Differences

This shows you the differences between two versions of the page.

--- private:teleconsnotes20181031 [2018/10/31 14:31] – hanany
+++ private:teleconsnotes20181031 [2018/10/31 16:00] (current) – kyoung
@@ Line 1: / Line 1: @@
 ====== Telecon Notes Oct. 31, 2018  ======
-Attendance:    \\
+Attendance: Nick B., Colin H., Charles L., Shaul, Raphael F., Jim B.  \\
 Notes by:  Karl Young \\
@@ Line 14: / Line 14: @@
       * Anything else? (for 12/31)
   * Key concerns: r
-    * well defined target
+    * well defined target; Feeney forecasts sigma(r)=1e-4; 5e-4 is 5sigma
     * foregrounds:
-      * State of Forecasts
+      * State of Forecasts; Baseline plan for Nov. 1
+        * [[https://zzz.physics.umn.edu/ipsig/foregroundstelecon20180830|Soumen analysis]] and {{:private:soumen_nilc_foregrounds.png?50|Figure}}
+        * [[https://www.dropbox.com/s/engghfxhidoe67q/Remazeilles_gnilc_pico.pdf?dl=0|GNILC]] and {{:private:mathieu_gnilc_foregrounds.png?50|Figure}}
+        * [[https://zzz.physics.umn.edu/ipsig/_media/remazeilles_commander_pico_psm.pdf|PSM]] and {{:private:commander_foregrounds_bb.png?50|Figure}}
       * need reviewers
+=== Notes ===
+Status of report
+JPL (Instrument)
+  * 1 chapter released. Amy says remainder (2-3) possible tomorrow.
+UMN (Science)
+  * Very close.
+    * NB: a few sentences to add in fundamental physics. running new cluster count constraints (dN/dz) now for PICO only case. Done in minutes, need to check results and would need some text added to explain.
+    * NB: dN/dz gives a nuetrino mass constraint. comparable precision to 15 meV (given model uncertainties) from lensing and tau.
+    * SH: what about dark energy?  Breadth is useful.
+      * NB: Just updated forecast calculation. not confident enough to say if PICO constraints at all comparable to LSST and similar.
+      * JB: suggest we at least say something. total absence might catch people's attention in the negative.  Even just a simple 'this is doable via cluster counts . . .' Something completely vanilla, non-quantitative is enough.
+      * NB: maybe can say something about high-z clusters can constrain sigma_8 across z. just 1 sentence.
+      * SH: agree with JB. even something small now, to be increased/deleted later is good.
+      * **NB to add a couple sentences on this**
+  * SH: currently nothing about CIB.  Recruited Olivier Dore to write 1/2 page (won't exist by Nov. 1 deadline).  Olivier has PICO specs.  Asked to do forecasts.
+    * **CL will call Olivier and apply pressure**  -- Success. **Olivier will write something by tonight**
+      * Olivier said it will be similar to Planck. (NB: why? CL: didn't ask. We'll see from his text.)
+  * Anything else? (for 12/31 or later)
+    * All: silence.  So we must have it all.
+Key concerns: r  SH: target not well defined in text and STM. Should refine.
+  * well defined target; Feeney forecasts sigma(r)=1e-4; 5e-4 is 5sigma
+    * RF: possible differences, lensing residual? Our noise levels agree very closely. so we should understand what the differences are.
+    * RF: also cross checked my number with Alex. Stephen's is different, should understand why.  Mostly larger lensing residual by Stephen. Should iron out what makes most sense.
+  * SH: for now propose quote: sigma(r) 1e-4, 5 sigma at r=5e-4.  Does this clearly reject a specific class of models?
+    * RF: problem is some of the models are open ended. so not a clear target that rejects an entire class of models.
+    * SH: to be clear.  so 1e-3 also doesn't reject entire class?
+      * RF: 1e-3 is nice because that is where Planck scale is characteristic scale (1e-3 rejects planck scale at 3sigma. sigma(r) is 5e-4. current S4 target).  But models exist below that as well. Looking to see if there is a similar line at 1e-4.  Going to 1e-4 is still interesting.  thinking about the field range is a likely angle.
+      * **RF will think to see if there is another argument to draw a line at our limits.**
+  * foregrounds: (SH: least mature section, only one that isn't complete.)
+    * State of Forecasts; Baseline plan for Nov. 1
+      * [[https://zzz.physics.umn.edu/ipsig/foregroundstelecon20180830|Soumen analysis]] and {{:private:soumen_nilc_foregrounds.png?50|Figure}}
+      * [[https://www.dropbox.com/s/engghfxhidoe67q/Remazeilles_gnilc_pico.pdf?dl=0|GNILC]] and {{:private:mathieu_gnilc_foregrounds.png?50|Figure}}
+      * [[https://zzz.physics.umn.edu/ipsig/_media/remazeilles_commander_pico_psm.pdf|PSM]] and {{:private:commander_foregrounds_bb.png?50|Figure}}
+    * Current status:
+      * Soumen has results. 85% delensed. NILC procedure. has the 6 maps from Clem. 5/6 have residuals below 3e-3.
+      * Mathieu results. GNILC. 4 maps done.
+      * Planning to add those plots to report.  Say results are encouraging. beginning of a long path.
+      * RF: Why 3e-3? not r=0 or lower r?
+        * SH: history. Clem started with 3e-3 and r=0 because that was from S4.  Soumen, Mathieu haven't got to r=0 case.  Asked Clem to do a lower r case.  Clem reluctant to put in more time when the current maps haven't all been used yet.
+        * RF: r=0 case also good. since that is sigma(r) case. This is higher priority than r=1e-4 to me.
+      * AK: who are these plots aimed at?  CMB experts? General science folks?  Current plots take a large amount of explaining for a small amount of confidence that PICO can do this.  Could consider text that paints a broader picture and 1 plot with reasonable complexity model.  Showing the failure case seems like a bad idea.
+        * SH: True. Other possible risk is we paint a picture that is to optimistic. Since haven't shown we can do this on full sky.
+        * AK: Showing negative results from a work in progress doesn't seem worthwhile.
+        * AK: foregrounds should focus on what access to high frequencies buys you. more positive statement.  Not excessively detailed/complicated plots. (SH: see third plot)
+    * Frequency breadth (3rd plot)
+      * SH: for 1 PySM model.  20-800 GHz fully reconstructs. smaller range fails.
+      * AK: yes. This is better argument. the frequency coverage is the strongest point to make.
+    * SH: Bottom line? showing the state of study and challenges is less useful than showing what works and state of complexity.
+      * AK: would phrase as 'show problems and show how they will be dealt with'
+    * need reviewers besides external team. rapid feedback.
+      * **AK will give comments at least. SH will ping when done.**
+Send full draft out in ~ 2 days.

User Tools

Differences

Page Tools