-
@vpuri3 this is the example code using https://docs.sciml.ai/SciMLOperators/dev/tutorials/fftw/
as a starting point for an fft pde (updated)
```julia
using SciMLOperators, LinearAlgebra, FFTW, O…
-
- [ ] [Is Cosine-Similarity of Embeddings Really About Similarity?](https://arxiv.org/html/2403.05440v1)
# Is Cosine-Similarity of Embeddings Really About Similarity?
**DESCRIPTION:**
Is Cosine-Si…
-
As far as I can tell, there is currently no way to do NMF (or other decomposition methods) with missing data. This could be useful for at least a couple of reasons:
1. Data imputation: Using factor…
-
I would like to tune hyper-parameters with implicit's `AlternatingLeastSquares`. Ideally, I would use cross-validation but it seems like there is no simple way to "fit" on training data and "predict" …
-
I'm trying to implement both recommendations and item-item similarities based on an `ImplicitFactorizationModel` trained on clickstream data. The former is trivial using the `predict` method, but the …
mszep updated
6 years ago
-
SciJS doesn't yet have a good linear solver. At the moment, we have the following simple package (thanks @substack) which works ok for smaller systems using an LU decomposition:
- https://www.npmjs.c…
-
I dug through the code and DifferentialEquations.jl could add quite a bit to the efficiency of the time stepping code. For reference, what it would really add is SSP methods:
http://docs.juliadiffe…
-
We have a quick implementation of ADI:
https://github.com/dlfivefifty/AlternatingDirectionImplicit.jl
This could live in this package.
-
**Submitting author:** @https://github.com/josemanuel22 (José Manuel de Frutos)
**Repository:** https://github.com/josemanuel22/ISL
**Branch with paper.md** (empty if default branch):
**Version:** v0…
-
@fprimeau As we discussed might happen, I can't even factorize the matrix on my laptop. FWIW, this is the sparsity pattern, which looks OK (but, yes quite big, with 10'000'000 non-zeros...):
![Scre…