Code Injection in jeikeilim/kindle

Description Kindle is an easy model build package for PyTorch. Building a deep learning model became so simple that almost all model can be made by copy and paste from other existing model codes, which is vulnerable to Arbitary Code Execution. Vulnerability Vulnerable to YAML deserialization atta...

in catalyst-team/catalyst

Description Catalyst is a PyTorch framework for Deep Learning research and development. It focuses on reproducibility, rapid experimentation, and codebase reuse so you can create something new rather than write another regular train loop. This package was vulnerable to Arbitrary code execution vi...

Code Injection in uber/petastorm

Description Petastorm is an open source data access library developed at Uber ATG. This library enables single machine or distributed training and evaluation of deep learning models directly from datasets in Apache Parquet format. Petastorm supports popular Python-based machine learning ML...

SoReL-20M: A Huge Dataset of 20 Million Malware Samples Released Online

Cybersecurity firms Sophos and ReversingLabs on Monday jointly released the first-ever production-scale malware research dataset to be made available to the general public that aims to build effective defenses and drive industry-wide improvements in security detection and response. "SoReL-20M"...

Security Bulletin: WML CE: Pillow before 7.1.0 has multiple out-of-bounds reads

Summary Pillow before 7.1.0 has multiple out-of-bounds reads in libImaging/FliDecode.c. PyTorch and TensorFlow use Pillow. Vulnerability Details CVEID: CVE-2020-10177 DESCRIPTION: Pillow could allow a remote attacker to obtain sensitive information, caused by multiple out-of-bounds reads in...

Security Bulletin: WML CE: In Pillow before 7.1.0, there is a Buffer Overflow

Summary In Pillow before 7.1.0, there are two Buffer Overflows in libImaging/TiffDecode.c. PyTorch and TensorFlow uses Pillow. Vulnerability Details CVEID: CVE-2020-10378 DESCRIPTION: Pillow could allow a remote attacker to obtain sensitive information, caused by an out-of-bounds read when readin...

Adversarial Robustness Toolbox: ART

The Adversarial Robustness Toolbox ART, an open source software library, supports both researchers and developers in defending deep neural networks against adversarial attacks, making AI systems more secure. Its purpose is to allow rapid crafting and analysis of attack and defense methods for...