Aurora Rossi

Results 20 issues of Aurora Rossi

issue #34, now `is_cyclic` works for undirected connected graphs.

Now both algorithms return a `Vector{Edges}`. I chose `Vector{Edges}` instead of `Vector{SimpleWeightedEdge}` because in this way the code is faster. In the case that weights are needed it is easy...

do not merge

I noticed that in some checks of my pull request the tests of ` \traversals\diffusion.jl` fail. I didn't modify that file. The first time that happened was in Julia 1...

With this PR I add the Bernoulli random graph and the $\rho$-correlated Bernoulli random graphs generators.

enhancement

See the description of PR https://github.com/JuliaGraphs/Graphs.jl/pull/194 by @SimonCoste

enhancement

Hello! when downloading the datasets contained [here](https://graphmining.ai/temporal_datasets), there is an error caused by the certificate of the website. ``` urllib.error.URLError: ```

Hello, I saw that the implementation of `A3T-GCN` is not the same as the [original paper](https://arxiv.org/pdf/2006.11583.pdf). Are there any particular reasons in the choices made?

Hello, when implementing the model of the paper https://arxiv.org/pdf/1811.05320.pdf why do you use a `GCNConv` for each `update` and `reset` gate and `candidate` state? Why not just take the output...

With this PR I add the `topk_nodes` and `topk_edges` features (see issue https://github.com/CarloLucibello/GraphNeuralNetworks.jl/issues/41).

As described for heterographs in the issue #388, `TemporalSnapshotGNNGraph`s also need adapted pooling. List of existing poolings that need to be extended: - [ ] GlobalPool - [ ] GlobalAttentionPool...

temporalgraphs