{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "a427e9d5",
   "metadata": {},
   "source": [
    "# E3 — ISIC 2017 UNet2D LR-Schedule Tie-Break (10 seeds)\n",
    "\n",
    "**Question.** With the architecture and learning rate already fixed (`classical + attention_gate`, `lr = 3e-4`), does the **LR decay schedule** change the model? We compare cosine annealing against ReduceLROnPlateau head-to-head.\n",
    "\n",
    "**Design.** 10 shared seeds (100–109). Within each seed both schedulers share weight initialisation and data ordering, so the *only* difference inside a seed pair is the schedule. We analyse the **paired** differences Δᵢ = cosineᵢ − plateauᵢ on the ISIC 2017 validation set.\n",
    "\n",
    "**Source.** `E3-isic2017-unet2d-cosanneal-10seeds.db` + `E3-isic2017-unet2d-reduceonplateauON-10seeds.db` — 20 runs (2 schedules × 10 seeds), all `FINISHED`.\n",
    "\n",
    "> **Scope.** This compares two *decay* schedules against each other. A 1-seed pilot found no quality difference between cosine and flat LR, so the scheduler was dropped before E4. A post-hoc unpaired comparison (E3 cosine vs E4 flat LR, 10 seeds each) on plateau Dice suggests cosine is better (Δ=+0.0042, BCa 95% CI [+0.0015, +0.0069]); the 1-seed pilot was underpowered. See Decision for caveats."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "686cc05d",
   "metadata": {},
   "source": [
    "## Executive summary\n",
    "\n",
    "- Paired across 10 shared seeds, Δ = cosine − plateau\n",
    "- Primary test = Wilcoxon signed-rank\n",
    "- CI = BCa bootstrap 95 % (10 000 resamples, `rng=42`)\n",
    "- d_z = Cohen's paired effect size.\n",
    "\n",
    "| Metric | Cosine | Plateau | Δ (cos−plat) | 95 % BCa CI | Wilcoxon p | d_z | Verdict |\n",
    "|---|---|---|---|---|---|---|---|\n",
    "| **Plateau Dice** — *primary* [1] | 0.8346 | 0.8331 | +0.0015 | [−0.0006, +0.0038] | 0.32 | +0.38 | tie |\n",
    "| Peak Dice | 0.8451 | 0.8459 | −0.0007 | [−0.0024, +0.0006] | 0.43 | −0.29 | tie |\n",
    "| Peak IoU | 0.7382 | 0.7392 | −0.0010 | [−0.0033, +0.0008] | 0.43 | −0.29 | tie |\n",
    "| Generalisation gap *(lower = better)* | 0.0957 | 0.0971 | −0.0014 | [−0.0054, +0.0039] | 0.70 | −0.18 | tie |\n",
    "| Throughput | 119.4 sps | 118.9 sps | +0.5 (+0.4 %) | [−1.7, +4.8] | 0.85 | +0.09 | tie |\n",
    "\n",
    "[1] Plateau Dice (last-10-epoch mean) is the primary quality metric; tested standalone (k = 1, α = 0.05). Secondary quality metrics (Peak Dice, Peak IoU, gen-gap) form a Holm family (k = 3).\n",
    "\n",
    "**Finding.**\n",
    "- **No metric reaches significance.** Every Wilcoxon p ≥ 0.32, every BCa CI straddles 0, every |d_z| ≤ 0.38 (negligible–small). The two schedules are statistically indistinguishable on all quality and cost axes.\n",
    "- The only consistent difference is **convergence timing, not quality**: cosine reaches its best checkpoint at epoch 125 on average vs 161 for plateau (≈ 36 epochs / ~22 % earlier), at equal throughput.\n",
    "\n",
    "**Decision: between the two decay schedules, prefer `cosine_annealing`** — not on Dice (a tie) but on earlier, deterministic convergence. The decay-vs-no-decay question stays open pending the flat-LR baseline."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "169c889e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-24T10:16:56.209817Z",
     "iopub.status.busy": "2026-06-24T10:16:56.209318Z",
     "iopub.status.idle": "2026-06-24T10:16:57.864760Z",
     "shell.execute_reply": "2026-06-24T10:16:57.864275Z"
    }
   },
   "outputs": [],
   "source": [
    "# ── Path bootstrap (must run before any SkiNet import) ───────────────────────\n",
    "import sys\n",
    "from pathlib import Path\n",
    "\n",
    "# Resolve the repo root regardless of kernel working directory.\n",
    "# VS Code injects __vsc_ipynb_file__; nbconvert sets CWD = notebook directory.\n",
    "try:\n",
    "    _nb_dir = Path(__vsc_ipynb_file__).resolve().parent  # VS Code interactive\n",
    "except NameError:\n",
    "    _nb_dir = Path().resolve()                            # nbconvert / CLI\n",
    "\n",
    "PROJECT_ROOT = _nb_dir.parents[1].resolve()\n",
    "sys.path.insert(0, str(PROJECT_ROOT))\n",
    "\n",
    "# ── Imports ───────────────────────────────────────────────────────────────────\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "from SkiNet.Utils.analysis.aggregation import load_runs\n",
    "from SkiNet.Utils.analysis.stats import build_comparison_table\n",
