No specific ML-related updates.

Variable classification effort: Charter drafted, sent to PI for review. Waiting.

DR1 is a week away.

Dima - 7 minutes to process a nights data on TPU

Jan: DR1 - cross matches with Gaia

Test: known CRTS RR Lyrae. 90% recovery

Outliers need to be looked at.

2 competing philosophies. Statistics, automatic classifications. domain knowledge.

any number of features

initial catalog - bright stars only (vsx, assasin)

Thomas: percentages (scores, probabilistic classifications)

Then to fainter stars.

Matthew: Trial area.

Ashish: cross match CRTS periodic variable

Ashish: same objects in different quadrants will have to be dealt with

Jan: use random sets from ZTF.

Ashish - datasets

Jan: unique ZTF ids?

Matthew: No - chips, quads, filters have different identifiers

Master source catalog needs to be created.

Eric: depuplicated merge done at UW and being sent to Dima

Matthew: that excludes Caltech data

Eric: isn't that to be excluded from current project?

Matthew: yes, we will then need to create a superset with Caltech data

Dima: naming scheme to be decided. Then we can see if caltech objects are separate

Matthew: there could be i-band dropouts

Brigitta: Nothing

Eric: literature review of features popped up

David: waiting for access. Jan?

Jan: light curves quality and so on

Scott: no

Thomas: no

Umaa: RF work to be written up

Matthew: data validation, classification schemes

Richard: no

Dima: nothing further

Ashish: hackathon

Jan: tasks were assigned including literature review; will send a summary and list

Minutes 2019-05-02

Attendees:

Project Updates:

Pipeline updates:

RB updates:

Zooniverse updates:

Creating datasets, variability catalog:

Papers:

Other topics:

-- Ashish Mahabal - 2019-05-01

Comments

<--/commentPlugin-->