From f5ec52cdbd4a80a1edb8b86b2745212a1b546038 Mon Sep 17 00:00:00 2001
From: allenhzy <allenhzy3@gmail.com>
Date: Wed, 4 Sep 2024 00:22:59 +0800
Subject: [PATCH 01/14] BEYOND Detector and test case

---
 art/defences/detector/evasion/__init__.py     |   2 +
 .../detector/evasion/beyond_detector.py       | 163 ++++++++++++++++
 .../detector/evasion/test_beyond_detector.py  | 180 ++++++++++++++++++
 3 files changed, 345 insertions(+)
 create mode 100644 art/defences/detector/evasion/beyond_detector.py
 create mode 100644 tests/defences/detector/evasion/test_beyond_detector.py

diff --git a/art/defences/detector/evasion/__init__.py b/art/defences/detector/evasion/__init__.py
index 26112a2afe..3546431c38 100644
--- a/art/defences/detector/evasion/__init__.py
+++ b/art/defences/detector/evasion/__init__.py
@@ -6,3 +6,5 @@
 from art.defences.detector.evasion.binary_input_detector import BinaryInputDetector
 from art.defences.detector.evasion.binary_activation_detector import BinaryActivationDetector
 from art.defences.detector.evasion.subsetscanning.detector import SubsetScanningDetector
+from art.defences.detector.evasion.beyond_detector import BeyondDetector
+
diff --git a/art/defences/detector/evasion/beyond_detector.py b/art/defences/detector/evasion/beyond_detector.py
new file mode 100644
index 0000000000..5f0cccd10a
--- /dev/null
+++ b/art/defences/detector/evasion/beyond_detector.py
@@ -0,0 +1,163 @@
+# MIT License
+#
+# Copyright (C) The Adversarial Robustness Toolbox (ART) Authors 2023
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated
+# documentation files (the "Software"), to deal in the Software without restriction, including without limitation the
+# rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit
+# persons to whom the Software is furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in all copies or substantial portions of the
+# Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE
+# WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+# TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+"""
+This module implements the abstract base class for all evasion detectors.
+"""
+from __future__ import absolute_import, division, print_function, unicode_literals, annotations
+
+import abc
+from typing import Any
+
+import numpy as np
+
+from art.defences.detector.evasion.evasion_detector import EvasionDetector
+
+class BeyondDetector(EvasionDetector):
+    """
+    BEYOND detector for adversarial samples detection.
+    This detector uses a combination of SSL and target model predictions to detect adversarial samples.
+    """
+    
+    defence_params = ["target_model", "ssl_model", "augmentations", "aug_num", "alpha", "K", "percentile"]
+
+    def __init__(self,
+        target_model,
+        ssl_model,
+        augmentations=None,
+        aug_num=50,
+        alpha=0.8,
+        K=20,
+        percentile=5) -> None:
+        """
+        Initialize the BEYOND detector.
+
+        :param target_model: The target model to be protected
+        :param ssl_model: The self-supervised learning model used for feature extraction
+        :param augmentation: data augmentations for generating neighborhoods
+        :param aug_num: Number of augmentations to apply to each sample (default: 50)
+        :param alpha: Weight factor for combining label and representation similarities (default: 0.8)
+        :param K: Number of top similarities to consider (default: 20)
+        :param percentile: using to calculate the threshold
+        """
+        super().__init__()
+        self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
+        self.target_model = target_model.to(self.device)
+        self.ssl_model = ssl_model.to(self.device)
+        self.aug_num = aug_num
+        self.alpha = alpha
+        self.K = K
+
+        self.backbone = ssl_model.backbone
+        self.classifier = ssl_model.classifier
+        self.projector = ssl_model.projector
+
+        self.img_augmentations = augmentations
+
+        self.percentile = percentile # determinate the threshold
+        self.threshold = None
+
+    
+    
+    def _multi_transform(self, img):
+        return torch.stack([self.img_augmentations(img) for _ in range(self.aug_num)], dim=1)
+
+    def _get_metrics(self, x: np.ndarray, batch_size: int = 128) -> tuple[dict, np.ndarray]:
+        """
+        Calculate similarities that combining label consistency and representation similarity for given samples
+        :param x: Input samples
+        :param batch_size: Batch size for processing
+        :return: A report similarities
+        """
+        samples = torch.from_numpy(x).to(self.device)
+        
+        self.target_model.eval()
+        self.backbone.eval()
+        self.classifier.eval()
