summaryrefslogtreecommitdiff
path: root/ROADMAP.md
diff options
context:
space:
mode:
authorYurenHao0426 <Blackhao0426@gmail.com>2026-07-22 19:16:10 -0500
committerYurenHao0426 <Blackhao0426@gmail.com>2026-07-22 19:16:10 -0500
commit4ad5bdfe3e6d4ab7fc1349686b43982c60f951d3 (patch)
treed0ae63dfa6e4a4dbdc3ac6574ef88be5abe65329 /ROADMAP.md
parent9e987aea12035425b3a673579b702b0ecf5631d4 (diff)
docs: register sealed dynamic confirmation
Diffstat (limited to 'ROADMAP.md')
-rw-r--r--ROADMAP.md6
1 files changed, 5 insertions, 1 deletions
diff --git a/ROADMAP.md b/ROADMAP.md
index 57b6710..ac3914c 100644
--- a/ROADMAP.md
+++ b/ROADMAP.md
@@ -532,7 +532,11 @@ feedback cosine is `0.904056`, all projection certificates pass, cost is
untouched. Because this is one short seed, the score stays 5/10. The
predeclared 200-epoch D3 validation endpoint is now open; only a complete D3
pass can make the 5-to-6 score update eligible and open independent test
-confirmation.
+confirmation. Before observing D3, `DYNAMIC_INNOVATION.md` and executable
+scripts freeze D4 as a five-seed paired clean-KP/dynamic test panel. D4 is
+hard-gated on a complete D3 pass, evaluates test exactly once per run, and can
+raise 6-to-7 only through its predeclared accuracy, paired noninferiority,
+alignment, leakage, query, and cost checks.
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