From 6a51b62293b994ac7223c9926a73a8f7f7b1c670 Mon Sep 17 00:00:00 2001 From: Patrick Nicodemus Date: Fri, 17 Jan 2025 10:21:57 -0500 Subject: [PATCH] Updated documentation --- README.md | 7 +++++++ docs/gw_variants.rst | 1 + 2 files changed, 8 insertions(+) diff --git a/README.md b/README.md index 206053a..3f2e8cc 100644 --- a/README.md +++ b/README.md @@ -53,3 +53,10 @@ Extensive documentation, including several tutorials, can be found in [CAJAL's r ## New in this release Version 1.5 of CAJAL incorporates two significant new pieces of functionality for the analysis of neuron morphology, unbalanced Gromov-Wasserstein (a matching tool which allows for partial or incomplete matches) and fused Gromov-Wasserstein (a tool which allows the incorporation of additional feature data about neuron nodes, e.g., soma or dendrite.) +These are discussed in detail in the documentation [in a worked example tutorial](https://cajal.readthedocs.io/en/stable/notebooks/Example_5.html) and, at a more theoretical level, [here](https://cajal.readthedocs.io/en/stable/gw_variants.html). + +- Cleaned up function signatures - making function signatures more uniform between Euclidean and geodesic functions, removing unused parameters +- Removed source of nondeterminism/non-reproducibility in Louvain clustering algorithm +- Added optional dependencies for usage in the Docker image +- Restored parallelism to SWC sampling +- Fixed bug in implementation of Weighted Nearest Neighbors algorithm diff --git a/docs/gw_variants.rst b/docs/gw_variants.rst index c2b678a..fdae63a 100644 --- a/docs/gw_variants.rst +++ b/docs/gw_variants.rst @@ -155,6 +155,7 @@ We let :math:`(X,\mu)`, :math:`(Y,\nu)` and :math:`\mathcal{G}`` be as before. Classical Gromov-Wasserstein treats cells purely geometrically, as shapes. In searching for good alignments between two neurons, it doesn't consider some important information present in cell morphology reconstructions, such as the labels for the soma and dendrite nodes. On biological grounds, it is reasonable to argue that a "good alignment" between two neurons should align the soma node to the soma node, align axon to axon, basal dendrites to basal dendrites, and apical dendrites to apical dendrites. Fused Gromov-Wasserstein is a construct which modifies classical Gromov-Wasserstein to add a penalty term for transport plans which align nodes of different types. By making the penalty term large, we can bias the search algorithm towards transport plans which reflect the additional information available in the cell structure. +For more details please see the [Fused Gromov-Wasserstein paper](https://arxiv.org/abs/1811.02834) due to Vayer, Chapel, Flamary et. al. The formula for the fused GW cost of a transport plan is