From ce529bb877d19fed54dc3d887f9fb0ddcb8ac875 Mon Sep 17 00:00:00 2001 From: Bas Nijholt <basnijholt@gmail.com> Date: Wed, 31 Jul 2019 14:17:21 +0000 Subject: [PATCH] add some remarks made in the first discussion --- plan.md | 15 +++++++++++++++ 1 file changed, 15 insertions(+) create mode 100644 plan.md diff --git a/plan.md b/plan.md new file mode 100644 index 0000000..14c719e --- /dev/null +++ b/plan.md @@ -0,0 +1,15 @@ +# Adaptive + +We propose to use local criteriums for sampling combined with global updates. +This defines a family of straightforwardly parallelizable algorithms which are useful for intermediary cost simulations. + +When your function evaluation is very expensive, full-scale Bayesian sampling will perform better, however, there is a broad class of simulations that are in the right regime for Adaptive to be beneficial. + +We can include things like: +* Asymptotically complexity of algorithms +* Setting of the problem, which classes of problems can be handled with Adaptive +* Loss-functions examples (maybe include [Adaptive quantum dots](https://chat.quantumtinkerer.tudelft.nl/chat/channels/adaptive-quantum-dots)) +* Trials, statistics (such as measuring timings) +* Line simplification algorithm as a general criterium +* Desirable properties of loss-functions +* List potential applications -- GitLab