8 stories
·
5 followers

pic.twitter.com/39xahJQJgj

1 Comment and 2 Shares
Read the whole story
m
92 days ago
reply
F1
dreadhead
105 days ago
reply
Vancouver Island, Canada
Share this story
Delete

My Childhood Dog Never Got Vaccinated and He’s Alive and Well on That Farm Upstate

1 Comment and 2 Shares

A lot of people have been talking about this vaccine lately, asking plenty of important questions. Does it defend against COVID-19? How effective is it against the Delta Variant? Is it worth getting a tracking device injected into your body that also makes you sterile? Did Jesus need a vaccine? Well, all those questions are pointless and I can prove it. My childhood dog never got vaccinated and for the last 17 years, he’s been absolutely crushing it on that farm upstate, according to my parents.

I’m a free thinker. A rebel without bias. A doctor without borders. Or a medical license. I play my own tune and that tune is the star-spangled banner. Except for when it’s Enter Sandman. If experience has taught me anything it’s never to be a guinea pig for the state and that dogs live to about 50 if provided the right environment.

The American public is very childish. They’ll believe any fairy tale you tell them. Like how there are some tiny organisms coming to kill us that we can’t see (how convenient) and our only protection is getting some juice put in your arm. Doesn’t that sound ridiculous. It sure does to me, my family, and my dog, who is currently vaccine-free, romping around in meadowy fields until the end of time. It’s time to grow up and resist, people!

Taking the lead from my immortal pup, I will never get the COVID vaccine. The only things I put in my body are time-tested and proven to be safe. Now, if you’ll excuse me, I need to go get my annual distemper booster.

The post My Childhood Dog Never Got Vaccinated and He’s Alive and Well on That Farm Upstate appeared first on The Hard Times.

Read the whole story
m
93 days ago
reply
T
dreadhead
93 days ago
reply
Vancouver Island, Canada
Share this story
Delete

Google’s New AI Photo Upscaling Tech is Jaw-Dropping

2 Comments and 3 Shares

Photo enhancing in movies and TV shows is often ridiculed for being unbelievable, but research in real photo enhancing is actually creeping more and more into the realm of science fiction. Just take a look at Google’s latest AI photo upscaling tech.

In a post titled “High Fidelity Image Generation Using Diffusion Models” published on the Google AI Blog (and spotted by DPR), Google researchers in the company’s Brain Team share about new breakthroughs they’ve made in image super-resolution.

In image super-resolution, a machine learning model is trained to turn a low-res photo into a detailed high-res photo, and potential applications of this range from restoring old family photos to improving medical imaging.

Google has been exploring a concept called “diffusion models,” which was first proposed in 2015 but which has, up until recently, taken a backseat to a family of deep learning methods called “deep generative models.” The company has found that its results with this new approach beat out existing technologies when humans are asked to judge.

The first approach is called SR3, or Super-Resolution via Repeated Refinement. Here’s the technical explanation:

“SR3 is a super-resolution diffusion model that takes as input a low-resolution image, and builds a corresponding high resolution image from pure noise,” Google writes. “The model is trained on an image corruption process in which noise is progressively added to a high-resolution image until only pure noise remains.

“It then learns to reverse this process, beginning from pure noise and progressively removing noise to reach a target distribution through the guidance of the input low-resolution image.”

Before (left) and after (right) upscaling with SR3.

SR3 has been found to work well on upscaling portraits and natural images. When used to do 8x upscaling on faces, it has a “confusion rate” of nearly 50% while existing methods only go up to 34%, suggesting that the results are indeed photo-realistic.

Before (left) and after (right) upscaling with SR3.
Before (left) and after (right) upscaling with SR3.

Here are other portraits upscaled from low-resolution originals:

A selection of portraits upscaled from low-res originals by AI.
A selection of portraits upscaled from low-res originals by AI.

Once Google saw how effective SR3 was in upscaling photos, the company went a step further with a second approach called CDM, a class-conditional diffusion model.

“CDM is a class-conditional diffusion model trained on ImageNet data to generate high-resolution natural images,” Google writes. “Since ImageNet is a difficult, high-entropy dataset, we built CDM as a cascade of multiple diffusion models. This cascade approach involves chaining together multiple generative models over several spatial resolutions: one diffusion model that generates data at a low resolution, followed by a sequence of SR3 super-resolution diffusion models that gradually increase the resolution of the generated image to the highest resolution.”

Google has published a set of examples showing low-resolution photos upscaled in a cascade. A 32×32 photo can be enhanced to 64×64 and then 256×256. A 64×64 photo can be upscaled to 256×256 and then 1024×1024.

As you can see, the results are impressive and the final photos, despite having some errors (such as gaps in the frames of glasses), would likely pass as actual original photographs for most viewers at first glance.

“With SR3 and CDM, we have pushed the performance of diffusion models to state-of-the-art on super-resolution and class-conditional ImageNet generation benchmarks,” Google researchers write. “We are excited to further test the limits of diffusion models for a wide variety of generative modeling problems.”

Read the whole story
m
93 days ago
reply
I
dreadhead
93 days ago
reply
Vancouver Island, Canada
Share this story
Delete
1 public comment
duerig
93 days ago
reply
This looks really neat and is potentially a great way to 'enhance' entertainment. My real worry about stuff like this is that the 'enhancement' is simply a plausible filling in of the picture. But since it is magical AI algorithms (tm), people will believe that it represents truth. Will we see people try to 'enhance' grainy security footage and then finger a suspect because the higher-resolution photo looks like them? This seems like a great way to give ourselves false certainty about things we need to remain skeptical of.
ChrisDL
93 days ago
Definitely. I really wish they had also added what the actual original looked like so you could compare it with the upscaled version.

Preferred Networks: A 500 W Custom PCIe Card using 3000 mm2 Silicon

2 Shares

Pushing the physical limits of hardware is always a fun story, but I was positively blown away when I saw this poster at Supercomputing. Preferred Networks, seemingly a spin-out from Tokyo University, is shoving several large chiplets into a single PCIe card for peak performance, peak power, and it looks like they’re ready to deploy over 4800 of these cards into a custom supercomputer.

Read the whole story
m
721 days ago
reply
Share this story
Delete

Who Keeps Europe’s Farm Billions Flowing? Often, Those Who Benefit

1 Share
Conflict-of-interest rules are uneven. Influence-peddling is murky. The massive E.U. subsidy program is built to keep the money moving.

Read the whole story
m
721 days ago
reply
Share this story
Delete

Robot Revolution: Amazon’s Move Signals End of Line for Many Cashiers

1 Comment
The next jobs set to disappear may be ones that are a bigger part of people’s daily lives: retail workers and cashiers in stores and restaurants.

Read the whole story
m
1627 days ago
reply
+
Share this story
Delete
Next Page of Stories