This is the last post in a mini-series on designing Gitlab CI/CD pipelines. We’ve discussed the basic anatomy of a .gitlab-ci.yml file, how to set up authentication tokens and files for building and pushing packages to a registry, and designing a Dockerfile for building images from a package in the context of a CI/CD pipeline.

This is the third post in a mini-series on designing Gitlab CI/CD pipelines. In the last post, we discussed setting up your .pypirc and .netrc files in the context of a Gitlab CI/CD pipeline to enable building and pushing packages to a package registry, as well as for installing code from a private registry.

This is the second post in a mini-series on designing Gitlab CI/CD pipelines. In order to build packages and push them to a remote package registry, we use the build and twine packages. build generates a package, and twine pushes this package to a registry (or “index”).

I recently developed a template workflow to help our team adopt a CI/CD-based development strategy. Many of our web applications and tools were based on simple repository structures. With growing datasets and ever-increasing use by outside teams, we found ourselves needing to add new features more frequently to many of these tools and believed that continuous integration and deployment could help us not just develop more quickly, but also more intelligently.

I was recently tasked with examining databases related to some computer vision tools that my company had acquired. Basically, the framework was as follows… Clients/users would sign up for some service with the goal in mind of building a model to classify a set of microscopy images.

We examine the utility of graph neural networks for the purpose of learning cortical segmentations. We show that attention-based transformer networks significantly outperform conventional GCN and linear feed-forward variants for the purpose of generating accurate reproducible cortical maps.

Using non-linear dimensionality reduction of functional brain connectivity patterns, and multivariate spatial statistics to characterize the functional embeddings, we analyze the spatial relationships between pairs of cortical regions to better examine how pairs of cortical regions connect and relate to one another.

In this paper, we develop a novel method based on dynamic mode decomposition (DMD) to extract resting-state networks from short windows of noisy, high-dimensional fMRI data, allowing RSNs from single scans to be resolved robustly at a temporal resolution of seconds. This automated DMD-based method is a powerful tool to characterize spatial and temporal structures of RSNs in individual subjects.

In this analysis, we propose the use of a library of training brains to build a statistical model of the parcellated cortical surface to act as templates for mapping new MRI data.

We present a framework for registering cortical surfaces based on tractography-informed structural connectivity. We define connectivity as a continuous kernel on the product space of the cortex, and develop a method for estimating this kernel from tractography fiber models.