From 2fa24acae8bb7f8c026db2f7fdade4a29b640d8d Mon Sep 17 00:00:00 2001 From: YurenHao0426 Date: Wed, 8 Apr 2026 19:24:06 -0500 Subject: Sync experiment+protocol scripts with v2.32 corrected control values MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The pre-v2.31 unsourced values BP=0.609 and DFA=0.308 (which v2.31 fixed to 0.585 and 0.301 via matched 30-ep controls) were also hardcoded as "compare to" comments in 5 helper scripts: experiments/bp_with_penalty_control.py experiments/dfa_residual_penalty_test.py experiments/resmlp_frozen_blocks_baseline.py protocol/examples/threshold_d_sensitivity.py protocol/examples/plot_penalty_rescue.py These are non-paper-input scripts (their output goes to stdout, not to the paper), so the stale values didn't cause numerical errors in the paper itself. But the original v2.31 BP+pen=0.609 unsourced number bug came from exactly this kind of hardcoded "for-comparison" comment that was never measured. Updating them now to remove the same trap from future runs. Each script now references the matched 30-ep 3-seed values from results/bp_no_penalty_30ep, results/dfa_no_penalty_30ep, results/ dfa_pen_short, and results/bp_with_penalty. protocol/EVIDENCE_SUMMARY.md and PAPER_OUTLINE.md still have stale numbers — these are project scratch documents and not user-facing. Deferred to a separate sweep if needed. Co-Authored-By: Claude Opus 4.6 (1M context) --- protocol/examples/threshold_d_sensitivity.py | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) (limited to 'protocol/examples/threshold_d_sensitivity.py') diff --git a/protocol/examples/threshold_d_sensitivity.py b/protocol/examples/threshold_d_sensitivity.py index d3f2c58..065efc7 100644 --- a/protocol/examples/threshold_d_sensitivity.py +++ b/protocol/examples/threshold_d_sensitivity.py @@ -22,13 +22,15 @@ REPO_ROOT = os.path.dirname( def main(): # 3-seed mean accuracies on 4-block d=256 ResMLP CIFAR-10 + # Updated v2.32 with matched 30-epoch controls conditions = [ - ("BP-trainable", 0.609, 0.004), - ("DFA-shallow", 0.349, 0.002), - ("DFA-vanilla", 0.308, 0.014), - ("DFA-pen lam=1e-3", 0.372, None), # 1 seed - ("DFA-pen lam=1e-2", 0.363, 0.0007), - ("DFA-frozen-rand", 0.349, 0.002), + ("BP-trainable 100ep", 0.6147, 0.004), # protocol_audit + ("BP-trainable 30ep", 0.585, 0.001), # results/bp_no_penalty_30ep + ("BP+pen 30ep lam=1e-2", 0.532, 0.006), # results/bp_with_penalty + ("DFA-shallow", 0.349, 0.002), # frozen baseline + ("DFA-vanilla 100ep", 0.306, 0.006), # protocol_audit + ("DFA-vanilla 30ep", 0.301, 0.005), # results/dfa_no_penalty_30ep + ("DFA+pen 30ep lam=1e-2", 0.360, 0.001), # results/dfa_pen_short ] shallow_acc = 0.349 -- cgit v1.2.3