Reference tune

[scarrazza]

We should take #1 and apply https://professor.hepforge.org/.

[scarrazza]

@mlazzarin to close this issue we should:

• push the reference tune runcard after merging New feature: Bins weighting #5.
• print the reference results here.

[mlazzarin]

I've written the results obtained with the runcards (or the weightfiles) that are in the folder reference_tune_AZ_Pythia8 in this draft, table 2:
Results_comparison.pdf

[alioli]

Non mi apre il PDF, è forse corrotto?

…

On Sat, May 4, 2019, 09:58 Marco Lazzarin ***@***.***> wrote: I've written the results obtained with the runcards (or the weightfiles) that are in the folder reference_tune_AZ_Pythia8 in this draft, table 2: Results_comparison.pdf <https://github.com/scarrazza/mcnntunes/files/3144105/Results_comparison.pdf> — You are receiving this because you were assigned. Reply to this email directly, view it on GitHub <#2 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAENHQHGL3E4XOIGKDNBOULPTU6YVANCNFSM4HCIAQPA> .

[mlazzarin]

Sul mio PC viene aperto. Ho provato a cambiare compilatore:
Results_comparison.pdf

[alioli]

Non va ancora. Sto provando dal telefono però. Proverò nel pomeriggio dal laptop

…

On Sat, May 4, 2019, 11:49 Marco Lazzarin ***@***.***> wrote: Sul mio PC viene aperto. Ho provato a cambiare compilatore: Results_comparison.pdf <https://github.com/scarrazza/mcnntunes/files/3144212/Results_comparison.pdf> — You are receiving this because you were assigned. Reply to this email directly, view it on GitHub <#2 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAENHQECZOEULEHVC5VTWPLPTVL2FANCNFSM4HCIAQPA> .

[mlazzarin]

Ok, se non funziona neanche sul laptop mi avvisi che tento in altri modi.

[scarrazza]

Thanks, the PDF is working fine for me.

[alioli]

Now is working fine for me too. The results seem solid. We should talk about how to proceed/wrap up the project. Next week I will be at DESY but I am reachable via Skype. Simone ᐧ

…

On Sat, May 4, 2019 at 12:59 PM Stefano Carrazza ***@***.***> wrote: Thanks, the PDF is working fine for me. — You are receiving this because you were assigned. Reply to this email directly, view it on GitHub <#2 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAENHQHEEAQPHQR7PFXDZH3PTVUBBANCNFSM4HCIAQPA> .

[scarrazza]

Simone, just to keep you updated, Marco is now implementing a closure test mechanism in mcnntune and professor in order to build a quality estimator. This will be extremely useful for testing and tuning the mcnntune output.

[alioli]

Ok, perfect. The only other missing part compared to Professor is the availability of Eigentunes, i.e. variations of 1 sigma in each direction with respect the best tune result. These are used (perhaps without a solid theoretical foundation) to estimate the uncertainty coming with the tune. Do we want to include something along these lines in MCNNTUNE ? ᐧ

…

On Sun, May 5, 2019 at 2:09 PM Stefano Carrazza ***@***.***> wrote: Simone, just to keep you updated, Marco is now implementing a closure test mechanism in mcnntune and professor in order to build a quality estimator. This will be extremely useful for testing and tuning the mcnntune output. — You are receiving this because you were assigned. Reply to this email directly, view it on GitHub <#2 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAENHQGAS6HEFXGUVCIQKFDPT3E6NANCNFSM4HCIAQPA> .

[scarrazza]

Yes, I think after the colsure implementation we have to provide reliable uncertainties.