SimplyUnknown

Thanks for the short explanation. That clears things up, and you make the First Light approach sound very cool.

As far as I can tell, the newsworthy part is that they achieve fusion using their "two-phase" which is supposedly different than conventional tokamaks such as ITER uses. I'm not sure how either of those technologies exactly work, but the article seems to suggest this is a cheaper way to build a fusion reactor. Then again, this is a press piece, so not exactly unbiased.

I agree that people have been reckless with population safety in past, but your comment seems to imply that the Millenial generation are somehow flawless with regards with this subject. While safety is now more stringent than in the past, I very much doubt that this is the case.

It is a filter in the sense that it can track the system state from one (or more) noisy measurements. The filter part thus reflects the input denoising.

Might be, but it's just a terribly bad look, especially when you offset it against news reports of working conditions at Amazon. Or more local, the young Dutch people not being able to buy any house in the current market. It is not related, but just rubs people in a very wrong way.

The Dutch don't really care about his boat, but they care about a bridge. This bridge, "De Hef", is a symbol of Rotterdam. People are proud of it. Several years ago it was temporarily dismantled and almost permanently removed because it is no longer used. But the people of Rotterdam were pissed and fought to turn it into a monument, which they did. The officials made a promise to never touch the bridge again, except for maintainance. And now some billionaire comes along who wants dismantle the bridge because immense superyacht can't fit underneath. That's why people care

Conditions may apply. If you're homeless or living in a country with ongoing war, disease or hunger the likelihood that you have to worry about any of these things goes up dramatically.

He is not disgraced for his political beliefs, but because he was caught red-handed involved in government corruption and political meddling by foreign nations[1].

[1] Short summary: https://en.wikipedia.org/wiki/Ibiza_affair

> - discourages vote selling (anonymity is one way to achieve this)

Maybe more general: It should be infeasible to perform fraud. Ideally the infeasibility should scale exponentially with the size of the election.

It's most certainly not obvious how the relation df(x)/dx = f(x), the ratio between the diameter and circumference of a circle , and the square root of -1 -- a "trick" to extend the number line with a second dimension -- are related, let alone be used to create an useful construct.

Essentialy, you do the same thing as an FFT for separation. A gaussian blur is a low-pass filter, and functionally the same as selecting all frequency component within a radius r of the DC component (assuming the DC component is in the center of the image). The radius r depends on the width of the gaussian kernel.

The only difference could be in computational time (might be that the FFT is faster) and will be different if you use a different kind of blurring (i.e. non-gaussian such as selecting a square region in the frequency domain).

I mean, we definitely can. We see stars that are thousands of lightyears away! However, maybe OP means "receive" as "response to a message we sent". This is indeed something we are able to do since ~100 years, so our communication bubble is currently ~100 lightyears.

Another explanation: Ryanair is a budget airline and was able to buy more 737-Maxes at a very low price, making their already cheap tickets even cheaper, allowing them to increase profit margins and cementing their status as the cheapest airline around even if that comes at the expense of safety.

Either way, both explanations are speculation and we'll have to see how that turns out.

One of the advantages of using the CPU rather than GPU for inference (especially with batch size 1) is that it doesn't need data transfer from host to device, which is a notoriously slow, asynchronous process. This could also explain the difference in total run time, if measured correctly.

Looks cool!

Two questions:

- Is this approach also learning the palette? It is kind presented as a given here but it is of course very important for a good dithering.

- The loss function might work better on spatially downsampled images. The downsampling causes a mix of the image colors making the dithered image look more like the original given a good dithering. This also naturally removes the variance that is now penalized in the loss function as this is blurred away.

> Is this really what a "done" software is?

I think it is. It basically solved all the problems that were there at its inception and the world moved on. It continues to offer a solution for the original problems (which legacy software most likely still has) and can still be used with minimal engineering effort.

I would call that done.

You mean like arxiv-sanity? As I understand it, it trains a SVM on papers you like to suggest papers that are on the same side of the hyperplane. Could be used as a quality classifier by only liking high-quality work.

No the resolution can be set at will, down from microns to centimeters (depending on the gradient system and sequence). But! The amount of signal per voxel depends on the magnetic field strength. So with tiny voxels there is hardly any signal at low field strength so you won't see anything and you will measure mostly noise.

There is some renewed interest in low field scanners. Past four decades have mainly been pushing for higher fields to improve SNR. But low field has a few advantages:

- no need for superconducting magnets. This is a big win for developing countries and portable systems.

- better geometric accuracy. While not an show stopper as it can be corrected for, but low field exhibits better geometric accuracy as you move away from isocenter, which is of extra concern with e.g. radiotherapy applications.

