arXiv subjects

Elisa Bertino

Publications and source records attributed to Elisa Bertino.

At least 37 records · Page 2Linked to original sources

arXiv2025

Automated Vulnerability Validation and Verification: A Large Language Model Approach

Software vulnerabilities remain a critical security challenge, providing entry points for attackers into enterprise networks. Despite advances in security practices, the lack of high-quality datasets capturing diverse exploit behavior limits effective vulnerability assessment and mitigation. This paper introduces an end-to-end multi-step pipeline leveraging generative AI, specifically large language models (LLMs), to address the challenges of orchestrating and reproducing attacks to known software vulnerabilities. Our approach extracts information from CVE disclosures in the National Vulnerability Database, augments it with external public knowledge (e.g., threat advisories, code snippets) using Retrieval-Augmented Generation (RAG), and automates the creation of containerized environments and exploit code for each vulnerability. The pipeline iteratively refines generated artifacts, validates attack success with test cases, and supports complex multi-container setups. Our methodology overcomes key obstacles, including noisy and incomplete vulnerability descriptions, by integrating LLMs and RAG to fill information gaps. We demonstrate the effectiveness of our pipeline across different vulnerability types, such as memory overflows, denial of service, and remote code execution, spanning diverse programming languages, libraries and years. In doing so, we uncover significant inconsistencies in CVE descriptions, emphasizing the need for more rigorous verification in the CVE disclosure process. Our approach is model-agnostic, working across multiple LLMs, and we open-source the artifacts to enable reproducibility and accelerate security research. To the best of our knowledge, this is the first system to systematically orchestrate and exploit known vulnerabilities in containerized environments by combining general-purpose LLM reasoning with CVE data and RAG-based context enrichment.

cs.CR↗

arXiv2025

TIMESAFE: Timing Interruption Monitoring and Security Assessment for Fronthaul Environments

5G and beyond cellular systems embrace the disaggregation of Radio Access Network (RAN) components, exemplified by the evolution of the fronthaul (FH) connection between cellular baseband and radio unit equipment. Crucially, synchronization over the FH is pivotal for reliable 5G services. In recent years, there has been a push to move these links to an Ethernet-based packet network topology, leveraging existing standards and ongoing research for Time-Sensitive Networking (TSN). However, TSN standards, such as Precision Time Protocol (PTP), focus on performance with little to no concern for security. This increases the exposure of the open FH to security risks. Attacks targeting synchronization mechanisms pose significant threats, potentially disrupting 5G networks and impairing connectivity. In this paper, we demonstrate the impact of successful spoofing and replay attacks against PTP synchronization. We show how a spoofing attack is able to cause a production-ready O-RAN and 5G-compliant private cellular base station to catastrophically fail within 2 seconds of the attack, necessitating manual intervention to restore full network operations. To counter this, we design a Machine Learning (ML)-based monitoring solution capable of detecting various malicious attacks with over 97.5% accuracy.

cs.NI↗

arXiv2025

LLMalMorph: On The Feasibility of Generating Variant Malware using Large-Language-Models

Large Language Models (LLMs) have transformed software development and automated code generation. Motivated by these advancements, this paper explores the feasibility of LLMs in modifying malware source code to generate variants. We introduce LLMalMorph, a semi-automated framework that leverages semantical and syntactical code comprehension by LLMs to generate new malware variants. LLMalMorph extracts function-level information from the malware source code and employs custom-engineered prompts coupled with strategically defined code transformations to guide the LLM in generating variants without resource-intensive fine-tuning. To evaluate LLMalMorph, we collected 10 diverse Windows malware samples of varying types, complexity and functionality and generated 618 variants. Our experiments demonstrate that LLMalMorph variants can effectively evade antivirus engines, achieving typical detection rate reductions of 10-15% across multiple complex samples. Furthermore, without explicitly targeting learning-based detectors, LLMalMorph attained attack success rates of up to 91% against a Machine Learning (ML) based malware detector. We also discuss the limitations of current LLM capabilities in generating malware variants from source code and assess where this emerging technology stands in the broader context of malware variant generation.

cs.CR↗

arXiv2025

Maximizing Information in Domain-Invariant Representation Improves Transfer Learning

