Table 4.
τ predictions from 10 000 CMB + SRoll2 test simulations generated with three different fiducial τ values.
| Test on SRoll2 simulations | ||||||
|---|---|---|---|---|---|---|
| 143 + 100 GHz | 143 + 100 GHz | |||||
| SRoll2 training | SRoll2 retraining | |||||
| fiducial τ |
 |
 |
σ(τNN) |
 |
 |
σ(τNN) |
| 0.05 | 0.0526 | 0.0059 | 0.0066 | 0.0508 | 0.0077 | 0.0091 |
| 0.06 | 0.0622 | 0.0062 | 0.0070 | 0.0606 | 0.0079 | 0.0088 |
| 0.07 | 0.0722 | 0.0064 | 0.0070 | 0.0707 | 0.0081 | 0.0087 |
Notes. These results correspond to two frequency channels, either training on SRoll2 from the start or retraining on SRoll2 maps. Displayed are the average posterior mean, average predicted standard deviation 
, and the scatter σ(τNN) calculated across the test simulations.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.