- more patient comfort as lower field strengths need less strong gradients thus less chance for PNS (involuntary muscle twitches due electric fields induced in the nerves)

No. More teslas are needed for more signal to improve the signal to noise ratio. And there is a lot of noise with portable non-shielded MRI. The only way to compensate for this is by having bigger voxels (3-5mm per voxel rather than the standard 1-2mm), or repeat scans and perform averaging. The more drastic way is to do funky image reconstruction with compressed sensing or deep learning.

The wider bore does indeed come at a cost. You have more geometric distortion and a less homogenous magnetic field at the edges of the bore. Also, you will need stronger magnetic gradients to form image, causing more energy disposition (i.e. Heating the patient) and higher probability of peripheral nerve stimulation (PNS, involuntary muscle twitches due electric fields induced in the nerve caused by MR gradients).

I also think that's partly survivor bias. You listen to the best hits of those decades, tricking you into thinking that all music from those decades was like that. In reality, there was a lot of junk in those times too but we don't listen to it anymore. The same will happen with the 10s and the current 20s in twenty years.

> but programming is also an essential skill for pushing the boundaries in just about any field now

This is definitely true. But it is important to remember that writing code is a means to an end and not the goal in many PhDs. You do it to get results and to many bears no more importance than measuring chemicals or prepping an animal for examination. In the eyes of professional coders this will lead to poor practices (hard to maintain, little testing, etc.) but up to a certain degree this is fine, IMO, because you don't need production-quality code to get results.

> You don't need to get a phd to push the boundaries of knowledge > A lot of work that generates real value is also pushing the boundaries of computer science.

I agree with both of these. A lot of cool stuff pushing the boundary of knowledge is happening outside of academia. But a lot of cool stuff is also happening within academia.

> Good scientific work often has real world impact. I agree, whether it is within or outside of academia. As I think that this is the goal of academia, I remain with my stance that if you aren't trying to do that you are in the wrong place.

Everything is political. It's a waste of time to pretend they aren't. However, in academia these effect are amplified due to perverse incentives (publication count over doing meaningful research) and power dynamics (senior advisor pushing his/her agenda over what is The Right Thing).

But it is important to remember what you are doing. In a PhD you should be pushing the boundary of knowledge. As a programmer in a company, you should be solving problems to generate value for your customer and your company.

If you aren't doing these things, you are in the wrong place.

This is a Big Deal in MRI image reconstruction. As the MR scanner samples in the frequency domain (or k-space, in the nomenclature) one can significantly accelarete MR exams by using C/S reconstruction. The only restriction is that the sampling mask is random. This causes incoherent artifacts in image space, which can be removed by denoising the image in a sparse domain, e.g. wavelets. See also the work by Lustig et al. https://onlinelibrary.wiley.com/doi/full/10.1002/mrm.21391

> True, it's even better than normal.

That's definitely not true.

> Let's not pretend that the extent of California's reaction to the coronavirus is even comparable to that of Netherlands, Japan

You are right, it's not. But that wasn't the point. But pretending that California is the only one suffering and taking measures while the Netherlands or Japan are not is also out of touch with reality.

I don't know if you live in the Netherlands, but stating "only certain businesses like brothels have been suspended" is a massive understatement of the impact of the "intelligent lockdown" as it is called.

Bars, gyms, schools have been closed for two months now. Only a few restaurants offer take-out and more offer delivery. Dining in is not possible. Work-from-home is default for everyone I know. Nursing homes have only just been allowing visitors again.

It is true that the Netherlands was not locked down, but people definitely are not going normal about their lives.

Funnily enough, yesterday an article appeared[1] claiming the number of suicides in the Netherlands has dropped up to 20 percent. Cause as of yet unknown. Also unclear whether this effect is temporary.

One hypothesis the article states is that in normal life people's sensory inputs are overwhelmed and cause anxiety. The sudden calmness the lockdown introduced might have a positive effect. This is, however, in contrast with the paired increased loneliness.

[1] https://nos.nl/artikel/2334626-tot-20-procent-minder-zelfdod...

Turns out, they do and the author of this wikipedia page is indeed the founder of the program: https://unnwo.org/

Sure, but you never know which way it will go but this is assuming the 737 MAX gets recertified by the FAA and operators are going to buy them because it is cost-effective or whatever. This means that the group that doesn't have a 737 MAX is going to be smaller and therefore harder/way more expensive to avoid.

I'm totally with you, but how do you prevent it? Usually you only know what type of plane you fly with after buying tickets. A solution might be to only buy tickets from operators that don't have with 737 MAX planes but this group might be getting smaller over long time.