summaryrefslogtreecommitdiff
path: root/ROADMAP.md
diff options
context:
space:
mode:
authorYurenHao0426 <Blackhao0426@gmail.com>2026-07-22 16:38:42 -0500
committerYurenHao0426 <Blackhao0426@gmail.com>2026-07-22 16:38:42 -0500
commit575530e7124c540e58571b50ef147a5f20e6cb54 (patch)
treec1058abde90f59d2e9b594661f1794b6e974ae18 /ROADMAP.md
parentd818fc26cdcbc00521e4cd4e347e74d5cd3e5670 (diff)
docs: explain mixed-traffic instability
Diffstat (limited to 'ROADMAP.md')
-rw-r--r--ROADMAP.md6
1 files changed, 6 insertions, 0 deletions
diff --git a/ROADMAP.md b/ROADMAP.md
index c425356..dd521e4 100644
--- a/ROADMAP.md
+++ b/ROADMAP.md
@@ -502,6 +502,12 @@ checks pass while all performance, finite, alignment, and tracking checks
fail. Per the frozen stop rule, MT-2 and MT-3 remain untouched, no weaker
traffic ratio is allowed, and this standard-ResNet accept path closes at 5/10.
+The training-only post-failure audit localizes the common instability to stem
+`W[0]` and its momentum at steps 23/69/74 for raw/matched/innovation,
+respectively, with no validation or test evaluation. This narrows any future
+independent method to operator-level gain control or dissipativity; more
+endpoint tuning of the same residual-RMS gate is not justified.
+
Prepare convolutional local-update primitives and ResNet-20/32/56 protocols early. Queue frozen
runs opportunistically on authorized idle GPUs. Because BurstCCN already reports CIFAR-10 and
ImageNet scaling, dataset scale alone is not novel. The oral-level target is a memorable joint