+        self.projector.eval()
+
+        number_batch = int(math.ceil(len(samples) / batch_size))
+        
+        similarities = []
+
+        with torch.no_grad():
+            for index in range(number_batch):
+                start = index * batch_size
+                end = min((index + 1) * batch_size, len(samples))
+
+                batch_samples = samples[start:end]
+                b, c, h, w = batch_samples.shape
+
+                trans_images = self._multi_transform(batch_samples).to(self.device)
+                ssl_backbone_out = self.backbone(batch_samples)
+
+                ssl_repre = self.projector(ssl_backbone_out)
+                ssl_pred = self.classifier(ssl_backbone_out)
+                ssl_label = torch.max(ssl_pred, -1)[1]
+
+                aug_backbone_out = self.backbone(trans_images.reshape(-1, c, h, w))
+                aug_repre = self.projector(aug_backbone_out)
+                aug_pred = self.classifier(aug_backbone_out)
+                aug_pred = aug_pred.reshape(b, self.aug_num, -1)
+
+                sim_repre = F.cosine_similarity(ssl_repre.unsqueeze(dim=1), aug_repre.reshape(b, self.aug_num, -1), dim=2)
+                sim_preds = F.cosine_similarity(F.one_hot(torch.argmax(ssl_label, dim=1), num_classes=ssl_pred.shape[-1]).unsqueeze(dim=1), aug_pred, dim=2)
+
+                similarities.append((self.alpha * sim_preds + (1-self.alpha)*sim_repre).sort(descending=True)[0].cpu().numpy())
+
+        similarities = np.concatenate(similarities, axis=0)
+        
+        return similarities
+
+
+    def fit(self, x: np.ndarray, y: np.ndarray, batch_size: int = 128, nb_epochs: int = 20, **kwargs) -> None:
+        """
+        Determine a threshold that covers 95% of clean samples.
+        :param x: Clean sample data
+        :param y: Clean sample labels (not used in this method)
+        :param batch_size: Batch size for processing
+        :param nb_epochs: Number of training epochs (not used in this method)
+        """
+        clean_similarities = self._get_metrics(x, batch_size)
+        
+        # 使用第K-1列的值来确定阈值
+        k_minus_one_metrics = clean_metrics[:, self.K-1]
+        
+        # 计算95%分位数作为阈值
+        self.threshold = np.percentile(k_minus_one_metrics, self.threshold)
+        
+        print(f"Threshold set to: {self.threshold}")
+
+    def detect(self, x: np.ndarray, batch_size: int = 128, **kwargs) -> tuple[dict, np.ndarray]:
+        """
+        Detect whether given samples are adversarial
+        :param x: Input samples
+        :param batch_size: Batch size for processing
+        :return: (report, is_adversarial):
+            where report containing detection results 
+            where is_adversarial is a boolean list indicating whether samples are adversarial or not
+        """
+        if self.threshold is None:
+            raise ValueError("Detector has not been fitted. Call fit() before detect().")
+        
+        similarities = self._get_metrics(x, batch_size)
+        
+        report = similarities[:, self.K-1]
+        is_adversarial = report < self.threshold
+        
+        return report, is_adversarial
diff --git a/tests/defences/detector/evasion/test_beyond_detector.py b/tests/defences/detector/evasion/test_beyond_detector.py
new file mode 100644
index 0000000000..9e0b6a3406
--- /dev/null
+++ b/tests/defences/detector/evasion/test_beyond_detector.py
@@ -0,0 +1,180 @@
+# MIT License
+#
+# Copyright (C) The Adversarial Robustness Toolbox (ART) Authors 2023
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated
+# documentation files (the "Software"), to deal in the Software without restriction, including without limitation the
+# rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit
+# persons to whom the Software is furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in all copies or substantial portions of the
+# Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE
+# WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+# TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+from __future__ import absolute_import, division, print_function, unicode_literals
+
+import logging
+import pytest
+import numpy as np
+
+import sys
+import os 
+
+from art.attacks.evasion.fast_gradient import FastGradientMethod
+from art.estimators.classification import PyTorchClassifier
+from art.defences.detector.evasion import BeyondDetector
+from art.utils import load_dataset, get_file
+
+
+from tests.utils import ARTTestException
+
+logger = logging.getLogger(__name__)
+
+import torch.nn as nn
+from torchvision import models, transforms
+
+class SimSiamWithCls(nn.Module):
+    '''
+    SimSiam with Classifier
+    '''
+    def __init__(self, arch='resnet18', feat_dim=2048, num_proj_layers=2):
