Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

TopoNets: High performing vision and language models with brain-like topography

arxiv.org

Post reply on HN
Print view View original

11–20 of 76 posts

Re: TopoNets: High performing vision and language models with brain-like topography

#11

Post by cma »

energy123 wrote: post #5

The main reason topography emerges in physical brains is because spatially distant connections are physically difficult and expensive in biological systems. Artificial neural nets have no such trade-off. So what's the motivation here? I can understand this might be a very good regularizer, so it could help with generalization error on small-data tasks. But hard to see why this should be on the critical path to AGI. A…

> The main reason topography emerges in physical brains is because spatially distant connections are physically difficult and expensive in biological systems.

The brain itself seems to have bottlenecks that aren't distance related, like hemispheres and the corpus callosum that are preserved over all placental mammals and other mammalian groups have something similar and still hemispheres. Maybe it's just an artifact of bilateral symmetry that is stuck in there from path dependence, or forcing a redundancy to make damage more recoverable, but maybe it has a big regularizing or alternatively specializing effect (regularization like dropout tends to force more distributed representations which seems kind of opposite to this work and other work like "Seeing is Believing: Brain-Inspired Modular Training for Mechanistic Interpretability," https://arxiv.org/abs/2305.08746 ).

Re: TopoNets: High performing vision and language models with brain-like topography

#12

Post by energy123 »

xpl wrote: post #6
energy123 wrote: post #5

The main reason topography emerges in physical brains is because spatially distant connections are physically difficult and expensive in biological systems. Artificial neural nets have no such trade-off. So what's the motivation here? I can understand this might be a very good regularizer, so it could help with generalization error on small-data tasks. But hard to see why this should be on the critical path to AGI. A…

> So what's the motivation here? Better interpretability, I suppose. Could give insights into how cognition works.

That's true, and interpretability is helpful for AI safety.

Re: TopoNets: High performing vision and language models with brain-like topography

#13

Post by mercer »

energy123 wrote: post #5

The main reason topography emerges in physical brains is because spatially distant connections are physically difficult and expensive in biological systems. Artificial neural nets have no such trade-off. So what's the motivation here? I can understand this might be a very good regularizer, so it could help with generalization error on small-data tasks. But hard to see why this should be on the critical path to AGI. A…

I imagine it could be easier to make sense of the 'biological' patterns that way? like, having bottlenecks or spatially-related challenges might have to be simulated too, to make sense of the ingested 'biological' information.

Re: TopoNets: High performing vision and language models with brain-like topography

#14

Post by vessenes »

I hate to dog on research papers. They’re work to write. That said, I think this paper is not likely to be of interest to AI researchers — instead it may be of interest to Neuroscience folks or other brain research types.

The lede — adding topography worsens networks at similar weights — is not only buried, it’s obscured with statements claiming that topo networks show less upheaval when scaled down, e.g. they are more efficient than similar weight networks.

It’s hard for me to see how both these things can be true — the graphs show the more topography is added, the worse the networks perform at the trained model sizes.

To have the second statement “They compress better and are therefore more efficient” also be true, I think you’d need to show a pretty remarkable claim, which is that while a model trained at the same scale as a llama architecture is worse, when you scale them both down, this model becomes not only better than the scaled down llama, but also better than a natively trained model at the new smaller scale.

There is no proof of this in the paper, and good reason to be skeptical of this idea based on the data presented.

That said, like a lot of ideas in AI, this .. works! You can train a model successfully imposing these outside structures on it, and that model doesn’t even suck very much. Which is a cool statement about complexity theory and the resilience of these architectures, in my opinion. But I don’t think it says much else about either the brain or underlying AI ‘truths’.

Re: TopoNets: High performing vision and language models with brain-like topography

#15

Post by mayukhdeb »

light_hue_1 wrote: post #9

They bury the part where inducing brain like structure hurts performance! This is a method to just hurt your network in exchange for nothing useful at all aside from some sketchy story that this is "brain like".