We propose MaxDIRep, a domain adaptation method that improves the decomposition of data representations into domain-independent and domain-dependent components. Existing methods, such as Domain-Separation Networks (DSN), use a weak orthogonality constraint between these components, which can lead to label-relevant features being partially encoded in the domain-dependent representation (DDRep) rather than the domain-independent representation (DIRep). As a result, information crucial for target-domain classification may be missing from the DIRep. MaxDIRep addresses this issue by applying a Kullback-Leibler (KL) divergence constraint to minimize the information content of the DDRep, thereby encouraging the DIRep to retain features that are both domain-invariant and predictive of target labels. Through geometric analysis and an ablation study on synthetic datasets, we show why DSN's weaker constraint can lead to suboptimal adaptation. Experiments on standard image benchmarks and a network intrusion detection task demonstrate that MaxDIRep achieves strong performance, works with pretrained models, and generalizes to non-image classification tasks.

cs.CV↗

arXiv2025

Castle: Causal Cascade Updates in Relational Databases with Large Language Models

This work introduces Castle, the first framework for schema-only cascade update generation using large language models (LLMs). Despite recent advances in LLMs for Text2SQL code generation, existing approaches focus primarily on SELECT queries, neglecting the challenges of SQL update operations and their ripple effects. Traditional CASCADE UPDATE constraints are static and unsuitable for modern, denormalized databases, which demand dynamic, context-aware updates. Castle enables natural language instructions to trigger multi-column, causally consistent SQL UPDATE statements, without revealing table content to the model. By framing UPDATE SQL generation as a divide-and-conquer task with LLMs' reasoning capacity, Castle can determine not only which columns must be directly updated, but also how those updates propagate through the schema, causing cascading updates -- all via nested queries and substructures that ensure data confidentiality. We evaluate it on real-world causal update scenarios, demonstrating its ability to produce accurate SQL updates, and thereby highlighting the reasoning ability of LLMs in automated DBMS.

cs.DB↗

arXiv2025

VWAttacker: A Systematic Security Testing Framework for Voice over WiFi User Equipments

We present VWAttacker, the first systematic testing framework for analyzing the security of Voice over WiFi (VoWiFi) User Equipment (UE) implementations. VWAttacker includes a complete VoWiFi network testbed that communicates with Commercial-Off-The-Shelf (COTS) UEs based on a simple interface to test the behavior of diverse VoWiFi UE implementations; uses property-guided adversarial testing to uncover security issues in different UEs systematically. To reduce manual effort in extracting and testing properties, we introduce an LLM-based, semi-automatic, and scalable approach for property extraction and testcase (TC) generation. These TCs are systematically mutated by two domain-specific transformations. Furthermore, we introduce two deterministic oracles to detect property violations automatically. Coupled with these techniques, VWAttacker extracts 63 properties from 11 specifications, evaluates 1,116 testcases, and detects 13 issues in 21 UEs. The issues range from enforcing a DH shared secret to 0 to supporting weak algorithms. These issues result in attacks that expose the victim UE's identity or establish weak channels, thus severely hampering the security of cellular networks. We responsibly disclose the findings to all the related vendors. At the time of writing, one of the vulnerabilities has been acknowledged by MediaTek with high severity.

cs.CR↗

arXiv2025

Gotta Detect 'Em All: Fake Base Station and Multi-Step Attack Detection in Cellular Networks

Fake base stations (FBSes) pose a significant security threat by impersonating legitimate base stations (BSes). Though efforts have been made to defeat this threat, up to this day, the presence of FBSes and the multi-step attacks (MSAs) stemming from them can lead to unauthorized surveillance, interception of sensitive information, and disruption of network services. Therefore, detecting these malicious entities is crucial to ensure the security and reliability of cellular networks. Traditional detection methods often rely on additional hardware, rules, signal scanning, changing protocol specifications, or cryptographic mechanisms that have limitations and incur huge infrastructure costs. In this paper, we develop FBSDetector-an effective and efficient detection solution that can reliably detect FBSes and MSAs from layer-3 network traces using machine learning (ML) at the user equipment (UE) side. To develop FBSDetector, we create FBSAD and MSAD, the first-ever high-quality and large-scale datasets incorporating instances of FBSes and 21 MSAs. These datasets capture the network traces in different real-world cellular network scenarios (including mobility and different attacker capabilities) incorporating legitimate BSes and FBSes. Our novel ML framework, specifically designed to detect FBSes in a multi-level approach for packet classification using stateful LSTM with attention and trace level classification and MSAs using graph learning, can effectively detect FBSes with an accuracy of 96% and a false positive rate of 2.96%, and recognize MSAs with an accuracy of 86% and a false positive rate of 3.28%. We deploy FBSDetector as a real-world solution to protect end-users through a mobile app and validate it in real-world environments. Compared to the existing heuristic-based solutions that fail to detect FBSes, FBSDetector can detect FBSes in the wild in real-time.

