Detecting Network Intrusion Using Binarized Neural Networks

Vreca, Jure; Ivanov, Iva; Papa, Gregor; Biasizzo, Anton

doi:10.1109/wf-iot51360.2021.9595961

Cited by 4 publications

(3 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We chose to implement FFTLayer and LMFELayer with floatingpoint arithmetic because of the existence of high-performance floating-point libraries and because modern high-performance hardware is optimized for floating-point operations. We provide a direct path from the FFTLayer and LMFELayer class definition to generated hardware via our Python/Chisel framework Chisel4ml [7]. The software architecture of Chisel4ml is shown in Figure 8.…”

Section: Python Integration-chisel4mlmentioning

confidence: 99%

See 1 more Smart Citation

Hardware–Software Co-Design of an Audio Feature Extraction Pipeline for Machine Learning Applications

Vreča,

Pilipović,

Biasizzo

2024

Electronics

View full text Add to dashboard Cite

Keyword spotting is an important part of modern speech recognition pipelines. Typical contemporary keyword-spotting systems are based on Mel-Frequency Cepstral Coefficient (MFCC) audio features, which are relatively complex to compute. Considering the always-on nature of many keyword-spotting systems, it is prudent to optimize this part of the detection pipeline. We explore the simplifications of the MFCC audio features and derive a simplified version that can be more easily used in embedded applications. Additionally, we implement a hardware generator that generates an appropriate hardware pipeline for the simplified audio feature extraction. Using Chisel4ml framework, we integrate hardware generators into Python-based Keras framework, which facilitates the training process of the machine learning models using our simplified audio features.

show abstract

Section: Python Integration-chisel4mlmentioning

confidence: 99%

“…Section 3 details the modules that compute the simplified MFCC features. In Section 4, we discuss Chisel4ml [7]. It is a Python/Chisel-based framework, which we developed, that is the basis of how we connect our hardware generators with the Python ecosystem.…”

Section: Introductionmentioning

confidence: 99%

Hardware–Software Co-Design of an Audio Feature Extraction Pipeline for Machine Learning Applications

Vreča,

Pilipović,

Biasizzo

2024

Electronics

View full text Add to dashboard Cite

show abstract

“…Network intrusion, or the unauthorized activate on a network, can lead to the compromise of not just data, but entire nodes and systems. A low power binary neural network (BNN) was developed for the detection of network intrusion with an 82% accuracy at a cost of 1.5 W [ 90 ]. This implementation analyzes network packets for fuzzers, backdoor attacks, denial of service attacks, reconnaissance attacks, and shellcode/worms.…”

Section: Mid-range Iot Solutionsmentioning

confidence: 99%

Review of State-of-the-Art FPGA Applications in IoT Networks

Magyari

Chen

2022

Sensors

View full text Add to dashboard Cite

Modern networks used for integrating custom Internet of Things (IoT) systems and devices have restrictions and requirements unique to their individual applications. These application specific demands require custom designed hardware to maximize throughput, security and data integrity whilst minimizing latency, power consumption, and form factor. Within this paper, we describe current, state-of-the-art works that utilize FPGAs for IoT network developments. We analyze two categories of works: those that prioritize reducing power consumption, and those that prioritize networking features. Further, we describe how future works can improve upon these designs and therefore improve the efficiency of resource-constrained IoT networks.

show abstract