From 5bc712478ecc8c12c422993789fbaee015f5bc4f Mon Sep 17 00:00:00 2001
From: jgrusewski <jgrusewski@localhost>
Date: Mon, 20 Apr 2026 21:38:53 +0200
Subject: [PATCH] =?UTF-8?q?refactor(A2):=20production=20=CE=B1=3D0.1,=20te?=
 =?UTF-8?q?sts=20opt=20in=20to=20=CE=B1=3D0.3=20via=20with=5Falpha()?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

LearningHealth::new() now uses α=0.1 (the production-appropriate ~10-sample
EMA window for 80-100 epoch runs). Tests that need to reach threshold
assertions inside a 5-iteration loop (warmup=3 + 2 EMA steps) use
LearningHealth::with_alpha(0.3) instead of bending production behaviour
to satisfy test convenience.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---
 .../ml/src/cuda_pipeline/learning_health.rs   | 20 +++++++++++++++----
 1 file changed, 16 insertions(+), 4 deletions(-)

diff --git a/crates/ml/src/cuda_pipeline/learning_health.rs b/crates/ml/src/cuda_pipeline/learning_health.rs
index 67389d251..38191a55e 100644
--- a/crates/ml/src/cuda_pipeline/learning_health.rs
+++ b/crates/ml/src/cuda_pipeline/learning_health.rs
@@ -86,12 +86,22 @@ pub struct LearningHealth {
 }
 
 impl LearningHealth {
+    /// Production constructor. α=0.1 gives ~10-sample EMA window — appropriate
+    /// smoothing for an 80–100 epoch training run where we want health to be
+    /// responsive but not whipsaw on single-epoch noise.
     pub fn new() -> Self {
+        Self::with_alpha(0.1)
+    }
+
+    /// Test/tuning constructor with an explicit α. Tests use higher α (≈0.3)
+    /// to satisfy threshold assertions inside a 5-iteration (warmup + 2 EMA
+    /// steps) test loop; production code should use `new()`.
+    pub fn with_alpha(alpha: f32) -> Self {
         Self {
             value: 0.5, // neutral start
             epoch: 0,
             components: NormalizedComponents::default(),
-            ema_alpha: 0.3,
+            ema_alpha: alpha,
             warmup_epochs: 3,
         }
     }
@@ -144,7 +154,9 @@ mod tests {
             grad_consistency: -0.3, // gradients contradict
             spectral_gap: 50.0,   // rank 1
         };
-        let mut health = LearningHealth::new();
+        // Tests use α=0.3 to reach thresholds inside the short 5-iteration loop;
+        // production new() uses α=0.1.
+        let mut health = LearningHealth::with_alpha(0.3);
         // Skip warmup
         for _ in 0..5 { health.update(&raw); }
         assert!(health.value <= 0.3, "Collapsed state should give health <= 0.3, got {}", health.value);
@@ -161,7 +173,7 @@ mod tests {
             grad_consistency: 0.9, // consistent direction
             spectral_gap: 1.5,     // balanced rank
         };
-        let mut health = LearningHealth::new();
+        let mut health = LearningHealth::with_alpha(0.3);
         for _ in 0..5 { health.update(&raw); }
         assert!(health.value >= 0.7, "Healthy state should give health >= 0.7, got {}", health.value);
     }
@@ -173,7 +185,7 @@ mod tests {
             grad_norm: 50_000.0, ens_disagreement: 0.0,
             grad_consistency: -0.5, spectral_gap: 100.0,
         };
-        let mut health = LearningHealth::new();
+        let mut health = LearningHealth::with_alpha(0.3);
         // First 3 updates are warmup — should stay at 0.5
         for _ in 0..3 {
             let v = health.update(&raw);