HIGHLIGHTS
- who: Marcus M. Noack from the Berkeley National Laboratory University at have published the research: Exact Gaussian processes for massive datasets via non-stationary sparsity-discovering kernels, in the Journal: Scientific Reports Scientific Reports of 10/06/2022
- what: The authors show that, by combining tailored kernel designs, HPC implementation, and constrained optimization, exact GPs can be scaled to datasets of any size, under the assumption of naturally-occurring sparsity. Instead of formulating kernels that learn well what points are dependent, the authors propose to consider kernels that are tailored to be capable of learning . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.