Michiel Stock

Results 40 comments of Michiel Stock

```julia julia> quiver([1,2], [3,4],quiver=((1, 2), (-1, 1))) julia> yaxis!(flip=true) ``` Gives

I have implemented Dijkstra for my optimization course. If you want I can implement this for this package? Would a [minimum spanning tree](https://en.wikipedia.org/wiki/Minimum_spanning_tree) of a weighted network be relevant here?

Dijkstra can be implemented implemented easily with a [time complexity](https://en.wikipedia.org/wiki/Dijkstra%27s_algorithm#Running_time) of O(|V|^2) without additional data structures. If the O(|E|+|V|\log |V|) version is needed, I will need a heap from DataStructures.jl....

At this point, I implemented Dijkstra from source to all nodes. It is substantially faster than Bellman-Ford for larger (> 100 nodes) networks. I also implemented a version of Dijkstra...

This is a large body of research using Kronecker factorizations in GP/kernel ridge regression, e.g. [GPatt](https://arxiv.org/pdf/1310.5288.pdf) and [Kronecker ridge regression](https://www.mitpressjournals.org/doi/abs/10.1162/neco_a_01096?journalCode=neco). Would love to discuss this or contribute!

Yes! I am here the whole Hackathon. Would like to get involved 🙂 On Thu, 25 Jul 2019, 16:48 willtebbutt, wrote: > Would be great to both discuss and collaborate...

I found it, based on the explanation in the bifurcation settings. My solution was (for a slightly different system): ```julia prob_sens = ODEForwardSensitivityProblem(oprob.f, [0.0, 0.1], tspan, [1.7, 29, 50, 10,...

I think [TensorOperations](https://jutho.github.io/TensorOperations.jl/stable/) does something like that. I am a bit hesitant to allow hidden matrices to be generated for this package, though it might make sense to have versions...

Your code does not seem to work for some examples and gives the wrong result for others. For example: ```julia gradient((A, B)->sum(A⊗B), A, B) gradient((A, B)->sum(kron(A,B)), A, B) ``` Most...

Kronecker is likely easier to use if you are just doing matrix algebra. TO is a bit more general and advanced and can likely attain somewhat higher performance. Take a...