SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2312.08550 · 2023

Harmonics of Learning: Universal Fourier Features Emerge in Invariant Networks

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 9 functions out of this paper's own repositories and ran 6 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
sophiaas/spectral-universality canonical 6 of 9
FunctionStatusWhere it lives
complex_array_to_rgb Ran sophiaas/spectral-universality/utils.py
pointer only (licence: NONE) · get_code("07f4e9b34e80e048")
direct_product Ran sophiaas/spectral-universality/groups.py
pointer only (licence: NONE) · get_code("6b6fd1f38b462a0c")
init_weights Ran sophiaas/spectral-universality/models_JAX.py
pointer only (licence: NONE) · get_code("b084e5a3eb7bc0c1")
matmul_complex Ran sophiaas/spectral-universality/models_torch.py
pointer only (licence: NONE) · get_code("e60f9d31ff4a53fa")
pad_eye Ran sophiaas/spectral-universality/models_torch.py
pointer only (licence: NONE) · get_code("a853a6fa01870647")
perm_matrices Ran sophiaas/spectral-universality/utils.py
pointer only (licence: NONE) · get_code("8ffe99d5ad67b54c")
pad_eye Not yet run sophiaas/spectral-universality/models_JAX.py
pointer only (licence: NONE) · get_code("3de5da36457066c7")
perm_frobenius Not yet run sophiaas/spectral-universality/utils.py
pointer only (licence: NONE) · get_code("d825189f35316425")
total_weight Not yet run sophiaas/spectral-universality/models_JAX.py
pointer only (licence: NONE) · get_code("14cdeef2a19d88e0")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

In this work, we formally prove that, under certain conditions, if a neural network is invariant to a finite group then its weights recover the Fourier transform on that group. This provides a mathematical explanation for the emergence of Fourier features -- a ubiquitous phenomenon in both biological and artificial learning systems. The results hold even for non-commutative groups, in which case the Fourier transform encodes all the irreducible unitary group representations. Our findings have consequences for the problem of symmetry discovery. Specifically, we demonstrate that the algebraic structure of an unknown group can be recovered from the weights of a network that is at least approximately invariant within certain bounds. Overall, this work contributes to a foundation for an algebraic learning theory of invariant neural network representations.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2312.08550")
get_code_for_paper("2312.08550")
have("2312.08550")

Connect an agent — have() is free.