bluemellophone
@bluemellophone@lemmy.world
- Comment on Cracker Barrel Outrage Was Almost Certainly Driven by Bots, Researchers Say 1 week ago:
The entire thing feels manufactured by design to get people talking about Cracker Barrel.
- Comment on Can you think of any now? 1 week ago:
No, you are correct. Hinton began researching ReLUs in 2010 and his students Alex Krizhevsky and Ilya Sutskever used it to train a much deeper network (AlexNet) to win the 2012 ILSVRC. The reason AlexNet was so groundbreaking was because it brought all of the gradient optimization improvements (SGD with momentum as popularized by Schmidhuber, and dropout), better activation functions (ReLU), a deeper network (8 layers), supervised training on very large datasets (necessary to learn good general-purpose convolutional kernels), and GPU acceleration into a single approach.
NNs, and specifically CNNs, won out because they were able to create more expressive and superior image feature representations over the hand-crafted features of competing algorithms. The proof was in the vastly better performance, it was a major jump when the ILSVRC was becoming saturated. Nobody was making nearly +10% improvements on that challenge back the , it blew everybody out of the water and made NNs and deep learning impossible to ignore.
- Comment on Can you think of any now? 1 week ago:
Before LeNet and AlexNet, SVMs were the best algorithms around. People used HOG+SVM, SIFT, SURF, ORB, older Haar / Viola-Jones features, template matching, random forests, Hough Transforms, sliding windows, deformable parts models… so many techniques that were made obsolete once the first deep networks became viable.
The problem is your schooling was correct at the time, but the march of research progress eventually saw 1) the creation of large, million-scale supervised datasets (ImageNet) and 2) larger / faster GPUs with more on-card memory.
- Comment on T-Mobile secretly records iPhone screens and claims it's being helpful. 4 months ago:
That would be a pretty big security hole in iOS if that was allowed, but it isn’t. Notification and other UI elements are rendered on top of the underlying app, which does not have access to or cannot see the full screen’s canvas. We can see practical implementations of this “snapshot” test feature in code:
- Comment on Infrared contact lenses let you see in the dark 4 months ago:
To counter the first theory, a contact cannot change the physical biology of the photon detecting cells (rods and cones) in the back of your eye. Nothing can, short of modifying your genetics.
So you can either become part mantis shrimp, or shift the wavelengths into the spectrum your biology already can absorb and interpret.
- Comment on Nintendo delays Switch 2 preorders over tariff concerns 5 months ago:
You seem to be missing the point, the Nintendo Switch is handily the most popular game console in US gaming history. It has officially sold more units than the PS2.
I’m sorry that you don’t appreciate the switch, it’s an amazing general purpose gaming console.
- Comment on Apple CEO Tim Cook Donating $1 Million to Trump's Inaugural Fund 8 months ago:
Previous administrations have used special government funds that are dedicated to the transition and inauguration. This comes with obvious restrictions and ethics requirements since it is taxpayer money.
Trump has decided to not use this money and get donations instead, which allows him and his incoming administration to skirt the normal mandatory reporting, disclosures, and rules.
- Comment on Juno for YouTube has been removed from the App Store 11 months ago:
Dang, I used to have so many Juno disks. What a flashback.