Our goal was never to optimize for performance. There's a long standing hypothesis that topographic structure in the human brain leads to metabolic efficiency. Thanks to topography in ANNs, we were able to test out this hypothesis in a computational setting.

> sketchy story this is "brain like".

we reproduce the hallmarks of functional organization seen in the visual and language cortex of the brain. I encourage you to read the paper before making such comments

Re: TopoNets: High performing vision and language models with brain-like topography

#16

Post by mayukhdeb »

TOMDM wrote: post #7
LZ_Khan wrote: post #2

Shouldn't there be a comparison in performance on common benchmarks to other models? Like a 7B toponet model vs a 7B Llama model? As a layperson I don't understand why topology is a thing to optimize for.

The only potential benefit shown in the paper is the topologically local models seem to be more resilient after pruning. So you may be able to prune a 7B model down to 6B while maintaining most of the capability.

> The only potential benefit

Other benefits:

1. Significantly lower dimensionality of internal representations 2. More interpretable (see: https://toponets.github.io)

> 7B model down to 6B

We remove ~80% of the parameters in topographic layers and retain the same performance in the model. The drop in parameter count is not significant because we did not experiment with applying TopoLoss in all of the layers of the model (did not align with the goal of the paper)

We are currently performing those strong sparsity experiments internally, and the results look very promising!

Re: TopoNets: High performing vision and language models with brain-like topography

#17

Post by TZubiri »

energy123 wrote: post #5

The main reason topography emerges in physical brains is because spatially distant connections are physically difficult and expensive in biological systems. Artificial neural nets have no such trade-off. So what's the motivation here? I can understand this might be a very good regularizer, so it could help with generalization error on small-data tasks. But hard to see why this should be on the critical path to AGI. A…

Maybe this would be relevant for datacenters with significant distance between machines, or multidatacenter systems.

Re: TopoNets: High performing vision and language models with brain-like topography

#18

Post by harles »

vlovich123 wrote: post #8
energy123 wrote: post #5

The main reason topography emerges in physical brains is because spatially distant connections are physically difficult and expensive in biological systems. Artificial neural nets have no such trade-off. So what's the motivation here? I can understand this might be a very good regularizer, so it could help with generalization error on small-data tasks. But hard to see why this should be on the critical path to AGI. A…

Unless GPUs work markedly differently somehow or there’s been some fundamental shift in computer architecture I’m not aware of, spatial locality is still a factor in computers. Aside from HW acceleration today, designs like Cebras would benefit heavily by reducing the amount of random access from accessing the weights (and thus freeing up cross-chip memory bandwidth for other things).

That could explain compute efficiency, but has nothing to do with the parameter efficiency pointed at in the paper.

Re: TopoNets: High performing vision and language models with brain-like topography

#19

Post by ziofill »

energy123 wrote: post #5

The main reason topography emerges in physical brains is because spatially distant connections are physically difficult and expensive in biological systems. Artificial neural nets have no such trade-off. So what's the motivation here? I can understand this might be a very good regularizer, so it could help with generalization error on small-data tasks. But hard to see why this should be on the critical path to AGI. A…

Perhaps they are more easily compressible? Once a bunch of nearby weights have similar roles one may not need all of them.

Re: TopoNets: High performing vision and language models with brain-like topography

#20

Post by mayukhdeb »

xpl wrote: post #6
energy123 wrote: post #5

The main reason topography emerges in physical brains is because spatially distant connections are physically difficult and expensive in biological systems. Artificial neural nets have no such trade-off. So what's the motivation here? I can understand this might be a very good regularizer, so it could help with generalization error on small-data tasks. But hard to see why this should be on the critical path to AGI. A…

> So what's the motivation here? Better interpretability, I suppose. Could give insights into how cognition works.

The motivation was to induce structure in the weights of neural nets and see if the functional organization that emerges aligns with that of the brain or not. Turns out, it does -- both for vision and language.

The gains in parameter efficiency was a surprise even to us when we first tried it out.

Post reply on HN