    "from SkiNet.Utils.analysis.reporting import show_run_table, show_comparison_table, show_family_verdicts\n",
    "from SkiNet.Utils.analysis.plotting import set_paper_style, plot_paired_slopegraph, plot_paired_forest\n",
    "from SkiNet.Utils.analysis.schema import VAL_DICE_MAX, VAL_DICE_TAIL_MEAN, VAL_IOU_MAX, GENERALIZATION_GAP_FINAL, SAMPLES_PER_SEC\n",
    "\n",
    "# ── Configuration — every tunable argument lives in this cell ────────────────\n",
    "FIG_DIR  = _nb_dir / '_static/model_selection'\n",
    "# Each scheduler arm was logged to its own MLflow DB (both use experiment_id=1\n",
    "# internally), so we load and label them separately, then concatenate.\n",
    "DB_DIR   = PROJECT_ROOT / 'mlruns'\n",
    "COS_DB   = DB_DIR / 'E3-isic2017-unet2d-cosanneal-10seeds.db'             # seeds 100–109\n",
    "PLAT_DB  = DB_DIR / 'E3-isic2017-unet2d-reduceonplateauON-10seeds.db'     # seeds 100–109\n",
    "\n",
    "COSINE, PLATEAU = 'cosine_annealing', 'reduce_on_plateau'\n",
    "PALETTE = {COSINE: '#d1495b', PLATEAU: '#30638e'}  # color palette for the two schedules\n",
    "\n",
    "ALPHA   = 0.05      # significance level for confidence intervals and hypothesis tests\n",
    "N_BOOT  = 10_000    # bootstrap samples for confidence intervals and p-values\n",
    "RNG     = np.random.default_rng(42)  # random number generator for reproducibility\n",
    "\n",
    "# Metric column names are imported from SkiNet.Utils.analysis.schema; the spec\n",
    "# lists below stay here because the choice of metrics, family sizes and display\n",
    "# names is specific to this E3 paired comparison.\n",
    "PRIMARY_METRIC    = VAL_DICE_TAIL_MEAN\n",
    "SECONDARY_METRICS = [VAL_DICE_MAX, VAL_IOU_MAX, GENERALIZATION_GAP_FINAL]\n",
    "# (metric, higher_is_better, family_size_k)\n",
    "METRICS_SPEC = [\n",
    "    (VAL_DICE_MAX,             True,  len(SECONDARY_METRICS)),\n",
    "    (VAL_IOU_MAX,              True,  len(SECONDARY_METRICS)),\n",
    "    (VAL_DICE_TAIL_MEAN,       True,  1),\n",
    "    (GENERALIZATION_GAP_FINAL, False, len(SECONDARY_METRICS)),\n",
    "    (SAMPLES_PER_SEC,          True,  1),\n",
    "]\n",
    "SLOPE_METRICS = [\n",
    "    (VAL_DICE_MAX,       'Peak Dice (best checkpoint)'),\n",
    "    (VAL_DICE_TAIL_MEAN, 'Plateau Dice (last-10-epoch mean)'),\n",
    "]\n",
    "# (display_name, metric, higher_is_better)\n",
    "FOREST_SPECS = [\n",
    "    ('Peak Dice',         VAL_DICE_MAX,             False),\n",
    "    ('Peak IoU',          VAL_IOU_MAX,              False),\n",
    "    ('Plateau Dice',      VAL_DICE_TAIL_MEAN,       False),\n",
    "    ('Gen-gap reduction', GENERALIZATION_GAP_FINAL, True),\n",
    "]\n",
    "\n",
    "# ── Presentation ─────────────────────────────────────────────────────────────\n",
    "set_paper_style(context='notebook')\n",
    "pd.set_option('display.width', 220)\n",
    "pd.set_option('display.float_format', '{:.4f}'.format)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "f39fb386",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-24T10:16:57.868734Z",
     "iopub.status.busy": "2026-06-24T10:16:57.868356Z",
     "iopub.status.idle": "2026-06-24T10:16:59.381338Z",
     "shell.execute_reply": "2026-06-24T10:16:59.380810Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded 10 runs: {'cosine_annealing': 10} | seeds: [100, 101, 102, 103, 104, 105, 106, 107, 108, 109]\n",
      "Loaded 10 runs: {'reduce_on_plateau': 10} | seeds: [100, 101, 102, 103, 104, 105, 106, 107, 108, 109]\n"
     ]
    }
   ],
   "source": [
    "# ── Load: 20 runs = 2 schedules × 10 seeds ───────────────────────────────────\n",
    "# Both DBs use experiment_id=1, so each is loaded with its own one-entry exp_map\n",
    "# and the labelled frames are concatenated into the paired (seed × schedule) table.\n",
    "runs = pd.concat([\n",
    "    load_runs(COS_DB,  exp_map={1: COSINE},  monitor='val_dice'),\n",
    "    load_runs(PLAT_DB, exp_map={1: PLATEAU}, monitor='val_dice'),\n",
    "], ignore_index=True)\n",
    "SEEDS, N = sorted(runs['seed'].unique()), runs['seed'].nunique()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cac097b4",
   "metadata": {},
   "source": [
    "## 1. Data\n",
    "\n",
    "One row per (seed, schedule). Columns:\n",
    "\n",
    "- **`val_dice_max`** — peak Dice of the best-saved checkpoint (the model that would actually be deployed).\n",
    "- **`val_dice_tail_mean` / `…_std`** — mean and SD of Dice over the last 10 epochs: the convergent *plateau* level and its noise.\n",
