Rudolf Blažek

Robust Hierarchical Linear Model Comparison for End-of-Utterance Detection under Noisy Environments

A simple and efficient algorithm for robust end of-utterance detection of speech signal in noisy ... more A simple and efficient algorithm for robust end of-utterance detection of speech signal in noisy environments is proposed in the paper. To detect speech-block end-points, we use entropy sequence of the input speech signal, and hierarchically compare the fit of two weighted linear models. The first model, M1, is very simple, it corresponds to a constant average entropy level for the speech signal in the entire window. The second model, M2, corresponds to a step like entropy change from one constant level to another, with a gradual transition between the levels. Model M2 is in fact a piecewise linear regression model with two horizontal lines connected by a third transitional line. We treat M1 as a linear model only to be able to describe it as a sub model of M2 and use methodology based on statistical sub model testing. The regression models are constructed so that their fit will differ the most near the speech-block end-points. The models are fitted in a interval of the entropy sequence, that correspond to several consecutive frames in a sliding time window. The interval with the greatest difference of the model fit is used to estimate the location of the speech-block boundaries. The performance of the proposed algorithm is compared with a conventional approach by Nokia. The experimental results show that the proposed method approach can significantly outperform the conventional-based approach in different SNR swith different noise types. Furthermore, we observe that the proposed algorithm has the potential to improve robustness not only on diverse noise-type conditions, but also on the extremely noise contaminated speech below 0 dB SNR.

<title>Interactive banks of Bayesian matched filters</title>

Proceedings of SPIE, Jul 13, 2000

<title>Effective adaptive spatial-temporal technique for clutter rejection in IRST</title>

Proceedings of SPIE, Jul 13, 2000

Predicting the Life Expectancy of Railway Fail-Safe Signaling Systems Using Dynamic Models with Censoring

In the presented work we predict the life expectancy of multi-part railway fail-safe signaling sy... more In the presented work we predict the life expectancy of multi-part railway fail-safe signaling systems. The monitored electronic track circuits detect train locations and movement in real time, and issue alerts and warnings to prevent collisions. Based on 10 years of failure reports from the manufacturer of systems used by Railway Infrastructure Administration in the Czech Republic, we establish estimates of time-to-failure distributions of their components. We modify and apply survival models for censored data with various parameters for which we propose and compare new estimators. Both left and right time-based censoring of the data is considered. This approach allows us to include in the analysis components that were in operation before the study started, as well as components that were functional after the end of the study. Special attention is paid to the correct treatment of missing and incomplete data in the analyzed reports. We compare models with constant and variable failure rates. Hypotheses testing methodology is used to select a model with the best fit for the analyzed data.

Detection of intrusions in information systems by sequential change-point methods

Statistical Methodology, Jul 1, 2006

Sequential multi-chart detection procedures for detecting changes in multichannel sensor systems ... more Sequential multi-chart detection procedures for detecting changes in multichannel sensor systems are developed. In the case of complete information on pre-change and post-change distributions, the detection algorithm represents a likelihood ratio-based multichannel generalization of Page's cumulative sum (CUSUM) test that is applied to general stochastic models that may include correlated and nonstationary observations. There are many potential application areas where it is necessary to consider multichannel generalizations and general statistical models. In this paper our main motivation for doing so is network security: rapid anomaly detection for an early detection of attacks in computer networks that lead to changes in network traffic. Moreover, this kind of application encourages the development of a nonparametric multichannel detection test that does not use exact pre-change (legitimate) and post-change (attack) traffic models. The proposed nonparametric method can be effectively applied to detect a wide variety of attacks such as denial-of-service attacks, worm-based attacks, port-scanning, and man-in-the-middle attacks. In addition, we propose a multichannel CUSUM procedure that is based on binary quantized data; this procedure turns out to be more efficient than the previous two algorithms in certain scenarios. All proposed detection algorithms are based on the change-point detection theory. They utilize the thresholding of test statistics to achieve a fixed rate of false alarms, while allowing changes in statistical models to be detected "as soon as possible". Theoretical frameworks for the performance analysis of detection procedures, as well

Download

Change-point detection method on 100 Gb/s ethernet interface

Adaptive Hardware Acceleration of Intrusion Detection • Current monitoring systems usually consis... more

Download

A novel approach to detection of intrusions in computer networks via adaptive sequential and batch-sequential change-point detection methods

IEEE Transactions on Signal Processing, Sep 1, 2006

Large-scale computer network attacks in their final stages can readily be identified by observing... more Large-scale computer network attacks in their final stages can readily be identified by observing very abrupt changes in the network traffic. In the early stage of an attack, however, these changes are hard to detect and difficult to distinguish from usual traffic fluctuations. Rapid response, a minimal false-alarm rate, and the capability to detect a wide spectrum of attacks are the crucial features of intrusion detection systems. In this paper, we develop efficient adaptive sequential and batch-sequential methods for an early detection of attacks that lead to changes in network traffic, such as denial-of-service attacks, worm-based attacks, portscanning, and man-in-the-middle attacks. These methods employ a statistical analysis of data from multiple layers of the network protocol to detect very subtle traffic changes. The algorithms are based on change-point detection theory and utilize a thresholding of test statistics to achieve a fixed rate of false alarms while allowing us to detect changes in statistical models as soon as possible. There are three attractive features of the proposed approach. First, the developed algorithms are self-learning, which enables them to adapt to various network loads and usage patterns. Secondly, they allow for the detection of attacks with a small average delay for a given false-alarm rate. Thirdly, they are computationally simple and thus can be implemented online. Theoretical frameworks for detection procedures are presented. We also give the results of the experimental study with the use of a network simulator testbed as well as real-life testing for TCP SYN flooding attacks.