+        
+        super(SimSiamWithCls, self).__init__()
+        self.backbone = models.resnet18()
+        out_dim = self.backbone.fc.weight.shape[1]
+        self.backbone.conv1 = nn.Conv2d(
+                    3, 64, kernel_size=3, stride=1, padding=2, bias=False
+                )
+        self.backbone.maxpool = nn.Identity()
+        self.backbone.fc = nn.Identity()
+        self.classifier = nn.Linear(out_dim, 10)
+
+        pred_hidden_dim = int(feat_dim / 4)
+
+        self.projector = nn.Sequential(
+            nn.Linear(out_dim, feat_dim, bias=False),
+            nn.BatchNorm1d(feat_dim),
+            nn.ReLU(),
+            nn.Linear(feat_dim, feat_dim, bias=False),
+            nn.BatchNorm1d(feat_dim),
+            nn.ReLU(),
+            nn.Linear(feat_dim, feat_dim),
+            nn.BatchNorm1d(feat_dim, affine=False),
+        )
+        self.projector[6].bias.requires_grad = False
+
+        self.predictor = nn.Sequential(
+            nn.Linear(feat_dim, pred_hidden_dim, bias=False),
+            nn.BatchNorm1d(pred_hidden_dim),
+            nn.ReLU(),
+            nn.Linear(pred_hidden_dim, feat_dim),
+        )
+
+    def forward(self, img, im_aug1=None, im_aug2=None):
+
+        r_ori = self.backbone(img)
+        if im_aug1 is None and im_aug2 is None:
+            cls = self.classifier(r_ori)
+            rep = self.projector(r_ori)
+            return {'cls': cls, 'rep':rep}
+        else:
+
+            r1 = self.backbone(im_aug1)
+            r2 = self.backbone(im_aug2)
+
+            z1 = self.projector(r1)
+            z2 = self.projector(r2)
+            # print("shape of z:", z1.shape)
+
+            p1 = self.predictor(z1)
+            p2 = self.predictor(z2)
+            # print("shape of p:", p1.shape)
+
+            return {'z1': z1, 'z2': z2, 'p1': p1, 'p2': p2}
+
+@pytest.fixture
+def get_cifar10():
+    """
+    Loads CIFAR10 dataset.
+    """
+    (x_train, y_train), (x_test, y_test), min_, max_ = load_cifar10()
+    return (x_train, y_train), (x_test, y_test), min_, max_
+
+
+@pytest.fixture
+def get_ssl_model(weights_path):
+    """
+    Loads the SSL model (SimSiamWithCls).
+    """
+    model = SimSiamWithCls()
+    model.load_state_dict(torch.load(weights_path))
+    return model
+
+@pytest.mark.only_with_platform("pytorch")
+def test_beyond_detector(art_warning, get_cifar10, get_ssl_model):
+    try:
+        # Load CIFAR10 data
+        (x_train, y_train), (x_test, y_test), min_, max_ = get_cifar10
+
+        # Load models
+        # Download pretrained weights from https://drive.google.com/drive/folders/1ieEdd7hOj2CIl1FQfu4-3RGZmEj-mesi?usp=sharing
+        target_model = models.resnet18()
+        target_model.load_state_dict(torch.load("./resnet_c10.pth"))
+        ssl_model = get_ssl_model()
+        ssl_model.load_state_dict(torch.load("./simsiam_c10.pth"))
+
+
+        # Generate adversarial samples
+        attack = FastGradientMethod(estimator=target_model, eps=0.05)
+        x_test_adv = attack.generate(x_test)
+
+        img_augmentations = transforms.Compose([
+            transforms.RandomResizedCrop(32, scale=(0.2, 1.)),
+            transforms.RandomHorizontalFlip(),
+            transforms.RandomApply([
+                transforms.ColorJitter(0.4, 0.4, 0.4, 0.1)  # not strengthened
+            ], p=0.8),
+            transforms.RandomGrayscale(p=0.2),
+            transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))
+        ])
+
+        # Initialize BeyondDetector
+        detector = BeyondDetector(
+            target_model=target_model,
+            ssl_model=ssl_model,
+            img_augmentation=img_augmentations,
+            aug_num=50,
+            alpha=0.8,
+            K=20,
+            percentile=5
+        )
+
+        # Fit the detector
+        detector.fit(x_train, y_train, batch_size=128)
+
+        # Apply detector on clean and adversarial test data
+        _, test_detection = detector.detect(x_test)
+        _, test_adv_detection = detector.detect(x_test_adv)
+
+        # Assert there is at least one true positive and negative
+        nb_true_positives = np.sum(test_adv_detection)
+        nb_true_negatives = len(test_detection) - np.sum(test_detection)
+        assert nb_true_positives > 0
+        assert nb_true_negatives > 0
+
+        # Calculate and print detection accuracy
+        clean_accuracy = 1 - np.mean(test_detection)
+        adv_accuracy = np.mean(test_adv_detection)
+        print(f"Clean Detection Accuracy: {clean_accuracy:.4f}")
+        print(f"Adversarial Detection Accuracy: {adv_accuracy:.4f}")
+
+    except ARTTestException as e:
+        art_warning(e)
+
+if __name__ == "__main__":
+
+    test_beyond_detector()
\ No newline at end of file

