Fraud Detection in Voice-Based Identity Authentication Applications and Services

Safavi, Saeid; Gan, Hock; Mporas, Iosif; Sotudeh, R.

doi:10.1109/icdmw.2016.0155

Cited by 16 publications

(6 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this work, the use of the Voxforge Spanish dataset raises some privacy concerns, since the donors accepted that their voices are part of a database destined to create language technologies however while maintaining anonymity. On the other hand, due to the type of information provided by each donor (mainly a personalised username), it could be easy to identify some of them and performing actions against them like identity fraud in certain speech systems [37]. Fortunately, some of this concerns were contemplated by the developers of the Voxforge datasets which allowed and option to opt-out, however it could hardly be enforced for all distributed copies [38].…”

Section: Ethical Discussionmentioning

confidence: 99%

Triplet loss based embeddings for forensic speaker identification in Spanish

Maqueda,

Alvarez-Jimenez,

Mena

et al. 2021

Preprint

View full text Add to dashboard Cite

With the advent of digital technology, it is more common that committed crimes or legal disputes involve some form of speech recording where the identity of a speaker is questioned [1]. In face of this situation, the field of forensic speaker identification has been looking to shed light on the problem by quantifying how much a speech recording belongs to a particular person in relation to a population. In this work, we explore the use of speech embeddings obtained by training a CNN using the triplet loss. In particular, we focus on the Spanish language which has not been extensively studies. We propose extracting the embeddings from speech spectrograms samples, then explore several configurations of such spectrograms, and finally, quantify the embeddings quality. We also show some limitations of our data setting which is predominantly composed by male speakers. At the end, we propose two approaches to calculate the Likelihood Radio given out speech embeddings and we show that triplet loss is a good alternative to create speech embeddings for forensic speaker identification.Keywords Triplet Loss • Speaker Identification • Forensic • Spanish Recently, Triplet loss has been proposed for speech tasks, for instance: the speaker verification task [6], for speaker turn [7], speech emotion classification [8], among other tasks. However, its applicability to forensic speaker identification has not been explored, particularly for the Spanish language. Forensic Speaker Identification (FSI) focuses on gathering and quantifying the evidence that will be presented in a court. FSI addresses the question if a specific recording registers or not, speech produced by a specific person [9,10]. The more basic scenario in FSI consists of two sample speech recordings, a reference sample and a questioned sample. For the reference sample, we always know the identity of

show abstract

Section: Ethical Discussionmentioning

confidence: 99%

Triplet loss based embeddings for forensic speaker identification in Spanish

Maqueda,

Alvarez-Jimenez,

Mena

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…For classification purposes we have used multiple functions provided by the Weka toolkit 2 . We have tried logistic classifier, decision tree, and multilayer perceptron (MLP) [14], [15], [16], [17].…”

Section: B Modeling Methodsmentioning

confidence: 99%

Predicting the perceived level of reverberation using machine learning

Safavi

Pearce

Wang

et al. 2018

2018 52nd Asilomar Conference on Signals, Systems, and Computers

Self Cite

View full text Add to dashboard Cite

Perceptual measures are usually considered more reliable than instrumental measures for evaluating the perceived level of reverberation. However, such measures are time consuming and expensive, and, due to variations in stimuli or assessors, the resulting data is not always statistically significant. Therefore, an (objective) measure of the perceived level of reverberation becomes desirable. In this paper, we develop a new method to predict the level of reverberation from audio signals by relating the perceptual listening test results with those obtained from a machine learned model. More specifically, we compare the use of a multiple stimuli test for within and between class architectures to evaluate the perceived level of reverberation. An expert set of 16 human listeners rated the perceived level of reverberation for a same set of files from different audio source types. We then train a machine learning model using the training data gathered for the same set of files and a variety of reverberation related features extracted from the data such as reverberation time, and direct to reverberation ratio. The results suggest that the machine learned model offers an accurate prediction of the perceptual scores. Index Terms-Reverberation time, Human subject test, Machine learning, MLP.

show abstract

“…About countermeasure and machine learning, the machine learning algorithm of article [12] could be used to find out the best hamming distance redistribution mapping by using neural dynamic programming. Safavi et al [13] researched the fraud detection in voice-based identity authentication applications and services. In article [14], based on maximum entropy clustering, two specialized criteriainter-view collaborative learning and intra-view weighted attributes are first devised as the bases.…”

Section: Introductionmentioning

confidence: 99%

Botnet Vulnerability Intelligence Clustering Classification Mining and Countermeasure Algorithm Based on Machine Learning

2019

View full text Add to dashboard Cite

Botnet detection, vulnerability mining and confrontation bring many challenges to network security, and become the main dangerous source threatening economic development and causing significant economic losses. To solve these problems, we combine machine learning with classification mining, and propose a botnet vulnerability mining and countermeasure algorithm and its system. Firstly, a botnet model based on machine learning and its construction method are designed according to the characteristics of Botnet such as group network attack, computer group controlled and distributed denial of service attack. Then, in order to achieve the global optimal evaluation and screening of Botnet detection by network vulnerability mining system, a network vulnerability combination classification mining algorithm and machine learning instruction set are designed. Then, based on the analysis of the vulnerability form and the vulnerability diffusion mode of botnet, a vulnerability mining and confrontation algorithm for botnet is proposed. Finally, through simulation experiments, the advantages of the proposed algorithm are proved in vulnerability identification rate, vulnerability clearance performance and botnet connectivity. For example, the average vulnerability detection rate is more than 87%, the vulnerability repair rate is higher than 86%, and the pass rate is still higher than 89%.

show abstract

Fraud Detection in Voice-Based Identity Authentication Applications and Services

Cited by 16 publications

References 11 publications

Triplet loss based embeddings for forensic speaker identification in Spanish

Triplet loss based embeddings for forensic speaker identification in Spanish

Predicting the perceived level of reverberation using machine learning

Botnet Vulnerability Intelligence Clustering Classification Mining and Countermeasure Algorithm Based on Machine Learning

Contact Info

Product

Resources

About