A giant PACMAN

Thingstätte

SISC published our paper on RESQPASS where subspaces are combined with constrained optimization to solver the Bounded Variable least squares problem. After a few iterations, where the bounds are discovered, the method behaves as a Krylov subspace. epubs.siam.org/doi/abs/10.1...

Hoboken

Baltimore

Washington - Judiciary square. There is something broken here.

Baltimore

these masses next to this building in Santiago de Chile are they earthquake dampers?

downtown Hoboken

J-class in southampton

What is more destructive? en.wikipedia.org/wiki/Monte_K... or en.wikipedia.org/wiki/Hambach...

Gaston Flater.

ESCOL at bois du cazier

Grateful to Householder committee for the opportunity to explain the ResQPASS algorithm at Householder symposium XXII at Cornell University --- Ithaca, NY USA. The paper, co-auther with Bas Symoens ( arxiv.org/abs/2302.13616 ), has been accepted for publication in SISC.

Is this hat "conclave" or "convlex"?

Do we need to start Flying universities? en.wikipedia.org/wiki/Flying_...

"Statistical Terror" or "statistisch schrikbewind". There was a period where there was a "statistisch schrikbewind" nl.wikipedia.org/wiki/Statist...

Folks, we're living in a Second Gilded Age. It can be easy to feel discouraged, but the First Gilded Age teaches us lessons on how we can fight back. Watch.

ResQPASS in the outer loop is very similar to GMRES or LSQR. But it can also deal with bounds on the variables, in contrast to these traditional Krylov subspace methods arxiv.org/abs/2302.13616 #RESQPASS

The ResQPASS algorithm creates a subspace similar to Krylov subspaces. But is more general since it can deal with bounds on the variables.