    "- **`val_dice_max_epoch`** — epoch at which the best checkpoint was reached (convergence-speed signal).\n",
    "- **`val_iou_max`** — peak IoU of the best checkpoint.\n",
    "- **`generalization_gap_final`** — final `train_dice − val_dice` (overfitting signal; lower is better).\n",
    "- **`samples_per_sec` / `duration_min`** — training throughput and wall-clock cost."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "d4128834",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-24T10:16:59.383948Z",
     "iopub.status.busy": "2026-06-24T10:16:59.383281Z",
     "iopub.status.idle": "2026-06-24T10:16:59.400509Z",
     "shell.execute_reply": "2026-06-24T10:16:59.399941Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>arch</th>\n",
       "      <th>seed</th>\n",
       "      <th>val_dice_max</th>\n",
       "      <th>val_dice_tail_mean</th>\n",
       "      <th>val_dice_tail_std</th>\n",
       "      <th>val_iou_max</th>\n",
       "      <th>generalization_gap_final</th>\n",
       "      <th>samples_per_sec</th>\n",
       "      <th>duration_min</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>cosine_annealing</td>\n",
       "      <td>100</td>\n",
       "      <td>0.8444</td>\n",
       "      <td>0.8384</td>\n",
       "      <td>0.0020</td>\n",
       "      <td>0.7374</td>\n",
       "      <td>0.0944</td>\n",
       "      <td>119.8566</td>\n",
       "      <td>47.3305</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>cosine_annealing</td>\n",
       "      <td>101</td>\n",
       "      <td>0.8475</td>\n",
       "      <td>0.8385</td>\n",
       "      <td>0.0011</td>\n",
       "      <td>0.7410</td>\n",
       "      <td>0.0902</td>\n",
       "      <td>115.3991</td>\n",
       "      <td>45.7549</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>cosine_annealing</td>\n",
       "      <td>102</td>\n",
       "      <td>0.8443</td>\n",
       "      <td>0.8366</td>\n",
       "      <td>0.0020</td>\n",
       "      <td>0.7372</td>\n",
       "      <td>0.0943</td>\n",
       "      <td>115.9677</td>\n",
       "      <td>45.7975</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>cosine_annealing</td>\n",
       "      <td>103</td>\n",
       "      <td>0.8474</td>\n",
       "      <td>0.8389</td>\n",
       "      <td>0.0020</td>\n",
       "      <td>0.7409</td>\n",
       "      <td>0.0887</td>\n",
       "      <td>116.6589</td>\n",
       "      <td>45.6267</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>cosine_annealing</td>\n",
       "      <td>104</td>\n",
       "      <td>0.8476</td>\n",
       "      <td>0.8323</td>\n",
       "      <td>0.0008</td>\n",
       "      <td>0.7413</td>\n",
       "      <td>0.1000</td>\n",
       "      <td>118.5947</td>\n",
       "      <td>45.8264</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>cosine_annealing</td>\n",
       "      <td>105</td>\n",
       "      <td>0.8473</td>\n",
       "      <td>0.8322</td>\n",
       "      <td>0.0021</td>\n",
       "      <td>0.7409</td>\n",
       "      <td>0.0923</td>\n",
       "      <td>119.3180</td>\n",
       "      <td>45.6379</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>cosine_annealing</td>\n",
       "      <td>106</td>\n",
       "      <td>0.8413</td>\n",
       "      <td>0.8292</td>\n",
       "      <td>0.0013</td>\n",
       "      <td>0.7335</td>\n",
       "      <td>0.1028</td>\n",
       "      <td>121.5028</td>\n",
       "      <td>45.7030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>cosine_annealing</td>\n",
       "      <td>107</td>\n",
       "      <td>0.8438</td>\n",
       "      <td>0.8326</td>\n",
       "      <td>0.0012</td>\n",
       "      <td>0.7379</td>\n",
       "      <td>0.0971</td>\n",
       "      <td>118.5776</td>\n",
       "      <td>45.7078</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>cosine_annealing</td>\n",
       "      <td>108</td>\n",
       "      <td>0.8431</td>\n",
       "      <td>0.8352</td>\n",
       "      <td>0.0018</td>\n",
       "      <td>0.7342</td>\n",
       "      <td>0.0977</td>\n",
       "      <td>129.1880</td>\n",
       "      <td>45.7106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>cosine_annealing</td>\n",
       "      <td>109</td>\n",
       "      <td>0.8447</td>\n",
       "      <td>0.8317</td>\n",
       "      <td>0.0016</td>\n",
       "      <td>0.7377</td>\n",
       "      <td>0.0994</td>\n",
       "      <td>118.5783</td>\n",
