neurodata / neurodata.io

Homepage for NeuroData
https://neurodata.io
Other
12 stars 17 forks source link

NeuroData Active PR List #570

Open jshinm opened 2 years ago

jshinm commented 2 years ago

The following is the active* list of current PR in progress (last updated 12/22/2021)

*Activity since 09/01/2021

<html xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:x="urn:schemas-microsoft-com:office:excel" xmlns="http://www.w3.org/TR/REC-html40">

Lead | Repository | PR Name | PR Url -- | -- | -- | -- adam2392 | scikit-learn | [TEST PR] Adding oblique trees (i.e. Forest-RC) to cythonized tree module | https://github.com//neurodata/scikit-learn/pull/11 adam2392 | scikit-learn | [TEST PR] Oblique forests | https://github.com//neurodata/scikit-learn/pull/10 KhelmholtzR | ProgLearn | Add Streaming Function | https://github.com//neurodata/ProgLearn/pull/526 parthgvora | ProgLearn | UF enhanced parameters to match GRF | https://github.com//neurodata/ProgLearn/pull/453 adam2392 | scikit-learn | Tom/grid to graph 26 | https://github.com//neurodata/scikit-learn/pull/8 LizaNaydanova | ProgLearn | Added neural network scene segmentation tutorial. | https://github.com//neurodata/ProgLearn/pull/527 KevinWang905 | ProgLearn | unsupervised contrastive loss | https://github.com//neurodata/ProgLearn/pull/522 Dante-Basile | ProgLearn | Supervised Contrastive Loss | https://github.com//neurodata/ProgLearn/pull/518 waleeattia | ProgLearn | Exploring task unaware vs aware BTE and generalization error experiment | https://github.com//neurodata/ProgLearn/pull/517 kfenggg | ProgLearn | Add notebook & functions for adversarial domain adaptation | https://github.com//neurodata/ProgLearn/pull/464 tliu68 | ProgLearn | Adding KNNClassificationDecider | https://github.com//neurodata/ProgLearn/pull/463 latasianguy | ProgLearn | Draft of SPORF | https://github.com//neurodata/ProgLearn/pull/454 jzheng17 | df-dn-paper | Implementation of Hyperparameter Tuning for the VIsion and Audition Folders | https://github.com//neurodata/df-dn-paper/pull/64 jzheng17 | df-dn-paper | hyperparamter tuning | https://github.com//neurodata/df-dn-paper/pull/62 sampan501 | hyppo | update MGC to use scipy as reference implementation in tests | https://github.com//neurodata/hyppo/pull/231 sampan501 | hyppo | Added kernel goodness of fit | https://github.com//neurodata/hyppo/pull/206 nhahn7 | SDTF | ENH Prepare SynF for Gaussian Simulation experiments (updated code) | https://github.com//neurodata/SDTF/pull/42 tathey1 | brainlit | Update config.yml | https://github.com//neurodata/brainlit/pull/328 tathey1 | brainlit | Renaming lddmm registration files | https://github.com//neurodata/brainlit/pull/320 dallen02 | brainlit | time sleep | https://github.com//neurodata/brainlit/pull/295 tathey1 | brainlit | get angle method | https://github.com//neurodata/brainlit/pull/241 mlemyco | brainlit | multiple radius | https://github.com//neurodata/brainlit/pull/114 Laknath1996 | kdg | Label Noise Experiments for KDN | https://github.com//neurodata/kdg/pull/23 dependabot | talks | Bump node-sass from 3.3.3 to 4.14.1 in /big-data-science | https://github.com//neurodata/talks/pull/28 dlee0156 | bilateral-connectome | Adding test comparing triangle subgraph counts | https://github.com//neurodata/bilateral-connectome/pull/11 Lawreros | m2g | update func output tree | https://github.com//neurodata/m2g/pull/408 Lawreros | m2g | [WIP] Bunch of bug fixes and features | https://github.com//neurodata/m2g/pull/398 kellymarchisio | goat | Fixes #1 - fix outlier handling in qap/qapot | https://github.com//neurodata/goat/pull/2 JBaijot | neuroparc | Add files via upload | https://github.com//neurodata/neuroparc/pull/40 jovo | neuroparc | updated readme trivially | https://github.com//neurodata/neuroparc/pull/24 adam2392 | SPORF | [WIP] An implementation of discontiguous sampling of the SRerf variant; we call it MTORF(?) | https://github.com//neurodata/SPORF/pull/353 adam2392 | SPORF | Bic fix urerf | https://github.com//neurodata/SPORF/pull/350 megh1241 | SPORF | Add Cython USPORF and outlier detection example | https://github.com//neurodata/SPORF/pull/349 MaiRajborirug | SPORF | update maintainer and update version to 2.0.5 | https://github.com//neurodata/SPORF/pull/347 MrAE | SPORF | Several suggestions on the distance metrics of SPORF that may boost the performance on geodesic learning | https://github.com//neurodata/SPORF/pull/346 ShumingXu | SPORF | update python to dev v2.0.5.1 for clarity | https://github.com//neurodata/SPORF/pull/342 MrAE | SPORF | [WIP] Add two feature sampling methods for graphs | https://github.com//neurodata/SPORF/pull/337 j1c | SPORF | fix string -> int | https://github.com//neurodata/SPORF/pull/336 dfrancisco1998 | notebooks | Simulations from papers for minka ZG comparison | https://github.com//neurodata/notebooks/pull/13 dfrancisco1998 | notebooks | Comparing GraphSAGE to CASC | https://github.com//neurodata/notebooks/pull/12 zeou1 | notebooks | ASE/LSE vs SVD/Spectral Comparison | https://github.com//neurodata/notebooks/pull/11 kareef928 | notebooks | Add experiments on spherical clusterings of DCSBM embeddings | https://github.com//neurodata/notebooks/pull/10 PerifanosPrometheus | notebooks | ZG vs minka comparison using OpenMl-cc2018 | https://github.com//neurodata/notebooks/pull/8 dfrancisco1998 | notebooks | clean up homepage, remove about | https://github.com//neurodata/notebooks/pull/4

jovo commented 2 years ago

@jshinm the priority functionalities are listed below, along with the POC (point of contact) for them:

stuff into sklearn

LDDMM into skimage (@vikramc1, @DevinCrowley)

@asaadeldin11 did i miss anything?

jovo commented 2 years ago

@jshinm please make a schedule with each of the POC's to complete the PRs.
of course, the timeline must be somewhat flexible due to its dependence on the core developers. but our turn around time is not dependent on them, and that is the main factor that we are concerned about.

jshinm commented 2 years ago

@jovo Thank you for this comprehensive list. I will go ahead and get in touch with each of the POC and start making concrete plans and rough timelines for each project. Up to this point, we've made some progress on oblique tree where I ran a benchmark for @adam2392 and he'll use that to push forward the next round of PR into sklearn. I will follow up with Adam and the rest of POCs when I get back on the 5th