Viewing profile — soraki_soladead

The latent representations of the data are like points on a surface. That surface is the manifold. We don't typically have the full manifold and can only sample points from it by e…

I might be misunderstanding your point but this conflates the distinguishing features of each. you mention expansion but autoencoders canonically compress their inputs. autoencoder…

This is awesome! Can someone in this field comment on the implications of sidestepping the cytoskeleton?

Possibly because SIGGRAPH is coming up and these were papers submitted to that conference.

The original NCA is probably a helpful intro: https://distill.pub/2020/growing-ca/

I think you're misremembering or misunderstanding Picard's argument. It isn't a tangent. Here's the transcript[0]. TL;DR Picard's initial arguments are pretty weak, even admitting …

The post links another that goes into the theory a little: https://shahriyarshahrabi.medium.com/in-the-valley-of-gods-s... Apparently a combination of Mie and Rayleigh scattering. …

I'm not saying you're wrong but then why do a big website and branding push. If they had someone in mind they'd bury it on a regular job posting. They specify early to mid career. …

> There is also no pride. Is the pride not in solving the users' problems? > nobody talks about it, treats it with interest, or pays above market rate to work on it. Definitely nee…

Context: https://xkcd.com/1053/ Then, if you're like me and read this years ago, play around with the Light Mode dropdown which was new to me. :)

We don't know however "It would take so long" is an anthropomorphic assumption of time scale.

Why would we settle for anything less than discontinuing both?

Roughly, when you train a model to make its predictions align to its own predictions in some way, you create a scenario where the simplest "correct" solution is to output a single …

also, BabyLM is more of a conference track / workshop than an open-repo competition which creates a different vibe

Alias-Free GANs? https://nvlabs.github.io/stylegan3/

We have lossless memory for models today. That's the training data. You could consider this the offline version of a replay buffer which is also typically lossless. The online, con…

You might enjoy this paper[0] which shows that recurrent position encodings recover grid cell representations and maps to path integration found in a popular model of the hippocamp…

Fwiw, that's SwiGLU in #3 above. Swi = Swish = silu. GLU is gated linear unit; the gate construction you describe.

It's not always about cost. Sometimes the ergonomics of a local machine are nicer.

Agree that's not a great look for the supervisor. Cyclists have a bad rep in SF because many (not all) ride quite dangerously. It's a common sight to see cyclists running four-way …

Cyclists do this all the time in SF. Afaik an "Idaho stop" is not legal here, despite it being common and often unsafe for obvious reasons.

FLOPs by perplexity by samples is an interesting way to compare this family of models.

https://github.com/cozodb/pycozo/blob/main/pycozo/test_build... Here's the python version of what I think you're looking for. Shouldn't be too difficult to port to rust.