       "      <td>45.7795</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>reduce_on_plateau</td>\n",
       "      <td>100</td>\n",
       "      <td>0.8467</td>\n",
       "      <td>0.8347</td>\n",
       "      <td>0.0056</td>\n",
       "      <td>0.7406</td>\n",
       "      <td>0.0942</td>\n",
       "      <td>116.5515</td>\n",
       "      <td>47.6221</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>reduce_on_plateau</td>\n",
       "      <td>101</td>\n",
       "      <td>0.8439</td>\n",
       "      <td>0.8319</td>\n",
       "      <td>0.0044</td>\n",
       "      <td>0.7366</td>\n",
       "      <td>0.0942</td>\n",
       "      <td>116.0661</td>\n",
       "      <td>45.8422</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>reduce_on_plateau</td>\n",
       "      <td>102</td>\n",
       "      <td>0.8449</td>\n",
       "      <td>0.8294</td>\n",
       "      <td>0.0073</td>\n",
       "      <td>0.7382</td>\n",
       "      <td>0.0982</td>\n",
       "      <td>119.6104</td>\n",
       "      <td>46.0099</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>reduce_on_plateau</td>\n",
       "      <td>103</td>\n",
       "      <td>0.8489</td>\n",
       "      <td>0.8358</td>\n",
       "      <td>0.0092</td>\n",
       "      <td>0.7427</td>\n",
       "      <td>0.0985</td>\n",
       "      <td>121.3753</td>\n",
       "      <td>45.8837</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>reduce_on_plateau</td>\n",
       "      <td>104</td>\n",
       "      <td>0.8458</td>\n",
       "      <td>0.8331</td>\n",
       "      <td>0.0062</td>\n",
       "      <td>0.7393</td>\n",
       "      <td>0.1072</td>\n",
       "      <td>119.0549</td>\n",
       "      <td>45.8842</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>reduce_on_plateau</td>\n",
       "      <td>105</td>\n",
       "      <td>0.8466</td>\n",
       "      <td>0.8354</td>\n",
       "      <td>0.0092</td>\n",
       "      <td>0.7401</td>\n",
       "      <td>0.0843</td>\n",
       "      <td>118.1467</td>\n",
       "      <td>46.3332</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>reduce_on_plateau</td>\n",
       "      <td>106</td>\n",
       "      <td>0.8473</td>\n",
       "      <td>0.8317</td>\n",
       "      <td>0.0051</td>\n",
       "      <td>0.7418</td>\n",
       "      <td>0.0981</td>\n",
       "      <td>120.2713</td>\n",
       "      <td>46.3837</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>reduce_on_plateau</td>\n",
       "      <td>107</td>\n",
       "      <td>0.8448</td>\n",
       "      <td>0.8350</td>\n",
       "      <td>0.0056</td>\n",
       "      <td>0.7373</td>\n",
       "      <td>0.0840</td>\n",
       "      <td>122.5502</td>\n",
       "      <td>46.2808</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>reduce_on_plateau</td>\n",
       "      <td>108</td>\n",
       "      <td>0.8437</td>\n",
       "      <td>0.8321</td>\n",
       "      <td>0.0074</td>\n",
       "      <td>0.7368</td>\n",
       "      <td>0.1091</td>\n",
       "      <td>116.2731</td>\n",
       "      <td>46.0836</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>reduce_on_plateau</td>\n",
       "      <td>109</td>\n",
       "      <td>0.8461</td>\n",
       "      <td>0.8317</td>\n",
       "      <td>0.0050</td>\n",
       "      <td>0.7384</td>\n",
       "      <td>0.1031</td>\n",
       "      <td>119.2312</td>\n",
       "      <td>46.0698</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 arch  seed  val_dice_max  val_dice_tail_mean  val_dice_tail_std  val_iou_max  generalization_gap_final  samples_per_sec  duration_min\n",
       "0    cosine_annealing   100        0.8444              0.8384             0.0020       0.7374                    0.0944         119.8566       47.3305\n",
       "1    cosine_annealing   101        0.8475              0.8385             0.0011       0.7410                    0.0902         115.3991       45.7549\n",
       "2    cosine_annealing   102        0.8443              0.8366             0.0020       0.7372                    0.0943         115.9677       45.7975\n",
       "3    cosine_annealing   103        0.8474              0.8389             0.0020       0.7409                    0.0887         116.6589       45.6267\n",
       "4    cosine_annealing   104        0.8476              0.8323             0.0008       0.7413                    0.1000         118.5947       45.8264\n",
