changelog
- Tue, 21 Mar 2023 20:31:01 +0200
- by Tuomo Valkonen <tuomov@iki.fi> [Tue, 21 Mar 2023 20:31:01 +0200] rev 25
- Implement non-negativity constraints for the conditional gradient methods
- Sun, 11 Dec 2022 23:25:53 +0200
- by Tuomo Valkonen <tuomov@iki.fi> [Sun, 11 Dec 2022 23:25:53 +0200] rev 24
- Support arbitrary regularisation terms; implement non-positivity-constrained regularisation.
* Fixes the conditional gradient methods that were incorrectly solving
positivity constrained subproblems although the infinite-dimensional
versions do not include such constraints.
* Introduces the `ExperimentV2` struct that has `regularisation` in place
of `α`. The `Experiment` struct is now deprecated.
* The L^2-squared experiments were switch to be unconstrained, as the
Franke-Wolfe implementations do not support constraints. (This would
be easy to add for the “fully corrective” variant, but is not
immediate for the “relaxed” variant.)