cs.CR↗

arXiv2025

LLM Agents Should Employ Security Principles

Large Language Model (LLM) agents show considerable promise for automating complex tasks using contextual reasoning; however, interactions involving multiple agents and the system's susceptibility to prompt injection and other forms of context manipulation introduce new vulnerabilities related to privacy leakage and system exploitation. This position paper argues that the well-established design principles in information security, which are commonly referred to as security principles, should be employed when deploying LLM agents at scale. Design principles such as defense-in-depth, least privilege, complete mediation, and psychological acceptability have helped guide the design of mechanisms for securing information systems over the last five decades, and we argue that their explicit and conscientious adoption will help secure agentic systems. To illustrate this approach, we introduce AgentSandbox, a conceptual framework embedding these security principles to provide safeguards throughout an agent's life-cycle. We evaluate with state-of-the-art LLMs along three dimensions: benign utility, attack utility, and attack success rate. AgentSandbox maintains high utility for its intended functions under both benign and adversarial evaluations while substantially mitigating privacy risks. By embedding secure design principles as foundational elements within emerging LLM agent protocols, we aim to promote trustworthy agent ecosystems aligned with user privacy expectations and evolving regulatory requirements.

cs.CR↗

arXiv2025

Standing Firm in 5G: A Single-Round, Dropout-Resilient Secure Aggregation for Federated Learning

Federated learning (FL) is well-suited to 5G networks, where many mobile devices generate sensitive edge data. Secure aggregation protocols enhance privacy in FL by ensuring that individual user updates reveal no information about the underlying client data. However, the dynamic and large-scale nature of 5G-marked by high mobility and frequent dropouts-poses significant challenges to the effective adoption of these protocols. Existing protocols often require multi-round communication or rely on fixed infrastructure, limiting their practicality. We propose a lightweight, single-round secure aggregation protocol designed for 5G environments. By leveraging base stations for assisted computation and incorporating precomputation, key-homomorphic pseudorandom functions, and t-out-of-k secret sharing, our protocol ensures efficiency, robustness, and privacy. Experiments show strong security guarantees and significant gains in communication and computation efficiency, making the approach well-suited for real-world 5G FL deployments.

cs.CR↗

arXiv2025

Efficient Full-Stack Private Federated Deep Learning with Post-Quantum Security

Federated learning (FL) enables collaborative model training while preserving user data privacy by keeping data local. Despite these advantages, FL remains vulnerable to privacy attacks on user updates and model parameters during training and deployment. Secure aggregation protocols have been proposed to protect user updates by encrypting them, but these methods often incur high computational costs and are not resistant to quantum computers. Additionally, differential privacy (DP) has been used to mitigate privacy leakages, but existing methods focus on secure aggregation or DP, neglecting their potential synergies. To address these gaps, we introduce Beskar, a novel framework that provides post-quantum secure aggregation, optimizes computational overhead for FL settings, and defines a comprehensive threat model that accounts for a wide spectrum of adversaries. We also integrate DP into different stages of FL training to enhance privacy protection in diverse scenarios. Our framework provides a detailed analysis of the trade-offs between security, performance, and model accuracy, representing the first thorough examination of secure aggregation protocols combined with various DP approaches for post-quantum secure FL. Beskar aims to address the pressing privacy and security issues FL while ensuring quantum-safety and robust performance.

cs.CR↗

arXiv2025

SoK: Leveraging Transformers for Malware Analysis

The introduction of transformers has been an important breakthrough for AI research and application as transformers are the foundation behind Generative AI. A promising application domain for transformers is cybersecurity, in particular the malware domain analysis. The reason is the flexibility of the transformer models in handling long sequential features and understanding contextual relationships. However, as the use of transformers for malware analysis is still in the infancy stage, it is critical to evaluate, systematize, and contextualize existing literature to foster future research. This Systematization of Knowledge (SoK) paper aims to provide a comprehensive analysis of transformer-based approaches designed for malware analysis. Based on our systematic analysis of existing knowledge, we structure and propose taxonomies based on: (a) how different transformers are adapted, organized, and modified across various use cases; and (b) how diverse feature types and their representation capabilities are reflected. We also provide an inventory of datasets used to explore multiple research avenues in the use of transformers for malware analysis and discuss open challenges with future research directions. We believe that this SoK paper will assist the research community in gaining detailed insights from existing work and will serve as a foundational resource for implementing novel research using transformers for malware analysis.

cs.CR↗

arXiv2025

How Feasible is Augmenting Fake Nodes with Learnable Features as a Counter-strategy against Link Stealing Attacks?