From d9130cf6a923bcd18227438687fbe169957cdabc Mon Sep 17 00:00:00 2001
From: allenhzy <allenhzy3@gmail.com>
Date: Thu, 5 Dec 2024 23:26:16 +0800
Subject: [PATCH 02/14] revision by comments

---
 .../detector/evasion/beyond_detector.py       | 19 ++++++--------
 .../detector/evasion/test_beyond_detector.py  | 25 +++----------------
 2 files changed, 10 insertions(+), 34 deletions(-)

diff --git a/art/defences/detector/evasion/beyond_detector.py b/art/defences/detector/evasion/beyond_detector.py
index 5f0cccd10a..8206c0278e 100644
--- a/art/defences/detector/evasion/beyond_detector.py
+++ b/art/defences/detector/evasion/beyond_detector.py
@@ -1,6 +1,6 @@
 # MIT License
 #
-# Copyright (C) The Adversarial Robustness Toolbox (ART) Authors 2023
+# Copyright (C) The Adversarial Robustness Toolbox (ART) Authors 2024
 #
 # Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated
 # documentation files (the "Software"), to deal in the Software without restriction, including without limitation the
@@ -16,21 +16,19 @@
 # TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 # SOFTWARE.
 """
-This module implements the abstract base class for all evasion detectors.
+This module implements the BEYOND detector for adversarial examples detection.
+| Paper link: https://openreview.net/pdf?id=S4LqI6CcJ3
 """
-from __future__ import absolute_import, division, print_function, unicode_literals, annotations
-
-import abc
-from typing import Any
-
 import numpy as np
 
 from art.defences.detector.evasion.evasion_detector import EvasionDetector
 
 class BeyondDetector(EvasionDetector):
     """
-    BEYOND detector for adversarial samples detection.
-    This detector uses a combination of SSL and target model predictions to detect adversarial samples.
+      BEYOND detector for adversarial samples detection.
+    This detector uses a combination of SSL and target model predictions to detect adversarial examples.
+    
+    | Paper link: https://openreview.net/pdf?id=S4LqI6CcJ3
     """
     
     defence_params = ["target_model", "ssl_model", "augmentations", "aug_num", "alpha", "K", "percentile"]
@@ -133,9 +131,6 @@ def fit(self, x: np.ndarray, y: np.ndarray, batch_size: int = 128, nb_epochs: in
         :param batch_size: Batch size for processing
         :param nb_epochs: Number of training epochs (not used in this method)
         """
-        clean_similarities = self._get_metrics(x, batch_size)
-        
-        # 使用第K-1列的值来确定阈值
         k_minus_one_metrics = clean_metrics[:, self.K-1]
         
         # 计算95%分位数作为阈值
diff --git a/tests/defences/detector/evasion/test_beyond_detector.py b/tests/defences/detector/evasion/test_beyond_detector.py
index 9e0b6a3406..7747e9de2b 100644
--- a/tests/defences/detector/evasion/test_beyond_detector.py
+++ b/tests/defences/detector/evasion/test_beyond_detector.py
@@ -21,11 +21,7 @@
 import pytest
 import numpy as np
 
-import sys
-import os 
-
 from art.attacks.evasion.fast_gradient import FastGradientMethod
-from art.estimators.classification import PyTorchClassifier
 from art.defences.detector.evasion import BeyondDetector
 from art.utils import load_dataset, get_file
 
@@ -88,22 +84,12 @@ def forward(self, img, im_aug1=None, im_aug2=None):
 
             z1 = self.projector(r1)
             z2 = self.projector(r2)
-            # print("shape of z:", z1.shape)
 
             p1 = self.predictor(z1)
             p2 = self.predictor(z2)
-            # print("shape of p:", p1.shape)
 
             return {'z1': z1, 'z2': z2, 'p1': p1, 'p2': p2}
 
-@pytest.fixture
-def get_cifar10():
-    """
-    Loads CIFAR10 dataset.
-    """
-    (x_train, y_train), (x_test, y_test), min_, max_ = load_cifar10()
-    return (x_train, y_train), (x_test, y_test), min_, max_
-
 
 @pytest.fixture
 def get_ssl_model(weights_path):
@@ -115,10 +101,10 @@ def get_ssl_model(weights_path):
     return model
 
 @pytest.mark.only_with_platform("pytorch")
-def test_beyond_detector(art_warning, get_cifar10, get_ssl_model):
+def test_beyond_detector(art_warning, load_cifar10, get_ssl_model):
     try:
         # Load CIFAR10 data
-        (x_train, y_train), (x_test, y_test), min_, max_ = get_cifar10
+        (x_train, y_train), (x_test, y_test), _, _ = load_cifar10()
 
