Explore >> Select a destination
|
You are here 
|
andrewkchan.dev | ||
| | | | |
nickmcd.me
|
|
| | | | | ||
| | | | |
blog.kummerlaender.eu
|
|
| | | | | Adrian Kummerländer's blog on software development, linux and open source | |
| | | | |
mht.wtf
|
|
| | | | | [AI summary] The text discusses an optimization technique for solving linear systems in a simulation context, particularly when dealing with large, sparse matrices. By recognizing that the Hessian matrix is not actually dense but can be decomposed into a sparse matrix plus a rank-1 update, the problem is transformed from a dense solve into two sparse solves, significantly improving performance. This approach leverages the Sherman-Morrison formula and linear algebra techniques to avoid constructing a dense matrix, thereby reducing computational time and resource usage. The benefits of this method are demonstrated through micro-benchmarking in Julia, showing a substantial speedup for increasing problem sizes. The text emphasizes the importance of analyzing the... | |
| | | | |
www.khanna.law
|
|
| | | You want to train a deep neural network. You have the data. It's labeled and wrangled into a useful format. What do you do now? | ||