       "5    cosine_annealing   105        0.8473              0.8322             0.0021       0.7409                    0.0923         119.3180       45.6379\n",
       "6    cosine_annealing   106        0.8413              0.8292             0.0013       0.7335                    0.1028         121.5028       45.7030\n",
       "7    cosine_annealing   107        0.8438              0.8326             0.0012       0.7379                    0.0971         118.5776       45.7078\n",
       "8    cosine_annealing   108        0.8431              0.8352             0.0018       0.7342                    0.0977         129.1880       45.7106\n",
       "9    cosine_annealing   109        0.8447              0.8317             0.0016       0.7377                    0.0994         118.5783       45.7795\n",
       "10  reduce_on_plateau   100        0.8467              0.8347             0.0056       0.7406                    0.0942         116.5515       47.6221\n",
       "11  reduce_on_plateau   101        0.8439              0.8319             0.0044       0.7366                    0.0942         116.0661       45.8422\n",
       "12  reduce_on_plateau   102        0.8449              0.8294             0.0073       0.7382                    0.0982         119.6104       46.0099\n",
       "13  reduce_on_plateau   103        0.8489              0.8358             0.0092       0.7427                    0.0985         121.3753       45.8837\n",
       "14  reduce_on_plateau   104        0.8458              0.8331             0.0062       0.7393                    0.1072         119.0549       45.8842\n",
       "15  reduce_on_plateau   105        0.8466              0.8354             0.0092       0.7401                    0.0843         118.1467       46.3332\n",
       "16  reduce_on_plateau   106        0.8473              0.8317             0.0051       0.7418                    0.0981         120.2713       46.3837\n",
       "17  reduce_on_plateau   107        0.8448              0.8350             0.0056       0.7373                    0.0840         122.5502       46.2808\n",
       "18  reduce_on_plateau   108        0.8437              0.8321             0.0074       0.7368                    0.1091         116.2731       46.0836\n",
       "19  reduce_on_plateau   109        0.8461              0.8317             0.0050       0.7384                    0.1031         119.2312       46.0698"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_run_table(runs)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "76834aa7",
   "metadata": {},
   "source": [
    "## 2. Statistical methods\n",
    "\n",
    "### 2.1 Paired design\n",
    "\n",
    "Within each seed, cosine and plateau share weight initialisation and data ordering, so everything that varies run-to-run **except the schedule** is held constant. Subtracting within the pair, Δᵢ = cosineᵢ − plateauᵢ, removes seed-to-seed noise and we analyse 10 paired differences.\n",
    "\n",
    "> **Scope caveat.** All seeds reuse a single fixed ISIC-2017 train/validation split; only initialisation varies — this is *not* k-fold cross-validation. Per Rainio et al. (2024), fixed-split p-values can understate variance, so treat every interval and p-value below as an **optimistic lower bound** on the true uncertainty. (For a null result this cuts the safe way: even the optimistic intervals fail to separate the schedules.)\n",
    "\n",
    "### 2.2 Hypothesis families\n",
    "\n",
    "Each family controls its own family-wise error rate at α = 0.05.\n",
    "\n",
    "| Family | Metric(s) | k | Per-metric threshold | Correction |\n",
    "|---|---|---|---|---|\n",
    "| **Primary** | `val_dice_tail_mean` | 1 | **0.05** | none |\n",
    "| **Secondary quality** | `val_dice_max`, `val_iou_max`, gen-gap | 3 | ≤ **0.0167** | Holm step-down |\n",
    "| **Training cost** | `samples_per_sec` | 1 | **0.05** | none |\n",
    "\n",
    "*Why plateau Dice is primary:* it measures the **stable, convergent** Dice level (mean of the last 10 epochs) — the quantity a schedule is supposed to improve — rather than a lucky single-epoch peak.\n",
    "\n",
    "### 2.3 Inference criteria\n",
    "\n",
    "Three complementary statistics on the same 10 paired differences — identical to the E2 tie-break (Wilcoxon signed-rank, BCa bootstrap 95 % CI, Cohen's d_z); see E2 §2.3 for the full derivation. At n = 10 the smallest achievable two-tailed Wilcoxon p is 2 / 2¹⁰ = **0.00195**, so the test has room to reject if a real effect existed."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3826f86b",
