diff options
| -rw-r--r-- | CLLN_SCALING.md | 6 |
1 files changed, 5 insertions, 1 deletions
diff --git a/CLLN_SCALING.md b/CLLN_SCALING.md index 3e14dbd..be98543 100644 --- a/CLLN_SCALING.md +++ b/CLLN_SCALING.md @@ -57,6 +57,9 @@ diagnostics. - fraction of task/device pairs reaching zero classification error; - restricted epochs and local updates to zero error, assigning the experiment horizon to runs that do not reach the target; +- local teaching-signal reads to zero error, counting both the task-period + edge signal and every SDIL neutral edge signal; SDIL therefore pays two + local scalar reads per updated edge in this ladder; - learning-curve area for classification error; - wall time and neutral observations as implementation costs. @@ -71,6 +74,8 @@ change. The headline wording is selected from the observed sign: - lower resource exponent: “SDIL reduces the updates-to-target scaling exponent by X%”; +- lower resource count without a resolved exponent change: “SDIL uses X% + fewer local teaching-signal reads at N edges”; - lower error-growth exponent: “SDIL reduces imperfection-induced error growth with network size by X%”; - overlapping interval: report no resolved scaling-rate difference. @@ -92,4 +97,3 @@ as independent confirmations. 5. The scaling conclusion is computed from all frozen sizes and task clusters, including failed runs. 6. Hardware-realistic claims use Part 3 only. - |