Literal Odds and Ends

[Speaker to Animals]

The coolest thing about deep learning, and also the shittiest, is that you usually cannot show what the fuck this network even learns. It's an awesome technique to solve many pattern recognition problems, but trying to sell that to business leaders is difficult when you cannot really show *why* it works. People often want strict logical proof that it will work in the field, but that's not really happening.

If I were to go back to the doctoral program, one area I might consider focusing on is feature extraction with a mind towards being able to prove why a given trained network actually works.

[SuburbanFarmer]

It definitely doesn't inspire confidence.

At that point, the executives are probably thinking "sounds like this just needs a good interface, then we fire the nerds".

[SuburbanFarmer]

What kind of fucking idiot is running the schools now? I’m supposed to find “friendly numbers” and draw pictures of a division problem??

How about I just teach my kid to do math, if you won’t, and you shove your pictures up your ass?

[Speaker to Animals]

What kind of fucking idiot is running the schools now? I’m supposed to find “friendly numbers” and draw pictures of a division problem??

How about I just teach my kid to do math, if you won’t, and you shove your pictures up your ass?

Divide their budget number by the half we could save from firing half the do-nothing teachers.

[de officiis]

Coolness. See helpful diagram on p.4.

Deep image reconstruction from human brain activity

Guohua Shen, Tomoyasu Horikawa, Kei Majima, and Yukiyasu Kamitani

1ATR Computational Neuroscience Laboratories, Kyoto 619-0288, Japan 2Kyoto University, Kyoto 606-8501, Japan

Abstract

Machine learning-based analysis of human functional magnetic resonance imaging (fMRI) patterns has enabled the visualization of perceptual content. However, it has been limited to the reconstruction with low-level image bases (Miyawaki et al., 2008; Wen et al., 2016) or to the matching to exemplars (Naselaris et al., 2009; Nishimoto et al., 2011). Recent work showed that visual cortical activity can be decoded (translated) into hierarchical features of a deep neural network (DNN) for the same input image, providing a way to make use of the information from hierarchical visual features (Horikawa & Kamitani, 2017). Here, we present a novel image reconstruction method, in which the pixel values of an image are optimized to make its DNN features similar to those decoded from human brain activity at multiple layers. We found that the generated images resembled the stimulus images (both natural images and artificial shapes) and the subjective visual content during imagery. While our model was solely trained with natural images, our method successfully generalized the reconstruction to artificial shapes, indicating that our model indeed ‘reconstructs’ or ‘generates’ images from brain activity, not simply matches to exemplars. A natural image prior introduced by another deep neural network effectively rendered semantically meaningful details to reconstructions by constraining reconstructed images to be similar to natural images. Furthermore, human judgment of reconstructions suggests the effectiveness of combining multiple DNN layers to enhance visual quality of generated images. The results suggest that hierarchical visual information in the brain can be effectively combined to reconstruct perceptual and subjective images.

[Speaker to Animals]

Yeah, it's kind of weird, but you can sort of go backwards and generate information too. We have seen people making those strange images in this way.

[Speaker to Animals]

I just realized how to make sense of the Star Trek commie narrative. We are the Mirror Universe. The characters in the main plot are the communists that invade our timeline and fuck it all up by making Mirror Spok an emperor.

In the correct (mirror) universe, Earth comes first. Humans are strong, prosperous and rule over their section of the galaxy. But in the Commieverse, humans are overrun by aliens, giving up their sovereignty in ridiculous galactic democracy schemes.

Somebody needs to reboot this thing and show it from our side (mirror universe) where humans put humans first, enforce strict border controls, and are thus able to become the strongest species in the galaxy.

[SuburbanFarmer]

I just realized how to make sense of the Star Trek commie narrative. We are the Mirror Universe. The characters in the main plot are the communists that invade our timeline and fuck it all up by making Mirror Spok an emperor.

In the correct (mirror) universe, Earth comes first. Humans are strong, prosperous and rule over their section of the galaxy. But in the Commieverse, humans are overrun by aliens, giving up their sovereignty in ridiculous galactic democracy schemes.

Somebody needs to reboot this thing and show it from our side (mirror universe) where humans put humans first, enforce strict border controls, and are thus able to become the strongest species in the galaxy.

I can shave my beard into a goatee at any moment, for maximum Evilness.

[Speaker to Animals]

A goatee is about the only thing I can easily grow out, so I barely make it into the empire.

[Okeefenokee]

Chinese dictator for life banned the letter N apparently.