         # Load models
         # Download pretrained weights from https://drive.google.com/drive/folders/1ieEdd7hOj2CIl1FQfu4-3RGZmEj-mesi?usp=sharing
@@ -146,7 +132,7 @@ def test_beyond_detector(art_warning, get_cifar10, get_ssl_model):
         detector = BeyondDetector(
             target_model=target_model,
             ssl_model=ssl_model,
-            img_augmentation=img_augmentations,
+            augmentations=img_augmentations,
             aug_num=50,
             alpha=0.8,
             K=20,
@@ -163,14 +149,9 @@ def test_beyond_detector(art_warning, get_cifar10, get_ssl_model):
         # Assert there is at least one true positive and negative
         nb_true_positives = np.sum(test_adv_detection)
         nb_true_negatives = len(test_detection) - np.sum(test_detection)
-        assert nb_true_positives > 0
-        assert nb_true_negatives > 0
 
-        # Calculate and print detection accuracy
         clean_accuracy = 1 - np.mean(test_detection)
         adv_accuracy = np.mean(test_adv_detection)
-        print(f"Clean Detection Accuracy: {clean_accuracy:.4f}")
-        print(f"Adversarial Detection Accuracy: {adv_accuracy:.4f}")
 
     except ARTTestException as e:
         art_warning(e)

From c2ec510bb2c091a24fa8ba2388d93db85217dab9 Mon Sep 17 00:00:00 2001
From: allenhzy <allenhzy3@gmail.com>
Date: Thu, 5 Dec 2024 23:35:47 +0800
Subject: [PATCH 03/14] revision by comments

---
 tests/defences/detector/evasion/test_beyond_detector.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tests/defences/detector/evasion/test_beyond_detector.py b/tests/defences/detector/evasion/test_beyond_detector.py
index 7747e9de2b..fa7435e1bf 100644
--- a/tests/defences/detector/evasion/test_beyond_detector.py
+++ b/tests/defences/detector/evasion/test_beyond_detector.py
@@ -1,6 +1,6 @@
 # MIT License
 #
-# Copyright (C) The Adversarial Robustness Toolbox (ART) Authors 2023
+# Copyright (C) The Adversarial Robustness Toolbox (ART) Authors 2024
 #
 # Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated
 # documentation files (the "Software"), to deal in the Software without restriction, including without limitation the

From 32cb990f91a35ebee073b7af4636ba159afa6577 Mon Sep 17 00:00:00 2001
From: allenhzy <allenhzy3@gmail.com>
Date: Sat, 7 Dec 2024 00:05:44 +0800
Subject: [PATCH 04/14] revision by comment

---
 art/defences/detector/evasion/beyond_detector.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/art/defences/detector/evasion/beyond_detector.py b/art/defences/detector/evasion/beyond_detector.py
index 8206c0278e..11e6527f84 100644
--- a/art/defences/detector/evasion/beyond_detector.py
+++ b/art/defences/detector/evasion/beyond_detector.py
@@ -133,7 +133,6 @@ def fit(self, x: np.ndarray, y: np.ndarray, batch_size: int = 128, nb_epochs: in
         """
         k_minus_one_metrics = clean_metrics[:, self.K-1]
         
-        # 计算95%分位数作为阈值
         self.threshold = np.percentile(k_minus_one_metrics, self.threshold)
         
         print(f"Threshold set to: {self.threshold}")

From 4628d64ba3a3413b291ddf7c248dd4e9a96701ac Mon Sep 17 00:00:00 2001
From: allenhzy <allenhzy3@gmail.com>
Date: Sat, 7 Dec 2024 00:08:22 +0800
Subject: [PATCH 05/14] remove print

---
 art/defences/detector/evasion/beyond_detector.py | 2 --
 1 file changed, 2 deletions(-)

diff --git a/art/defences/detector/evasion/beyond_detector.py b/art/defences/detector/evasion/beyond_detector.py
index 11e6527f84..848fab3b2b 100644
--- a/art/defences/detector/evasion/beyond_detector.py
+++ b/art/defences/detector/evasion/beyond_detector.py
@@ -134,8 +134,6 @@ def fit(self, x: np.ndarray, y: np.ndarray, batch_size: int = 128, nb_epochs: in
         k_minus_one_metrics = clean_metrics[:, self.K-1]
         
         self.threshold = np.percentile(k_minus_one_metrics, self.threshold)
-        
-        print(f"Threshold set to: {self.threshold}")
 
     def detect(self, x: np.ndarray, batch_size: int = 128, **kwargs) -> tuple[dict, np.ndarray]:
         """

