Merge remote-tracking branch 'origin/main' into dev_1.4.1

Author: Beat Buesser

README-cn.md

@@ -15,9 +15,13 @@
 [![PyPI - Python Version](https://img.shields.io/pypi/pyversions/adversarial-robustness-toolbox)](https://pypi.org/project/adversarial-robustness-toolbox/)
 [![slack-img](https://img.shields.io/badge/chat-on%20slack-yellow.svg)](https://ibm-art.slack.com/)
 
-对抗性鲁棒性工具箱（ART）是用于机器学习安全性的Python库。 ART提供的工具使开发人员和研究人员可以针对以下方面评估，捍卫，认证和验证机器学习模型和应用程序：
-逃避，中毒，提取和推断的对抗性威胁。 ART支持所有数据类型的所有流行的机器学习框架（TensorFlow，Keras，PyTorch，MXNet，scikit-learn，XGBoost，LightGBM，CatBoost，GPy等）
-（图像，表格，音频，视频等）和机器学习任务（分类，对象检测，生成，认证等）。
+
+对抗性鲁棒性工具箱（ART）是用于机器学习安全性的Python库。 ART提供的工具可
+开发人员和研究人员针对以下方面捍卫和评估机器学习模型和应用程序：
+逃避，中毒，提取和推断的对抗性威胁。 ART支持所有流行的机器学习框架
+（TensorFlow，Keras，PyTorch，MXNet，scikit-learn，XGBoost，LightGBM，CatBoost，GPy等），所有数据类型
+（图像，表格，音频，视频等）和机器学习任务（分类，物体检测，语音识别，
+生成，认证等）。
 
 <p align="center">
   <img src="docs/images/adversarial_threats_attacker.png?raw=true" width="400" title="ART logo">

README.md

@@ -18,11 +18,11 @@
 [中文README请按此处](README-cn.md)
 
 Adversarial Robustness Toolbox (ART) is a Python library for Machine Learning Security. ART provides tools that enable
-developers and researchers to evaluate, defend, certify and verify Machine Learning models and applications against the
+developers and researchers to defend and evaluate Machine Learning models and applications against the
 adversarial threats of Evasion, Poisoning, Extraction, and Inference. ART supports all popular machine learning frameworks
 (TensorFlow, Keras, PyTorch, MXNet, scikit-learn, XGBoost, LightGBM, CatBoost, GPy, etc.), all data types
-(images, tables, audio, video, etc.) and machine learning tasks (classification, object detection, generation,
-certification, etc.).
+(images, tables, audio, video, etc.) and machine learning tasks (classification, object detection, speech recognition,
+generation, certification, etc.).
 
 <p align="center">
   <img src="docs/images/adversarial_threats_attacker.png?raw=true" width="400" title="ART logo">

art/estimators/object_detection/tensorflow_faster_rcnn.py

@@ -100,7 +100,7 @@ def __init__(
                               `first_stage_localization_loss`, `first_stage_objectness_loss`,
                               `second_stage_localization_loss`, `second_stage_classification_loss`.
         """
-        import tensorflow as tf
+        import tensorflow as tf  # lgtm [py/repeated-import]
 
         # Super initialization
         super().__init__(

requirements.txt

@@ -1,7 +1,7 @@
 # base
 numpy==1.19.2
 scipy==1.4.1
-matplotlib==3.3.1
+matplotlib==3.3.2
 scikit-learn==0.22.2
 six==1.15.0
 Pillow==7.2.0
@@ -11,7 +11,7 @@ pydub==0.24.1
 resampy==0.2.2
 ffmpeg-python==0.2.0
 cma==3.0.3
-pandas==1.1.1
+pandas==1.1.2
 
 # frameworks
 h5py==2.10.0
@@ -27,7 +27,7 @@ catboost==0.24.1
 GPy==1.9.9
 lightgbm==3.0.0
 xgboost==1.2.0
-kornia==0.3.1
+kornia==0.4.0
 
 # other
 pytest==5.4.1