joshua vogelstein

did we include GMA in here yet? if not, lets?

1

https://ieeexplore.ieee.org/abstract/document/6247923 to solve the QCQP, we can potentially leverage https://scipy.github.io/devdocs/generated/scipy.optimize.quadratic_assignment.html

- [ ] benchmark against TensorFlow - [ ] plot performance vs. matrix size - [ ] plot performance vs # cores - [ ] ratio of relative performance across...

## Q2 2017 - [x] ase/laplacian supporting directed/undirected, etc. - [x] LOL - [ ] semipar (https://arxiv.org/abs/1403.7249) @youngser - [ ] idt (https://github.com/youngser/behaviotypes/blob/master/doidt.r) @mrae working on this code - [...

Two population of graph test

1

Let's say we have 2 populations of matched graphs, and we want to test whether elements of P or B are different across the two populations (P is the IER...

symmetrize binary graph

2

same as 'avg', but then round somehow? or make it clear that we report weighted graph?

compare CASC to GraphSAGE on the experiments they performed in their papers

8

https://arxiv.org/pdf/1706.02216.pdf

fix mismatch between 'simulation' and 'model'

3

seems like model means 'estimation'? also, documentation is somewhat incomplete. for example, i would think the simulation documentation would have sub-sections for each model, just like the models documentation?

add fast 2-sample HHG test

1

If we don't include the fast HHG independence test, which is fine, there is also a fast 2-sample test that should definitely be included _Originally posted by @sampan501 in https://github.com/neurodata/hyppo/issues/106#issuecomment-667273786_

default mtry and "projection type" for low-dimensional data

2

i never seriously considered defaults for really low dimensions. it occurs to me that it might make sense for mtry to never be much smaller than 100, and rather than...

see https://people.eecs.berkeley.edu/~brecht/papers/08.rah.rec.nips.pdf