From 2e7b1a453c6ae752d17477ba181b528272a385c4 Mon Sep 17 00:00:00 2001
From: allenhzy <allenhzy3@gmail.com>
Date: Sat, 7 Dec 2024 00:19:41 +0800
Subject: [PATCH 06/14] adding Type of arguments

---
 .../detector/evasion/beyond_detector.py       | 22 ++++++++++++-------
 1 file changed, 14 insertions(+), 8 deletions(-)

diff --git a/art/defences/detector/evasion/beyond_detector.py b/art/defences/detector/evasion/beyond_detector.py
index 848fab3b2b..ae1cba33e0 100644
--- a/art/defences/detector/evasion/beyond_detector.py
+++ b/art/defences/detector/evasion/beyond_detector.py
@@ -20,6 +20,12 @@
 | Paper link: https://openreview.net/pdf?id=S4LqI6CcJ3
 """
 import numpy as np
+from typing import TYPE_CHECKING
+if TYPE_CHECKING:
+    from art.utils import CLASSIFIER_NEURALNETWORK_TYPE
+
+
+logger = logging.getLogger(__name__)
 
 from art.defences.detector.evasion.evasion_detector import EvasionDetector
 
@@ -34,13 +40,13 @@ class BeyondDetector(EvasionDetector):
     defence_params = ["target_model", "ssl_model", "augmentations", "aug_num", "alpha", "K", "percentile"]
 
     def __init__(self,
-        target_model,
-        ssl_model,
-        augmentations=None,
-        aug_num=50,
-        alpha=0.8,
-        K=20,
-        percentile=5) -> None:
+        target_model: "CLASSIFIER_NEURALNETWORK_TYPE",
+        ssl_model: "CLASSIFIER_NEURALNETWORK_TYPE",
+        augmentations: Optional[Callable] = None,
+        aug_num: int=50,
+        alpha: float=0.8,
+        K:int=20,
+        percentile:int=5) -> None:
         """
         Initialize the BEYOND detector.
 
@@ -72,7 +78,7 @@ def __init__(self,
 
     
     
-    def _multi_transform(self, img):
+    def _multi_transform(self, img: torch.Tensor) -> torch.Tensor:
         return torch.stack([self.img_augmentations(img) for _ in range(self.aug_num)], dim=1)
 
     def _get_metrics(self, x: np.ndarray, batch_size: int = 128) -> tuple[dict, np.ndarray]:

From 5a92ea31f7bd47a22e186ce9d20b931920b6a56a Mon Sep 17 00:00:00 2001
From: allenhzy <allenhzy3@gmail.com>
Date: Sat, 7 Dec 2024 00:23:29 +0800
Subject: [PATCH 07/14] method comments

---
 art/defences/detector/evasion/beyond_detector.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/art/defences/detector/evasion/beyond_detector.py b/art/defences/detector/evasion/beyond_detector.py
index ae1cba33e0..9b1ac81a90 100644
--- a/art/defences/detector/evasion/beyond_detector.py
+++ b/art/defences/detector/evasion/beyond_detector.py
@@ -84,6 +84,7 @@ def _multi_transform(self, img: torch.Tensor) -> torch.Tensor:
     def _get_metrics(self, x: np.ndarray, batch_size: int = 128) -> tuple[dict, np.ndarray]:
         """
         Calculate similarities that combining label consistency and representation similarity for given samples
+
         :param x: Input samples
         :param batch_size: Batch size for processing
         :return: A report similarities
@@ -132,6 +133,7 @@ def _get_metrics(self, x: np.ndarray, batch_size: int = 128) -> tuple[dict, np.n
     def fit(self, x: np.ndarray, y: np.ndarray, batch_size: int = 128, nb_epochs: int = 20, **kwargs) -> None:
         """
         Determine a threshold that covers 95% of clean samples.
+
         :param x: Clean sample data
         :param y: Clean sample labels (not used in this method)
         :param batch_size: Batch size for processing
@@ -144,6 +146,7 @@ def fit(self, x: np.ndarray, y: np.ndarray, batch_size: int = 128, nb_epochs: in
     def detect(self, x: np.ndarray, batch_size: int = 128, **kwargs) -> tuple[dict, np.ndarray]:
         """
         Detect whether given samples are adversarial
+        
         :param x: Input samples
         :param batch_size: Batch size for processing
         :return: (report, is_adversarial):

From 7902e01438aba7d4c2fa2c22a60713c1255ae665 Mon Sep 17 00:00:00 2001
From: allenhzy <allenhzy3@gmail.com>
Date: Sat, 7 Dec 2024 00:26:32 +0800
Subject: [PATCH 08/14] delete unused import