   "metadata": {},
   "source": [
    "## 3. Results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "b96ac07e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-24T10:16:59.402517Z",
     "iopub.status.busy": "2026-06-24T10:16:59.402323Z",
     "iopub.status.idle": "2026-06-24T10:16:59.447179Z",
     "shell.execute_reply": "2026-06-24T10:16:59.445887Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>cosine</th>\n",
       "      <th>plateau</th>\n",
       "      <th>Δ (cosine−plateau)</th>\n",
       "      <th>95% BCa CI</th>\n",
       "      <th>wilcoxon_p</th>\n",
       "      <th>sig</th>\n",
       "      <th>d_z</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>val_dice_max</th>\n",
       "      <td>0.8451</td>\n",
       "      <td>0.8459</td>\n",
       "      <td>-0.0007</td>\n",
       "      <td>[-0.0024, +0.0006]</td>\n",
       "      <td>0.4316</td>\n",
       "      <td></td>\n",
       "      <td>-0.2900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>val_iou_max</th>\n",
       "      <td>0.7382</td>\n",
       "      <td>0.7392</td>\n",
       "      <td>-0.0010</td>\n",
       "      <td>[-0.0033, +0.0008]</td>\n",
       "      <td>0.4316</td>\n",
       "      <td></td>\n",
       "      <td>-0.2900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>val_dice_tail_mean</th>\n",
       "      <td>0.8346</td>\n",
       "      <td>0.8331</td>\n",
       "      <td>+0.0015</td>\n",
       "      <td>[-0.0006, +0.0038]</td>\n",
       "      <td>0.3223</td>\n",
       "      <td></td>\n",
       "      <td>0.3800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>generalization_gap_final</th>\n",
       "      <td>0.0957</td>\n",
       "      <td>0.0971</td>\n",
       "      <td>-0.0014</td>\n",
       "      <td>[-0.0054, +0.0039]</td>\n",
       "      <td>0.6953</td>\n",
       "      <td></td>\n",
       "      <td>-0.1800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>samples_per_sec</th>\n",
       "      <td>119.3642</td>\n",
       "      <td>118.9131</td>\n",
       "      <td>+0.4511</td>\n",
       "      <td>[-1.6984, +4.7605]</td>\n",
       "      <td>0.8457</td>\n",
       "      <td></td>\n",
       "      <td>0.0900</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           cosine  plateau Δ (cosine−plateau)          95% BCa CI wilcoxon_p sig     d_z\n",
       "val_dice_max               0.8451   0.8459            -0.0007  [-0.0024, +0.0006]     0.4316     -0.2900\n",
       "val_iou_max                0.7382   0.7392            -0.0010  [-0.0033, +0.0008]     0.4316     -0.2900\n",
       "val_dice_tail_mean         0.8346   0.8331            +0.0015  [-0.0006, +0.0038]     0.3223      0.3800\n",
       "generalization_gap_final   0.0957   0.0971            -0.0014  [-0.0054, +0.0039]     0.6953     -0.1800\n",
       "samples_per_sec          119.3642 118.9131            +0.4511  [-1.6984, +4.7605]     0.8457      0.0900"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Primary  val_dice_tail_mean  (k=1, α=0.05):\n",
      "  p=0.3223  →  retain H0\n",
      "\n",
      "Holm step-down  secondary family  (k=3, α_adj=0.0167):\n",
      "                              p  threshold  reject\n",
      "test                                              \n",
      "val_dice_max             0.4316     0.0167   False\n",
      "val_iou_max              0.4316     0.0250   False\n",
      "generalization_gap_final 0.6953     0.0500   False\n",
      "\n",
      "Throughput  samples_per_sec  (k=1, α=0.05):\n",
      "  p=0.8457  →  retain H0\n"
     ]
    }
   ],
   "source": [
    "results = build_comparison_table(\n",
    "    runs, METRICS_SPEC,\n",
    "    arch_a=COSINE, arch_b=PLATEAU, seeds=SEEDS,\n",
    "    alpha=ALPHA, n_resamples=N_BOOT, random_state=RNG,\n",
    ")\n",
    "show_comparison_table(results, label_a='cosine', label_b='plateau')\n",
    "show_family_verdicts(results, PRIMARY_METRIC, SECONDARY_METRICS, alpha=ALPHA)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cd871996",
   "metadata": {},
   "source": [
    "## 4. Figures"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "aa1648ff",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-24T10:16:59.449958Z",
     "iopub.status.busy": "2026-06-24T10:16:59.449486Z",
     "iopub.status.idle": "2026-06-24T10:17:00.504785Z",
