[HN Gopher] Multimodal Neurons in Artificial Neural Networks
___________________________________________________________________
Multimodal Neurons in Artificial Neural Networks
Author : todsacerdoti
Score : 61 points
Date : 2021-03-04 20:13 UTC (2 hours ago)
(HTM) web link (openai.com)
(TXT) w3m dump (openai.com)
| gallerdude wrote:
| I've always thought it's wild how we can apply one concept to so
| many different types of things. For example, if I say something
| is "soft," you probably think of the opposite of firmness. But at
| the same time, I can describe a person as "soft," and the same
| descriptor can say something meaningful about their character.
|
| Seeing the Spider-Man neuron work on multiple types (pictures,
| drawings, text), makes it seem like we can teach AI to learn
| these same type connections.
|
| And if we scale up the network size enough, what if we could see
| these types through the equivalent of a being with 1000IQ? What
| connection types are the most effective for a being like that?
| Can we even understand them? Maybe they would be deep, and
| archetypical in the way that Odysseus and Harry Potter are the
| same, despite the fact that one is an ancient Greek king, and the
| other is a modern British wizard. Even more interestingly, maybe
| the connections would be completely inexplicable to us, with no
| apparent rhyme or reason perceptible to humans.
| colah3 wrote:
| I'm really excited about the dream that we'll be able to learn
| from neural networks. Shan Cater and Michael Nielsen wrote a
| really inspiring article on this
| (https://distill.pub/2017/aia/). I also wrote something about
| this a while back
| (http://colah.github.io/posts/2015-01-Visualizing-
| Representat...).
|
| One of the amazing things about this project exploring CLIP was
| seeing some hints of this. For example, one day I was studying
| one of the Africa neurons and it generated the text "IMBEWU" --
| it turns out this is a popular TV show in South Africa
| (https://en.wikipedia.org/wiki/Imbewu:_The_Seed). That's a
| trivial example, but it begins to hint at something
| interesting.
|
| I'd really love to see what a domain expert analyzing CLIP
| would make of things. For example, I'd love to hear what
| ethnographers think of the region neurons, or what historians
| think of the time period neurons. Especially for future, larger
| models.
| biasdose wrote:
| I'm impressed with OpenAI confronting this head on.
|
| "Our model, despite being trained on a curated subset of the
| internet, still inherits its many unchecked biases and
| associations."
|
| If these models find themselves into production environment - if
| they are good enough and profitable enough - they will eventually
| become legacy systems quietly perpetuating the biases of past
| times.
| kowlo wrote:
| The typographic attacks are great fun. Labelling an apple as a
| toaster is all it takes!
| HPsquared wrote:
| It needs a 'shenanigans' neuron.
| iujjkfjdkkdkf wrote:
| Can someone give more technical detail on what they are showing
| with the "neurons"?
|
| They say "Each neuron is represented by a feature visualization
| with a human-chosen concept labels to help quickly provide a
| sense of each neuron", and these neurons are selected from the
| final layer. I don't think I understand this.
| the8472 wrote:
| Start with random input, then incrementally optimize the input
| to maximize the activation of one of the nodes in the graph,
| the neuron. The visualization is one of those inputs that hit a
| maximum.
___________________________________________________________________
(page generated 2021-03-04 23:00 UTC)