---
 tests/defences/detector/evasion/test_beyond_detector.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tests/defences/detector/evasion/test_beyond_detector.py b/tests/defences/detector/evasion/test_beyond_detector.py
index fa7435e1bf..7da225de9e 100644
--- a/tests/defences/detector/evasion/test_beyond_detector.py
+++ b/tests/defences/detector/evasion/test_beyond_detector.py
@@ -21,16 +21,16 @@
 import pytest
 import numpy as np
 
+import torch.nn as nn
+
 from art.attacks.evasion.fast_gradient import FastGradientMethod
 from art.defences.detector.evasion import BeyondDetector
-from art.utils import load_dataset, get_file
 
 
 from tests.utils import ARTTestException
 
 logger = logging.getLogger(__name__)
 
-import torch.nn as nn
 from torchvision import models, transforms
 
 class SimSiamWithCls(nn.Module):

From d5a0cd9bab02a11a3ddbc8ddc9e630e6642f680c Mon Sep 17 00:00:00 2001
From: allenhzy <allenhzy3@gmail.com>
Date: Tue, 10 Dec 2024 16:13:51 +0800
Subject: [PATCH 09/14] follow codeQL and remove unused variable

---
 art/defences/detector/evasion/beyond_detector.py        | 3 +--
 tests/defences/detector/evasion/test_beyond_detector.py | 2 --
 2 files changed, 1 insertion(+), 4 deletions(-)

diff --git a/art/defences/detector/evasion/beyond_detector.py b/art/defences/detector/evasion/beyond_detector.py
index 9b1ac81a90..fdfa5aaf79 100644
--- a/art/defences/detector/evasion/beyond_detector.py
+++ b/art/defences/detector/evasion/beyond_detector.py
@@ -17,6 +17,7 @@
 # SOFTWARE.
 """
 This module implements the BEYOND detector for adversarial examples detection.
+
 | Paper link: https://openreview.net/pdf?id=S4LqI6CcJ3
 """
 import numpy as np
@@ -25,8 +26,6 @@
     from art.utils import CLASSIFIER_NEURALNETWORK_TYPE
 
 
-logger = logging.getLogger(__name__)
-
 from art.defences.detector.evasion.evasion_detector import EvasionDetector
 
 class BeyondDetector(EvasionDetector):
diff --git a/tests/defences/detector/evasion/test_beyond_detector.py b/tests/defences/detector/evasion/test_beyond_detector.py
index 7da225de9e..9cca7f54a4 100644
--- a/tests/defences/detector/evasion/test_beyond_detector.py
+++ b/tests/defences/detector/evasion/test_beyond_detector.py
@@ -29,8 +29,6 @@
 
 from tests.utils import ARTTestException
 
-logger = logging.getLogger(__name__)
-
 from torchvision import models, transforms
 
 class SimSiamWithCls(nn.Module):

From f197bbef68001b551331c8a65b0eb34fb0119238 Mon Sep 17 00:00:00 2001
From: Beat Buesser <49047826+beat-buesser@users.noreply.github.com>
Date: Wed, 11 Dec 2024 22:47:26 +0100
Subject: [PATCH 10/14] Update
 tests/defences/detector/evasion/test_beyond_detector.py

---
 tests/defences/detector/evasion/test_beyond_detector.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/tests/defences/detector/evasion/test_beyond_detector.py b/tests/defences/detector/evasion/test_beyond_detector.py
index 9cca7f54a4..4d8737e1d3 100644
--- a/tests/defences/detector/evasion/test_beyond_detector.py
+++ b/tests/defences/detector/evasion/test_beyond_detector.py
@@ -17,7 +17,6 @@
 # SOFTWARE.
 from __future__ import absolute_import, division, print_function, unicode_literals
 
-import logging
 import pytest
 import numpy as np
 

From a7afbd1b575cf67251256a77e048cc0212d283a1 Mon Sep 17 00:00:00 2001
From: Beat Buesser <49047826+beat-buesser@users.noreply.github.com>
Date: Wed, 11 Dec 2024 22:47:37 +0100
Subject: [PATCH 11/14] Update art/defences/detector/evasion/beyond_detector.py

---
 art/defences/detector/evasion/beyond_detector.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/art/defences/detector/evasion/beyond_detector.py b/art/defences/detector/evasion/beyond_detector.py
index fdfa5aaf79..e420759745 100644
--- a/art/defences/detector/evasion/beyond_detector.py
+++ b/art/defences/detector/evasion/beyond_detector.py
@@ -41,7 +41,7 @@ class BeyondDetector(EvasionDetector):
     def __init__(self,
         target_model: "CLASSIFIER_NEURALNETWORK_TYPE",
         ssl_model: "CLASSIFIER_NEURALNETWORK_TYPE",
-        augmentations: Optional[Callable] = None,
+        augmentations: Callable | None,
         aug_num: int=50,
         alpha: float=0.8,
         K:int=20,

