30 research outputs found
QuakeSense, a LoRa-compliant Earthquake Monitoring Open System
No full textDetecting disruptive events, such as earthquakes, using environmental monitoring systems is a particularly promising, but rather challenging, opportunity. The Internet of Things (IoT) can play a significant role in characterizing and predicting seismic events. The present contribution introduces QuakeSense, an open-source earthquake and weather monitoring system. The implemented IoT system is configured as a Long Range (LoRa)based star topology with a fully energy-autonomous sensor node. The system leverages some of the most useful features of two emerging IoT technologies, e.g., LoRa and Message Queue Telemetry Transport (MQTT), and enables the near real-time monitoring of seismic events through a web-based interface. An experimental campaign has been carried out to verify the current consumption and, therefore, the battery lifetime of the sensor node. Moreover, LoRa parameters have been extensively tested as to evaluate performances in several configurations. The obtained results in terms of latency and Packet Delivery Ratio (PDR) demonstrated the reliability of the proposal
SourceBroken: A large-scale analysis on the (un)reliability of SourceRank in the PyPI ecosystem
No full textRaze to the ground: query-efficient adversarial HTML attacks on machine-learning phishing webpage detectors
No full textMachine-learning phishing webpage detectors (ML-PWD) have been shown to suffer from adversarial manipulations of the HTML code of the input webpage. Nevertheless, the attacks recently proposed have demonstrated limited effectiveness due to their lack of optimizing the usage of the adopted manipulations, and they focus solely on specific elements of the HTML code. In this work, we overcome these limitations by first designing a novel set of fine-grained manipulations which allow to modify the HTML code of the input phishing webpage without compromising its maliciousness and visual appearance, i.e., the manipulations are functionality- and rendering-preserving by design. We then select which manipulations should be applied to bypass the target detector by a query-efficient black-box optimization algorithm. Our experiments show that our attacks are able to raze to the ground the performance of current state-of-the-art ML-PWD using just 30 queries, thus overcoming the weaker attacks developed in previous work, and enabling a much fairer robustness evaluation of ML-PWD
Raze to the ground: Query-efficient adversarial HTML attacks on machine-learning phishing webpage detectors
No full textAdversarial ModSecurity: Countering Adversarial SQL Injections with Robust Machine Learning
No full text
ModSec-Learn: Boosting ModSecurity with Machine Learning
No full textModSecurity is widely recognized as the standard open-source Web Application Firewall (WAF), maintained by the OWASP Foundation. It detects malicious requests by matching them against the Core Rule Set (CRS), identifying well-known attack patterns. Each rule is manually assigned a weight based on the severity of the corresponding
attack, and a request is blocked if the sum of the weights of matched rules exceeds a given threshold. However, we argue that this strategy is largely ineffective against web attacks, as detection is only based on heuristics
and not customized on the application to protect. In this work, we overcome this issue by proposing a machine-learning model that uses the CRS rules as input features. Through training, ModSec-Learn is able to tune
the contribution of each CRS rule to predictions, thus adapting the severity level to the web applications to protect. Our experiments show that ModSec-Learn achieves a significantly better trade-off between detection
and false positive rates. Finally, we analyze how sparse regularization can reduce the number of rules that are relevant at inference time, by discarding more than 30% of the CRS rules. We release our open-source
code and the dataset at https://github.com/pralab/modsec-learn and https://github.com/pralab/http-traffic-dataset, respectively
Neural network splitter : optimal decomposition of a neural network and its distribution on multiple microcontrollers
No full textLAUREA MAGISTRALEL’implementazione di una rete neurale (NN) su dispositivi a bassa potenza e con risorse limitate rappresenta un problema critico nello sviluppo di sistemi IoT intelligenti ed autonomi a causa degli aggressivi vincoli computazionali e di memoria. Per questo motivo, le soluzioni di Machine Learning (ML) rivolte a piccoli dispositivi devono essere progettate tenendo presente i vincoli legati alla memoria e alla capacità di elaborazione che caratterizzano tali dispositivi. In questa tesi, introduciamo una nuova metodologia di progettazione basata su un approccio distribuito, il quale ha come obiettivo partizionare automaticamente l’esecuzione di una NN su più dispositivi eterogenei molto limitati. Tale metodologia è formalizzata come un problema di ottimizzazione in cui o la latenza di inferenza è minimizzata oppure il throughput è massimizzato, tenendo in considerazione le capacità di memoria e di calcolo dei dispositivi. La metodologia è valutata su diverse architetture di reti neurali e su microcontrollori (MCUs) utilizzando tre algoritmi, vale a dire il Full Search (FS), il Dichotomich Search (DS) ed il Branch-and-Bound (B&B). I risultati ottenuti hanno mostrato che il B&B ha performato in modo di gran lunga migliore rispetto agli altri, in quanto è stato sempre in grado di trovare la soluzione ottima nel minor numero di iterazioni. Con questo lavoro, cerchiamo di proporre nuove soluzioni di ML che offrano una bassa decision-latency, autonomia ed un’elevata efficienza energetica.The deployment of Neural Network (NN) models on low-power and resource-constrained devices represents a critical bottleneck in the development of intelligent and autonomous Internet of Things (IoT) systems due to the aggressive computational and memory constraints. For this reason, Machine Learning (ML) solutions addressing tiny devices must be designed having in mind constraints on memory and processing capability characterizing such devices. In this thesis, we introduce a novel design methodology based on a distributed approach, which aims at automatically partitioning the execution of a NN over multiple heterogeneous tiny devices. Such a methodology is formalized as an optimization problem where either the inference latency is minimized or the throughput is maximized, within the devices’ memory and computing capabilities. The methodology is evaluated over different NN architectures and microcontrollers (MCUs) by using three algorithms, namely Full Search (FS), Dichotomic Search (DS), and Branch-and-Bound (B&B). The obtained results showed that the B&B outperformed the others as it was always able to find the optimal solution in the lowest number of computing steps. With this work, we aim at enabling novel ML solutions which offer low decisionlatency, autonomy, and high energy efficiency
End-of-life wood quality of mooring poles
No full textStructural EngineeringCivil Engineering and Geoscience
