AI generates images from our brains

[twillisdc]

Granted, this is from VICE but this is super weird. I'm of the belief that once technology is released to the public, it's been used a long time by the powers that be and this seems no different.

Researchers Use AI to Generate Images Based on People's Brain Activity

Researchers found that they could reconstruct high-resolution images from brain activity.

www.vice.com

 

[ThinPicking]

Worth a read at least. Dodgy design and optimistic description of results would be my assessment, to put it lightly. They're good enough to include "limitations" but it seems framed around the questionable idea a better model is forthcoming.

"statistical significance" they say.
"a small number of participants" they say... what's small, 2?

Hyperrealistic neural decoding for reconstructing faces from fMRI activations via the GAN latent space - Scientific Reports

Neural decoding can be conceptualized as the problem of mapping brain responses back to sensory stimuli via a feature space. We introduce (i) a novel experimental paradigm that uses well-controlled yet highly naturalistic stimuli with a priori known feature representations and (ii) an...

doi.org

This is extremely cool none the less. Anything involving MRI is to me.

 

[Peater]

I thought most NPCs didn't have a minds eye. I'm not being mean they actually said they couldn't visualise an apple.

 

[ThinPicking]

Earlier similars for the interested.

At this point I'm not sure what's more dangerous, research in and practical applications of model based SI or the "science communication" memes around it.

Generic decoding of seen and imagined objects using hierarchical visual features - Nature Communications

Machine learning algorithms can decode objects that people see or imagine from their brain activity. Here the authors present a predictive decoder combined with deep neural network representations that generalizes beyond the training set and correctly identifies novel objects that it has never...

doi.org

Redirecting

Deep image reconstruction from human brain activity

Author summary Machine learning-based analysis of human functional magnetic resonance imaging (fMRI) patterns has enabled the visualization of perceptual content. However, prior work visualizing perceptual contents from brain activity has failed to combine visual information of multiple...

doi.org

Reconstructing faces from fMRI patterns using deep generative neural networks - Communications Biology

VanRullen and Reddy apply a state-of-the-art AI technique to brain decoding. After learning to translate multi-voxel fMRI activity patterns into the activation space of a deep generative neural network, each particular face viewed, or even imagined, by a human subject in the scanner can be...

doi.org

Redirecting

Redirecting