     "shell.execute_reply": "2026-06-24T10:17:00.504402Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1210x660 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_paired_slopegraph(\n",
    "    runs, SLOPE_METRICS,\n",
    "    arch_a=COSINE, arch_b=PLATEAU, seeds=SEEDS, palette=PALETTE,\n",
    "    title=f'Fig 1 — Per-seed paired comparison (n={N})',\n",
    "    save_path=FIG_DIR / 'E3_fig1_paired_slopegraph.png',\n",
    ");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a369cef2",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-24T10:17:00.506617Z",
     "iopub.status.busy": "2026-06-24T10:17:00.506454Z",
     "iopub.status.idle": "2026-06-24T10:17:00.957948Z",
     "shell.execute_reply": "2026-06-24T10:17:00.957338Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 990x572 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_paired_forest(\n",
    "    results, FOREST_SPECS,\n",
    "    arch_a=COSINE, arch_b=PLATEAU, n=N, palette=PALETTE,\n",
    "    title=f'Fig 2 — Paired cosine−plateau differences, BCa 95% CI (n={N})',\n",
    "    save_path=FIG_DIR / 'E3_fig2_forest_paired_diff.png',\n",
    ");"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "26ac7cc7",
   "metadata": {},
   "source": [
    "## 4b. Post-hoc: cosine annealing vs flat LR (unpaired)\n",
    "\n",
    "E4 used flat LR (`use_lr_scheduler: false`) based on a 1-seed pilot. This is an unpaired comparison of E3 cosine (10 seeds) against E4 flat LR (10 seeds, same architecture and lr=3e-4, different experiment). Seeds are the same numbers (100–109) but runs were independent, so differences are unpaired. The strip plot shows the two distributions; the forest plot shows the mean difference with BCa 95 % CI."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "972f23d6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 550x440 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import sqlite3, re as _re\n",
    "\n",
    "# ── load E4 flat-LR plateau Dice ─────────────────────────────────────────────\n",
    "E4_DB_DIR = PROJECT_ROOT / 'mlruns'\n",
    "e4_plateau = {}\n",
    "for _db in sorted(E4_DB_DIR.glob('E4-isic2017-unet2d-thres-sweep*.db')):\n",
    "    _con = sqlite3.connect(_db)\n",
    "    for _uuid, _name in _con.execute(\"SELECT run_uuid, name FROM runs WHERE status='FINISHED'\").fetchall():\n",
    "        _m = _re.search(r'seed(\\d+)', _name)\n",
    "        if not _m: continue\n",
    "        _vals = [r[0] for r in _con.execute(\n",
    "            \"SELECT value FROM metrics WHERE run_uuid=? AND key='val_dice' ORDER BY step\",\n",
    "            (_uuid,)).fetchall()]\n",
    "        e4_plateau[int(_m.group(1))] = float(np.mean(_vals[-10:]))\n",
    "\n",
    "flat_vals = np.array([e4_plateau[s] for s in SEEDS])\n",
    "cos_vals  = np.array([runs.loc[runs['arch'] == COSINE, 'val_dice_tail_mean'].sort_values().values[i]\n",
    "                      for i, s in enumerate(SEEDS)])\n",
    "# pull cosine tail means in seed order\n",
    "cos_df = runs[runs['arch'] == COSINE].set_index('seed')['val_dice_tail_mean']\n",
    "cos_vals = np.array([cos_df[s] for s in SEEDS])\n",
    "\n",
    "FLAT_COL = '#edae49'\n",
    "\n",
    "import matplotlib.pyplot as _plt\n",
    "_rng = np.random.default_rng(0)\n",
    "_jitter = 0.06\n",
    "fig5, ax5 = _plt.subplots(figsize=(5, 4))\n",
    "ax5.scatter(np.zeros(len(cos_vals))  + _rng.uniform(-_jitter, _jitter, len(cos_vals)),\n",
    "            cos_vals, color=PALETTE[COSINE], s=60, zorder=3, label='cosine annealing (E3)')\n",
    "ax5.scatter(np.ones(len(flat_vals))  + _rng.uniform(-_jitter, _jitter, len(flat_vals)),\n",
    "            flat_vals, color=FLAT_COL, s=60, zorder=3, label='flat LR (E4)')\n",
    "ax5.hlines(np.mean(cos_vals),  -0.25, 0.25, colors=PALETTE[COSINE], linewidths=2)\n",
    "ax5.hlines(np.mean(flat_vals),  0.75, 1.25, colors=FLAT_COL,        linewidths=2)\n",
    "ax5.set_xticks([0, 1])\n",
    "ax5.set_xticklabels(['Cosine annealing\\n(E3)', 'Flat LR\\n(E4)'])\n",
    "ax5.set_ylabel('Plateau Dice (last-10-epoch mean)')\n",
    "ax5.set_title(f'Fig 5 — Cosine vs flat LR (unpaired, n={N})', fontsize=10)\n",
    "ax5.legend(fontsize=8, frameon=False)\n",
    "ax5.set_xlim(-0.5, 1.5)\n",
    "ax5.text(0, np.mean(cos_vals)+0.0008, f'{np.mean(cos_vals):.4f}', ha='center', fontsize=8, color=PALETTE[COSINE])\n",
    "ax5.text(1, np.mean(flat_vals)+0.0008, f'{np.mean(flat_vals):.4f}', ha='center', fontsize=8, color=FLAT_COL)\n",
    "fig5.tight_layout()\n",
    "fig5.savefig(FIG_DIR / 'E3_fig5_cosine_vs_flat_strip.png', dpi=150)\n",
