When Machine Learning Meets Vulnerability Discovery: Challenges and Lessons Learned

In recent years, machine learning has demonstrated impressive results in various fields, including software vulnerability detection. Nonetheless, using machine learning to identify software vulnerabilities presents new challenges, especially regarding the scale of data involved, which was not a...

CVE-2025-50461

A deserialization vulnerability exists in Volcengine's verl 3.0.0, specifically in the scripts/modelmerger.py script when using the "fsdp" backend. The script calls torch.load with weightsonly=False on user-supplied .pt files, allowing attackers to execute arbitrary code if a maliciously crafted...

microcode_ctl: From CVEorg collector

New Spectre-v2 attack classes have been discovered within CPU architectures that enable self-training exploitation of speculative execution within the same privilege domain. These novel techniques bypass existing hardware and software mitigations, including IBPB, eIBRS, and BHINO, by leveraging...

Unlearning at Scale: Implementing the Right to Be Forgotten in Large Language Models

We study the right to be forgotten GDPR Art. 17 for large language models and frame unlearning as a reproducible systems problem. Our approach treats training as a deterministic program and logs a minimal per-microbatch record ordered ID hash, RNG seed, learning-rate value, optimizer-step counter...

RMSL: Weakly-Supervised Insider Threat Detection with Robust Multi-Sphere Learning

Insider threat detection aims to identify malicious user behavior by analyzing logs that record user interactions. Due to the lack of fine-grained behavior-level annotations, detecting specific behavior-level anomalies within user behavior sequences is challenging. Unsupervised methods face high...

CVE-2025-23306

NVIDIA Megatron-LM for all platforms contains a vulnerability in the megatron/training/ arguments.py component where an attacker could cause a code injection issue by providing a malicious input. A successful exploit of this vulnerability may lead to code execution, escalation of privileges,...

CVE-2025-23306

NVIDIA Megatron-LM for all platforms contains a vulnerability in the megatron/training/ arguments.py component where an attacker could cause a code injection issue by providing a malicious input. A successful exploit of this vulnerability may lead to code execution, escalation of privileges,...

CVE-2025-23306

NVIDIA Megatron-LM for all platforms contains a vulnerability in the megatron/training/ arguments.py component where an attacker could cause a code injection issue by providing a malicious input. A successful exploit of this vulnerability may lead to code execution, escalation of privileges,...

CVE-2025-23306

CVE-2025-23306 affects NVIDIA Megatron-LM, specifically the megatron/training/arguments.py component. The vulnerability permits code injection through malicious input, with potential outcomes including code execution, privilege escalation, information disclosure, and data tampering. Public refere...

PT-2025-33047 · Nvidia · Nvidia Megatron-Lm

Name of the Vulnerable Software and Affected Versions: NVIDIA Megatron-LM affected versions not specified Description: NVIDIA Megatron-LM contains an issue in the megatron/training/arguments.py component that may allow an attacker to inject code by providing malicious input. A successful exploit...

NVIDIA Megatron-LM 代码注入漏洞

NVIDIA Megatron-LM is a PyTorch-based distributed training framework from NVIDIA that specializes in training large Transformer language models. NVIDIA Megatron-LM suffers from a code injection vulnerability that originates in the megatron/training/arguments.py component, which can be exploited b...

NVIDIA Megatron-LM 代码注入漏洞

NVIDIA Megatron-LM is a PyTorch-based distributed training framework from NVIDIA that specializes in training large Transformer language models. NVIDIA Megatron-LM suffers from a code injection vulnerability that originates in a tool component and can be exploited by an attacker to modify the...

INE Named to Training Industry’s 2025 Top 20 Online Learning Library List

Cary, United States, 11th August 2025, CyberNewsWire...

VOIDFace: a Privacy-Preserving Multi-Network Face Recognition with Enhanced Security

Advancement of machine learning techniques, combined with the availability of large-scale datasets, has significantly improved the accuracy and efficiency of facial recognition. Modern facial recognition systems are trained using large face datasets collected from diverse individuals or public...

CVE-2025-8752

A vulnerability was found in wangzhixuan spring-shiro-training up to 94812c1fd8f7fe796c931f4984ff1aa0671ab562. It has been declared as critical. This vulnerability affects unknown code of the file /role/add. The manipulation leads to command injection. The attack can be initiated remotely. The...

CVE-2025-8752

The CVE-2025-8752 entry concerns the wangzhixuan spring-shiro-training project (up to commit 94812c1fd8f7fe796c931f4984ff1aa0671ab562). The vulnerability is in the /role/add code path and is due to a command injection vulnerability. It is exploitable remotely and has been publicly disclosed. The ...

CVE-2025-8752 wangzhixuan spring-shiro-training add command injection

A vulnerability was found in wangzhixuan spring-shiro-training up to 94812c1fd8f7fe796c931f4984ff1aa0671ab562. It has been declared as critical. This vulnerability affects unknown code of the file /role/add. The manipulation leads to command injection. The attack can be initiated remotely. The...

CVE-2025-8752 wangzhixuan spring-shiro-training add command injection

A vulnerability was found in wangzhixuan spring-shiro-training up to 94812c1fd8f7fe796c931f4984ff1aa0671ab562. It has been declared as critical. This vulnerability affects unknown code of the file /role/add. The manipulation leads to command injection. The attack can be initiated remotely. The...

PT-2025-32435 · Wangzhixuan · Spring-Shiro-Training

Name of the Vulnerable Software and Affected Versions: wangzhixuan spring-shiro-training up to 94812c1fd8f7fe796c931f4984ff1aa0671ab562 Description: A critical issue exists in wangzhixuan spring-shiro-training. The vulnerability is due to command injection in the /role/add file. This allows for...

A Real-Time, Self-Tuning Moderator Framework for Adversarial Prompt Detection

Ensuring LLM alignment is critical to information security as AI models become increasingly widespread and integrated in society. Unfortunately, many defenses against adversarial attacks and jailbreaking on LLMs cannot adapt quickly to new attacks, degrade model responses to benign prompts, or...