Graph Neural Networks (GNNs) are widely used and deployed for graph-based prediction tasks. However, as good as GNNs are for learning graph data, they also come with the risk of privacy leakage. For instance, an attacker can run carefully crafted queries on the GNNs and, from the responses, can infer the existence of an edge between a pair of nodes. This attack, dubbed as a "link-stealing" attack, can jeopardize the user's privacy by leaking potentially sensitive information. To protect against this attack, we propose an approach called "$(N)$ode $(A)$ugmentation for $(R)$estricting $(G)$raphs from $(I)$nsinuating their $(S)$tructure" ($NARGIS$) and study its feasibility. $NARGIS$ is focused on reshaping the graph embedding space so that the posterior from the GNN model will still provide utility for the prediction task but will introduce ambiguity for the link-stealing attackers. To this end, $NARGIS$ applies spectral clustering on the given graph to facilitate it being augmented with new nodes -- that have learned features instead of fixed ones. It utilizes tri-level optimization for learning parameters for the GNN model, surrogate attacker model, and our defense model (i.e. learnable node features). We extensively evaluate $NARGIS$ on three benchmark citation datasets over eight knowledge availability settings for the attackers. We also evaluate the model fidelity and defense performance on influence-based link inference attacks. Through our studies, we have figured out the best feature of $NARGIS$ -- its superior fidelity-privacy performance trade-off in a significant number of cases. We also have discovered in which cases the model needs to be improved, and proposed ways to integrate different schemes to make the model more robust against link stealing attacks.

cs.LG↗

arXiv2024

Revisiting Concept Drift in Windows Malware Detection: Adaptation to Real Drifted Malware with Minimal Samples

In applying deep learning for malware classification, it is crucial to account for the prevalence of malware evolution, which can cause trained classifiers to fail on drifted malware. Existing solutions to address concept drift use active learning. They select new samples for analysts to label and then retrain the classifier with the new labels. Our key finding is that the current retraining techniques do not achieve optimal results. These techniques overlook that updating the model with scarce drifted samples requires learning features that remain consistent across pre-drift and post-drift data. The model should thus be able to disregard specific features that, while beneficial for the classification of pre-drift data, are absent in post-drift data, thereby preventing prediction degradation. In this paper, we propose a new technique for detecting and classifying drifted malware that learns drift-invariant features in malware control flow graphs by leveraging graph neural networks with adversarial domain adaptation. We compare it with existing model retraining methods in active learning-based malware detection systems and other domain adaptation techniques from the vision domain. Our approach significantly improves drifted malware detection on publicly available benchmarks and real-world malware databases reported daily by security companies in 2024. We also tested our approach in predicting multiple malware families drifted over time. A thorough evaluation shows that our approach outperforms the state-of-the-art approaches.

cs.CR↗

arXiv2024

ZT-SDN: An ML-powered Zero-Trust Architecture for Software-Defined Networks

Zero Trust (ZT) is a security paradigm aiming to curtail an attacker's lateral movements within a network by implementing least-privilege and per-request access control policies. However, its widespread adoption is hindered by the difficulty of generating proper rules due to the lack of detailed knowledge of communication requirements and the characteristic behaviors of communicating entities under benign conditions. Consequently, manual rule generation becomes cumbersome and error-prone. To address these problems, we propose ZT-SDN, an automated framework for learning and enforcing network access control in Software-Defined Networks. ZT-SDN collects data from the underlying network and models the network "transactions" performed by communicating entities as graphs. The nodes represent entities, while the directed edges represent transactions identified by different protocol stacks observed. It uses novel unsupervised learning approaches to extract transaction patterns directly from the network data, such as the allowed protocol stacks and port numbers and data transmission behavior. Finally, ZT-SDN uses an innovative approach to generate correct access control rules and infer strong associations between them, allowing proactive rule deployment in forwarding devices. We show the framework's efficacy in detecting abnormal network accesses and abuses of permitted flows in changing network conditions with real network datasets. Additionally, we showcase ZT-SDN's scalability and the network's performance when applied in an SDN environment.

cs.CR↗

arXiv2024

Adversarial Domain Adaptation for Metal Cutting Sound Detection: Leveraging Abundant Lab Data for Scarce Industry Data

Cutting state monitoring in the milling process is crucial for improving manufacturing efficiency and tool life. Cutting sound detection using machine learning (ML) models, inspired by experienced machinists, can be employed as a cost-effective and non-intrusive monitoring method in a complex manufacturing environment. However, labeling industry data for training is costly and time-consuming. Moreover, industry data is often scarce. In this study, we propose a novel adversarial domain adaptation (DA) approach to leverage abundant lab data to learn from scarce industry data, both labeled, for training a cutting-sound detection model. Rather than adapting the features from separate domains directly, we project them first into two separate latent spaces that jointly work as the feature space for learning domain-independent representations. We also analyze two different mechanisms for adversarial learning where the discriminator works as an adversary and a critic in separate settings, enabling our model to learn expressive domain-invariant and domain-ingrained features, respectively. We collected cutting sound data from multiple sensors in different locations, prepared datasets from lab and industry domain, and evaluated our learning models on them. Experiments showed that our models outperformed the multi-layer perceptron based vanilla domain adaptation models in labeling tasks on the curated datasets, achieving near 92%, 82% and 85% accuracy respectively for three different sensors installed in industry settings.

