Independent scholarship · Open accessThe monthly digest

Indexed text attachment

differential-topology-of-parametrized-neural-networks.tex

Source for Differential Topology of Parametrized Neural Networks

View paper · Download original file

Lines 954–1053 of 1074. Source text is displayed literally.

954For a sphere of radius \(R>0\), rescaling replaces the positive955entries by \(R\) and \(2R\), respectively. The bound holds for every956parameter of an appropriate family. Its proof chooses an957antipodal pair separately at each parameter and needs no958continuous choice of witnesses.959 960Uniform and average losses have different implications. For example,961at width \(n\) choose stereographic coordinates \(\sigma\) from the962complement of the north pole of \(S^n\) to \(\mathbb R^n\).963Multiply \(\sigma\) by a smooth cutoff equal to one outside a small964polar cap \(C\) and equal to zero near the pole, extending the965encoder by zero there. The smooth decoder \(d=\sigma^{-1}\) gives966exact recovery off \(C\). For normalized spherical measure \(\mu\),967the squared error averaged over the sphere is at most \(4\mu(C)\),968which can be made arbitrarily small while the uniform error is \(2\).969Thus the uniform obstruction does not force a positive lower bound970on this average across all smooth maps of width \(n\).971 972A uniform Lipschitz bound prevents this concentration. If the973reconstruction has Lipschitz constant at most \(L\) with respect to974chord distance, its error function has Lipschitz constant at most975\(1+L\). For ambient width \(m\le n\), some point has error at least976\(1\), so every point within chord distance \(1/(2(1+L))\) of it977has error at least \(1/2\). Its mean squared error is at least978one quarter of the normalized measure of such a cap, a positive979bound at fixed \(L\). This estimate states the additional analytic980hypothesis under which the pointwise obstruction controls981a population loss.982\par983\section{Conclusion}984 985The analysis separates three conditions on a parametrized network.986The available weight directions determine where the total jet987map is transverse. The symmetry representation determines which988normal derivatives a fixed realization can have at a stabilizer.989The topology of the domain and latent factorization determines990which global reconstructions are possible. A model can satisfy991one of these conditions while exhibiting an obstruction governed992by another.993 994Examples on the circle and torus make these distinctions explicit.995Ordinary point counts jump at the fold discriminant. They can996also remain constant while boundary sheets are permuted along a997regular parameter loop. The resulting persistence module over998the parameter retains that permutation representation. In the999equivariant torus model the squared residual controls normal1000movement, while the tangent complex records an excess target1001direction. The autoencoder estimates bound uniform reconstruction1002error and give a separate mean-square bound under a Lipschitz1003constraint.1004 1005Relative evaluation and descent explain how these calculations1006pass between local presentations of a model. The assumptions1007used for the calculations are geometric: smoothness of the1008relevant charts, specified perturbation directions, properness1009for topology of fibers, and an explicit choice of decoder range.1010They apply to a realized architecture once those hypotheses1011have been checked.1012\par1013\balance1014\begin{thebibliography}{99}1015\small\raggedright1016 1017\bibitem{higher-hirsch}1018Anonymous. \emph{Differential Topology in Higher Categories:1019From Regular Values and Cobordism to Derived Smooth Families}.1020Submitted manuscript, 2026.1021 1022\bibitem{previous-networks}1023Anonymous. Representation theory of parametrized neural models1024on spaces. \emph{Higher Categorical Semantics}, 2026.1025\href{https://higher-categorical-semantics.pages.dev/research/representation-theory-of-parametrized-neural-models-on-spaces-ae10d3ed}{Published article}.1026 1027\bibitem{batson-autoencoders}1028J. Batson, C. G. Haaf, Y. Kahn, and D. A. Roberts.1029Topological obstructions to autoencoding.1030\emph{Journal of High Energy Physics} 04 (2021), 280.1031\href{https://doi.org/10.1007/JHEP04(2021)280}{DOI}.1032 1033\bibitem{chazal-persistence}1034F. Chazal, V. de Silva, M. Glisse, and S. Oudot.1035\emph{The Structure and Stability of Persistence Modules}.1036Springer, 2016.1037\href{https://doi.org/10.1007/978-3-319-42545-0}{DOI}.1038 1039\bibitem{cohen-steiner-stability}1040D. Cohen-Steiner, H. Edelsbrunner, and J. Harer.1041Stability of persistence diagrams.1042\emph{Discrete \& Computational Geometry} 37 (2007), 103--120.1043\href{https://doi.org/10.1007/s00454-006-1276-5}{DOI}.1044 1045\bibitem{fong-spivak-tuyeras}1046B. Fong, D. I. Spivak, and R. Tuy\'eras.1047Backprop as functor: A compositional perspective on supervised learning.1048In \emph{Proceedings of LICS}, 2019.1049\href{https://hdl.handle.net/1721.1/123513}{Author manuscript}.1050 1051\bibitem{hatcher-at}1052A. Hatcher. \emph{Algebraic Topology}.1053Cambridge University Press, 2002.