Extracting Personal Information from Large Language Models Like GPT-2

Researchers have been able to find all sorts of personal information within GPT-2. This information was part of the training data, and can be extracted with the right sorts of queries. Paper: "Extracting Training Data from Large Language Models." Abstract: It has become common to publish large...

New Framework Released to Protect Machine Learning Systems From Adversarial Attacks

Microsoft, in collaboration with MITRE, IBM, NVIDIA, and Bosch, has released a new open framework that aims to help security analysts detect, respond to, and remediate adversarial attacks against machine learning ML systems. Called the Adversarial ML Threat Matrix, the initiative is an attempt to...

Cyberattacks against machine learning systems are more common than you think

Machine learning ML is making incredible transformations in critical areas such as finance, healthcare, and defense, impacting nearly every aspect of our lives. Many businesses, eager to capitalize on advancements in ML, have not scrutinized the security of their ML systems. Today, along with...

Cyberattacks against machine learning systems are more common than you think

Machine learning ML is making incredible transformations in critical areas such as finance, healthcare, and defense, impacting nearly every aspect of our lives. Many businesses, eager to capitalize on advancements in ML, have not scrutinized the security of their ML systems. Today, along with...

From unstructured data to actionable intelligence: Using machine learning for threat intelligence

The security community has become proficient in using indicators of compromise IoC feeds for threat intelligence. Automated feeds have simplified the task of extracting and sharing IoCs. However, IoCs like IP addresses, domain names, and file hashes are in the lowest levels of the threat...

Tools for capturing and analyzing keyboard input paired with microphone capture

The main goal is to exploit the sound produced by pressing keyboard keys as a side channel in order to guess the content of the text being typed. To achieve this, the algorithm takes as input a training set, consisting of an audio recording, together with the corresponding keys being typed during...

GyoiThon - A Growing Penetration Test Tool Using Machine Learning

GyoiThon is a growing penetration test tool using Machine Learning. GyoiThon identifies the software installed on web server OS, Middleware, Framework, CMS, etc... based on the learning data. After that, it executes valid exploits for the identified software using Metasploit. Finally, it generate...

Extracting Secrets from Machine Learning Systems

This is fascinating research about how the underlying training data for a machine-learning system can be inadvertently exposed. Basically, if a machine-learning system trains on a dataset that contains secret information, in some cases an attacker can query the system to extract that secret...

MADLIRA - Malware detection using learning and information retrieval for Android

MADLIRA is a tool for Android malware detection. It consists in two components: TFIDF component and SVM learning component. In gerneral, it takes an input a set of malwares and benwares and then extracts the malicious behaviors TFIDF component or computes training model SVM classifier. Then, it...