cs.LG↗

arXiv2024

Uncovering Attacks and Defenses in Secure Aggregation for Federated Deep Learning

Federated learning enables the collaborative learning of a global model on diverse data, preserving data locality and eliminating the need to transfer user data to a central server. However, data privacy remains vulnerable, as attacks can target user training data by exploiting the updates sent by users during each learning iteration. Secure aggregation protocols are designed to mask/encrypt user updates and enable a central server to aggregate the masked information. MicroSecAgg (PoPETS 2024) proposes a single server secure aggregation protocol that aims to mitigate the high communication complexity of the existing approaches by enabling a one-time setup of the secret to be re-used in multiple training iterations. In this paper, we identify a security flaw in the MicroSecAgg that undermines its privacy guarantees. We detail the security flaw and our attack, demonstrating how an adversary can exploit predictable masking values to compromise user privacy. Our findings highlight the critical need for enhanced security measures in secure aggregation protocols, particularly the implementation of dynamic and unpredictable masking strategies. We propose potential countermeasures to mitigate these vulnerabilities and ensure robust privacy protection in the secure aggregation frameworks.

cs.CR↗

arXiv2024

CellularLint: A Systematic Approach to Identify Inconsistent Behavior in Cellular Network Specifications

In recent years, there has been a growing focus on scrutinizing the security of cellular networks, often attributing security vulnerabilities to issues in the underlying protocol design descriptions. These protocol design specifications, typically extensive documents that are thousands of pages long, can harbor inaccuracies, underspecifications, implicit assumptions, and internal inconsistencies. In light of the evolving landscape, we introduce CellularLint--a semi-automatic framework for inconsistency detection within the standards of 4G and 5G, capitalizing on a suite of natural language processing techniques. Our proposed method uses a revamped few-shot learning mechanism on domain-adapted large language models. Pre-trained on a vast corpus of cellular network protocols, this method enables CellularLint to simultaneously detect inconsistencies at various levels of semantics and practical use cases. In doing so, CellularLint significantly advances the automated analysis of protocol specifications in a scalable fashion. In our investigation, we focused on the Non-Access Stratum (NAS) and the security specifications of 4G and 5G networks, ultimately uncovering 157 inconsistencies with 82.67% accuracy. After verification of these inconsistencies on open-source implementations and 17 commercial devices, we confirm that they indeed have a substantial impact on design decisions, potentially leading to concerns related to privacy, integrity, availability, and interoperability.

cs.CR↗

arXiv2024

FlowMur: A Stealthy and Practical Audio Backdoor Attack with Limited Knowledge

Speech recognition systems driven by DNNs have revolutionized human-computer interaction through voice interfaces, which significantly facilitate our daily lives. However, the growing popularity of these systems also raises special concerns on their security, particularly regarding backdoor attacks. A backdoor attack inserts one or more hidden backdoors into a DNN model during its training process, such that it does not affect the model's performance on benign inputs, but forces the model to produce an adversary-desired output if a specific trigger is present in the model input. Despite the initial success of current audio backdoor attacks, they suffer from the following limitations: (i) Most of them require sufficient knowledge, which limits their widespread adoption. (ii) They are not stealthy enough, thus easy to be detected by humans. (iii) Most of them cannot attack live speech, reducing their practicality. To address these problems, in this paper, we propose FlowMur, a stealthy and practical audio backdoor attack that can be launched with limited knowledge. FlowMur constructs an auxiliary dataset and a surrogate model to augment adversary knowledge. To achieve dynamicity, it formulates trigger generation as an optimization problem and optimizes the trigger over different attachment positions. To enhance stealthiness, we propose an adaptive data poisoning method according to Signal-to-Noise Ratio (SNR). Furthermore, ambient noise is incorporated into the process of trigger generation and data poisoning to make FlowMur robust to ambient noise and improve its practicality. Extensive experiments conducted on two datasets demonstrate that FlowMur achieves high attack performance in both digital and physical settings while remaining resilient to state-of-the-art defenses. In particular, a human study confirms that triggers generated by FlowMur are not easily detected by participants.

cs.CR↗