- Leo: For snapshot bug, I talked to Rachael about kill signal and now know what to do going forward. But didn’t have chance to do... Will do by next week
* [x] Make plots for IMBH stitch bank
- Shio: Debnandini, Chad, met and discussed
generate a bank that is an extension to the all sky bank
mass ratio is extended to 20 instead of 10
chad had concerns about freq
Deb will look into that, negative chi values.
Can close action item
- Jim: utility function to stich two banks together?
- Shio: not yet, template ids needs to be adjusted
* [x] Make sure focus session is at the beginning of next week's call for Stefano
## Agenda / Minutes
* Announcements (5 minutes)
- Please check the rota for next week's call
- Shio: working meeting, not official, focus sessions push to Jan.
- Confirmation of next week's focus session
- O4a rota: Sign up [here](https://git.ligo.org/gstlal/projects/-/wikis/O4a-GstLAL-Analysis-ROTA)
* Focus session
- Stefano presented results from precessing searches
- VT improvement in high q and high chi_p region
- Leo: distribution of injection set?
- Debnandini: How are you sorting templates for SVD?
- Stefano: mu sorting has best SVD compression (~0.01)
- Leo: Why V1 has less scatter in inj SNR vs rec. SNR plot?
- Stefano: have not looked into that too much
- Shio: Explore changing values in making SVD bank
- Leo: use mu3?
- Shio: flow? max duration cut? might lower the number of templates?
- Stefano: templates are short
- Jim: metric evaluation?
- Stefano: https://arxiv.org/abs/2302.00436
- Ryan: unlikely LVK will do something on O4 time scale. Run independently and find a precessing signal that the collaboration didn’t find.
- Sarah: PyCBC NSBH precession?
- Ryan: not sure about actual ongoing works
- Chad: how big is the template for it to be meaningful?
- Stefano: 1M at least
- Chad: Banksim check? How much are we missing?
- Chad: target highest discovery potential
- Leo: Scatter in inj SNR vs. dec. SNR match the high q parameter space? Check SNR below 4 signals?
- Chad: bank is perfect, run on OSG, plausible to make discoveries
- Stefano: how to make injections
- Sarah: talked with maya fishbach about this
- Ryan: use gwtc-2 or -3 injections? already precessing
- Chad: choose injections that are more or less in the bank
- Cody: When do you need comments by?
- Stefano: second week of Jan?
- Cody votes yes to make this a GstLAL paper
- Chad: Making discovery is important. how to proceed? where to put effort in?
- Stefano: making injection sets are straight forward, most important is computing power, not experienced with OSG. one chunk in a week?
- Chad: HL only search? V add not much to VT
* Last week's East call ([link](https://wiki.ligo.org/CBC/Searches/GstLALEastAgenda20231207))
- Shio: presented an MR. this week’s east call is last in the year
* Quick updates (45 minutes)
- Operations (5 minutes)
- LL CBC operations
- Rachael: Quiet week. L out of lock since power outage last thurs. Tracked down bug in SNR optimizer, that causes plots not uploaded. Affecting Edward and EW. Fixed during maintenance. Bug fix needs to go in. Prathamesh and Becca can handle that
- O4 Dev (30 minutes)
- Low latency integrated testing and Monitoring
- Shomik: Working on fixing Sankey diagram. Event labels in SNR optimizer
- O4b online production lead team updates ([issue](https://git.ligo.org/gstlal/projects/-/issues/795))
- AllSky
- Shomik: need help getting Bob run on Nemo. Rick is running. Testing online extinction model. bob is edward on NEMO. No retractions in MDC 24
- Amanda: have trouble getting bob to run all the jobs at the same time. Taken by OSG resources? Talking to Duncan. EW is having same issues as other analyses
- Chad to Amanda: What the computing config is such that priority matters? Normally: dedicated slots. Come to PSU’s CI meeting?
- EW
- Leo: background plot in EW?
- Amanda: cannot see the plots
- Leo: just run background plotter program, doesn’t need to be uploaded, make sure background plot is working
- Cort: unrelated, going to merge get_env at tomorrow’s east call.
- SSM
- IMBH
- GWTC-4
- Computing
- Likelihood ratio, background and foreground sampling