From 94c6cedef491f55d6033fb73162d622ff9ece870 Mon Sep 17 00:00:00 2001
From: Beat Buesser <49047826+beat-buesser@users.noreply.github.com>
Date: Wed, 11 Dec 2024 22:47:49 +0100
Subject: [PATCH 12/14] Update art/defences/detector/evasion/beyond_detector.py

---
 art/defences/detector/evasion/beyond_detector.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/art/defences/detector/evasion/beyond_detector.py b/art/defences/detector/evasion/beyond_detector.py
index e420759745..af4bf81058 100644
--- a/art/defences/detector/evasion/beyond_detector.py
+++ b/art/defences/detector/evasion/beyond_detector.py
@@ -20,6 +20,8 @@
 
 | Paper link: https://openreview.net/pdf?id=S4LqI6CcJ3
 """
+from __future__ import annotations
+
 import numpy as np
 from typing import TYPE_CHECKING
 if TYPE_CHECKING:

From 318f9deac42c94a4e6283b91f8b3120090534da5 Mon Sep 17 00:00:00 2001
From: allenhzy <allenhzy3@gmail.com>
Date: Thu, 12 Dec 2024 22:31:21 +0800
Subject: [PATCH 13/14] remove unused import

---
 art/defences/detector/evasion/beyond_detector.py     | 12 ++++++++----
 .../detector/evasion/test_beyond_detector.py         |  4 +---
 2 files changed, 9 insertions(+), 7 deletions(-)

diff --git a/art/defences/detector/evasion/beyond_detector.py b/art/defences/detector/evasion/beyond_detector.py
index fdfa5aaf79..66b73abab3 100644
--- a/art/defences/detector/evasion/beyond_detector.py
+++ b/art/defences/detector/evasion/beyond_detector.py
@@ -20,14 +20,18 @@
 
 | Paper link: https://openreview.net/pdf?id=S4LqI6CcJ3
 """
-import numpy as np
+from __future__ import annotations
+
 from typing import TYPE_CHECKING
-if TYPE_CHECKING:
-    from art.utils import CLASSIFIER_NEURALNETWORK_TYPE
 
+import numpy as np
 
 from art.defences.detector.evasion.evasion_detector import EvasionDetector
 
+if TYPE_CHECKING:
+    from art.utils import CLASSIFIER_NEURALNETWORK_TYPE
+
+
 class BeyondDetector(EvasionDetector):
     """
       BEYOND detector for adversarial samples detection.
@@ -41,7 +45,7 @@ class BeyondDetector(EvasionDetector):
     def __init__(self,
         target_model: "CLASSIFIER_NEURALNETWORK_TYPE",
         ssl_model: "CLASSIFIER_NEURALNETWORK_TYPE",
-        augmentations: Optional[Callable] = None,
+        augmentations: Callable | None,
         aug_num: int=50,
         alpha: float=0.8,
         K:int=20,
diff --git a/tests/defences/detector/evasion/test_beyond_detector.py b/tests/defences/detector/evasion/test_beyond_detector.py
index 9cca7f54a4..b4f2ced48d 100644
--- a/tests/defences/detector/evasion/test_beyond_detector.py
+++ b/tests/defences/detector/evasion/test_beyond_detector.py
@@ -17,19 +17,17 @@
 # SOFTWARE.
 from __future__ import absolute_import, division, print_function, unicode_literals
 
-import logging
 import pytest
 import numpy as np
 
 import torch.nn as nn
+from torchvision import models, transforms
 
 from art.attacks.evasion.fast_gradient import FastGradientMethod
 from art.defences.detector.evasion import BeyondDetector
 
-
 from tests.utils import ARTTestException
 
-from torchvision import models, transforms
 
 class SimSiamWithCls(nn.Module):
     '''

From f6e337162d2775a01e379d63dddc35b9e12aeda3 Mon Sep 17 00:00:00 2001
From: allenhzy <allenhzy3@gmail.com>
Date: Thu, 12 Dec 2024 22:41:32 +0800
Subject: [PATCH 14/14] follow import standard

---
 art/defences/detector/evasion/beyond_detector.py | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/art/defences/detector/evasion/beyond_detector.py b/art/defences/detector/evasion/beyond_detector.py
index e43c3778b9..66b73abab3 100644
--- a/art/defences/detector/evasion/beyond_detector.py
+++ b/art/defences/detector/evasion/beyond_detector.py
@@ -22,10 +22,6 @@
 """
 from __future__ import annotations
 
-<<<<<<< HEAD
-=======
-import numpy as np
->>>>>>> 94c6cedef491f55d6033fb73162d622ff9ece870
 from typing import TYPE_CHECKING
 
 import numpy as np