    "_plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "b118d9f5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 550x242 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Δ=+0.0042, BCa 95% CI [+0.0015, +0.0069]\n"
     ]
    }
   ],
   "source": [
    "# BCa bootstrap CI on unpaired mean difference (cosine - flat)\n",
    "from scipy.stats import norm as _norm\n",
    "_obs = np.mean(cos_vals) - np.mean(flat_vals)\n",
    "_rng2 = np.random.default_rng(42)\n",
    "_boots = np.array([\n",
    "    np.mean(_rng2.choice(cos_vals, len(cos_vals), replace=True)) -\n",
    "    np.mean(_rng2.choice(flat_vals, len(flat_vals), replace=True))\n",
    "    for _ in range(N_BOOT)\n",
    "])\n",
    "_z0 = _norm.ppf(np.mean(_boots < _obs))\n",
    "_jk_a = np.array([np.mean(np.delete(cos_vals, i)) for i in range(len(cos_vals))])\n",
    "_jk_b = np.array([np.mean(np.delete(flat_vals, i)) for i in range(len(flat_vals))])\n",
    "_jk   = np.concatenate([_jk_a - np.mean(cos_vals), _jk_b - np.mean(flat_vals)])\n",
    "_acc  = np.sum((-_jk)**3) / (6*(np.sum(_jk**2))**1.5)\n",
    "_a1   = _norm.cdf(_z0 + (_z0 + _norm.ppf(0.025)) / (1 - _acc*(_z0 + _norm.ppf(0.025))))\n",
    "_a2   = _norm.cdf(_z0 + (_z0 + _norm.ppf(0.975)) / (1 - _acc*(_z0 + _norm.ppf(0.975))))\n",
    "_ci_lo, _ci_hi = np.percentile(_boots, 100*_a1), np.percentile(_boots, 100*_a2)\n",
    "\n",
    "fig6, ax6 = _plt.subplots(figsize=(5, 2.2))\n",
    "ax6.errorbar(_obs, 0, xerr=[[_obs - _ci_lo], [_ci_hi - _obs]],\n",
    "             fmt='o', color=PALETTE[COSINE], capsize=5, markersize=7, linewidth=2)\n",
    "ax6.axvline(0, color='#888888', linewidth=1, linestyle='--')\n",
    "ax6.set_yticks([])\n",
    "ax6.set_xlabel('Δ plateau Dice (cosine − flat LR)')\n",
    "ax6.set_title(f'Fig 6 — Mean difference BCa 95% CI (unpaired, n={N} each)', fontsize=10)\n",
    "ax6.text(_obs, 0.35, f'Δ={_obs:+.4f}\\n[{_ci_lo:+.4f}, {_ci_hi:+.4f}]',\n",
    "         ha='center', va='bottom', fontsize=8)\n",
    "ax6.set_ylim(-0.6, 0.9)\n",
    "fig6.tight_layout()\n",
    "fig6.savefig(FIG_DIR / 'E3_fig6_cosine_vs_flat_forest.png', dpi=150)\n",
    "_plt.show()\n",
    "print(f'Δ={_obs:+.4f}, BCa 95% CI [{_ci_lo:+.4f}, {_ci_hi:+.4f}]')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f154ac25",
   "metadata": {},
   "source": [
    "## 5. Decision\n",
    "\n",
    "> **Between the two decay schedules, use `cosine_annealing` (T_max = 300, η_min = 1×10⁻⁶) at `lr = 3e-4`.**\n",
    "\n",
    "| Priority | Criterion | Cosine | Plateau | Status (n = 10) |\n",
    "|---|---|---|---|---|\n",
    "| 1 | **Plateau Dice — *primary*** | 0.8346 | 0.8331 | tie — Δ = +0.0015, p = 0.32, CI [−0.0006, +0.0038] spans 0 |\n",
    "| 2 | Peak Dice / IoU | 0.8451 | **0.8459** | tie — Δ ≈ −0.0007 / −0.0010, p = 0.43, CI spans 0 |\n",
    "| 3 | Generalisation gap | 0.096 | 0.097 | tie — p = 0.70, CI spans 0 |\n",
    "| 4 | Training throughput | 119.4 sps | 118.9 sps | tie — Δ = +0.5 sps (+0.4 %), p = 0.85 |\n",
    "| — | Convergence epoch *(tie-breaker)* | **125.5** | 161.0 | cosine reaches best checkpoint ≈ 36 epochs earlier |\n",
    "\n",
    "**Rationale.** On every quality and cost axis the two schedules are statistically indistinguishable (no metric survives its family threshold; all CIs straddle 0; all |d_z| ≤ 0.38). With quality tied, the deciding factor is operational: cosine is a **deterministic, validation-independent** schedule whose LR trajectory is fully specified by epoch count, and it reaches its best checkpoint ~22 % earlier. \n",
    "\n",
    "**Decay-vs-no-decay (post-hoc).** A 1-seed pilot showed no quality difference between cosine and flat LR, so the scheduler was dropped before E4. Post-hoc unpaired comparison of E3 cosine (10 seeds) against E4 flat LR (10 seeds, same architecture and lr, different experiment purpose) on plateau Dice: Δ=+0.0042, BCa 95% CI [+0.0015, +0.0069] — entirely above zero (p=0.017, unpaired rank test). The 1-seed pilot was underpowered and missed this effect.\n",
    "\n",
    "> ⚠ **The E4 decision to drop the scheduler was made on insufficient evidence.** This comparison is unpaired — E3 and E4 were separate experiments with different purposes — so the finding is indicative rather than conclusive. A dedicated paired 10-seed flat-vs-cosine experiment would confirm it. Until then, E4 onward uses flat LR (`use_lr_scheduler: false`) as a pragmatic choice that is retrospectively questionable on plateau Dice."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}