have no idea why this is an advantage, maybe i'm all wrong about this) and b) being an application specific piece of hardware each node works in parallel (this is trumped as the primary reason for the speed potential in the article).

... Big disclaimer: I am massively speculating because the use case is not made super clear.

Re:Ok I'll try:

[knarfling]

From what i can quickly gather from the article:

This is all based on magnonics, which in short - is the use of magnetic spin for binary storage and or logic. This device focuses on the later...

From what I could tell from the article, it appears to focus on both. The device allows them to quickly create and store a pattern. It then allows another pattern to be created and quickly compare the patterns.

It does this by constructing a matrix of magnetic nodes that are effectively interconnected to neighbours (moor?) via spatial magnetic-spin sensitivity, these interconnects form the dynamic logic processing ability of the matrix.

I think that this is somewhat like a (soft) convolutional artificial neural network for image recognition, these are constructed out of a 2d or 3d matrix of nodes with weighted interconnects in a moor-neighbourhood arrangement. The difference here i guess is that a) it's done with magnetic spin (i really have no idea why this is an advantage, maybe i'm all wrong about this) and b) being an application specific piece of hardware each node works in parallel (this is trumped as the primary reason for the speed potential in the article).

... Big disclaimer: I am massively speculating because the use case is not made super clear.

From what I could tell, the advantage of the magnetic spin is that with an 8 terminal node, it can quickly create a matrix of 1,000 bits of data. It can then compare that matrix with a stored matrix. With several nodes working in parallel, it takes the same time to compare 1,000 bits of data as 10,000,000. The potential (and right now, only potential) is that a computing device with this hardware integrated in could compare an image with a stored image much faster than conventional circuits could make the comparison and result in a better matching algorithm.