Download

Predicting the Life Expectancy of Railway Fail-Safe Signaling Systems Using Dynamic Models with Censoring

2017 IEEE International Conference on Software Quality, Reliability and Security (QRS), 2017

In the presented work we predict the life expectancy of multi-part railway fail-safe signaling sy... more In the presented work we predict the life expectancy of multi-part railway fail-safe signaling systems. The monitored electronic track circuits detect train locations and movement in real time, and issue alerts and warnings to prevent collisions. Based on 10 years of failure reports from the manufacturer of systems used by Railway Infrastructure Administration in the Czech Republic, we establish estimates of time-to-failure distributions of their components. We modify and apply survival models for censored data with various parameters for which we propose and compare new estimators. Both left and right time-based censoring of the data is considered. This approach allows us to include in the analysis components that were in operation before the study started, as well as components that were functional after the end of the study. Special attention is paid to the correct treatment of missing and incomplete data in the analyzed reports. We compare models with constant and variable failu...

Security Analytics in the Big Data Era

2017 IEEE International Conference on Software Quality, Reliability and Security Companion (QRS-C), 2017

This paper discusses the reality of the state-of-theartof existing information security systems t... more This paper discusses the reality of the state-of-theartof existing information security systems that often providesenseless functions based on “buzz-words”. It points out real-liferequirements that these systems tend to ignore. It proposes thatdynamically changing understandable use cases should becreated in collaboration with corporate management to addressobjectives that are essential for the businesses. Big data analyticmethodologies should be utilized to assist the design of efficientimplementations of these use cases. Big data approaches shouldalso be used for heuristic detection of unknown attacks andanomalies, for data enrichment, and for post-hoc forensicanalysis. The main goal for information security systems that areusable in real-life should be that corporate decision makers areonly provided trustworthy and actionable insights that arerelevant to the business and provided services.

<title>Effective adaptive spatial-temporal technique for clutter rejection in IRST</title>

Signal and Data Processing of Small Targets 2000, 2000

Change-Point Detection Method in 100 Gb/s Networks

Adaptive Spatial-Temporal Method for Clutter Rejection and Switching Multiple Model Based Target Tracking Algorithms

Challenges in Spatial-Temporal Method for Electronic Scene Stabilization and Clutter Removal, and Nonlinear Filtering for Track Before Detect

Response of Authors to Discussions on `Detection of Intrusions in Information Systems by Sequential Change-point Methods

Robust Hierarchical Linear Model Comparison for End-of-Utterance Detection under Noisy Environments

2012 International Symposium on Biometrics and Security Technologies, 2012

A simple and efficient algorithm for robust end of-utterance detection of speech signal in noisy ... more A simple and efficient algorithm for robust end of-utterance detection of speech signal in noisy environments is proposed in the paper. To detect speech-block end-points, we use entropy sequence of the input speech signal, and hierarchically compare the fit of two weighted linear models. The first model, M1, is very simple, it corresponds to a constant average entropy level for the speech signal in the entire window. The second model, M2, corresponds to a step like entropy change from one constant level to another, with a gradual transition between the levels. Model M2 is in fact a piecewise linear regression model with two horizontal lines connected by a third transitional line. We treat M1 as a linear model only to be able to describe it as a sub model of M2 and use methodology based on statistical sub model testing. The regression models are constructed so that their fit will differ the most near the speech-block end-points. The models are fitted in a interval of the entropy sequence, that correspond to several consecutive frames in a sliding time window. The interval with the greatest difference of the model fit is used to estimate the location of the speech-block boundaries. The performance of the proposed algorithm is compared with a conventional approach by Nokia. The experimental results show that the proposed method approach can significantly outperform the conventional-based approach in different SNR swith different noise types. Furthermore, we observe that the proposed algorithm has the potential to improve robustness not only on diverse noise-type conditions, but also on the extremely noise contaminated speech below 0 dB SNR.

Interactive banks of Bayesian matched filters

SPIE Proceedings, 2000

ABSTRACT There exist a number of powerful methods for detecting small low observable targets with... more ABSTRACT There exist a number of powerful methods for detecting small low observable targets with stationary dynamics in image sequences provided by IR and other imaging sensors (see e.g. 12 ). However, these methods need to be extended to handle maneuvering targets. In this paper, we demonstrate that banks of interacting Bayesian filters (BIBF) can be utilized for this purpose. We are considering target dynamics modeled by jump-linear systems. In contrast to previous studies, we do not assume that the mode jump process is a Markov chain. In particular, we allow the probabilities of jumps to be conditioned on the state variable. Then, we present a computationally e#cient (real time) algorithm for detection and tracking of low observable agile targets. A comparison of BIBF and IMM approaches is carried out in a simple example. Key words: Bayesian, interactive banks filters, IMM, matched filters, track-before-detect. 1. INTRODUCTION This paper is concerned with the detection and tracking of s...

The quickest sequential detection of intrusions in computer networks

Interface, 2003

Rapid response, minimal false alarm rate, and the capability to detect a wide spectrum of attacks... more

A novel approach to detection of denialofservice attacks via adaptive sequential and batchsequential changepoint detection methods

Workshop on Information …

... Here, one particular run of a simulated UDP DoS attack is shown. ... V. DoS Attack Simulation... more

A novel approach to detection of intrusions in computer networks via adaptive sequential and batch-sequential change-point detection methods

IEEE Transactions on Signal Processing, 2006

Large-scale computer network attacks in their final stages can readily be identified by observing... more Large-scale computer network attacks in their final stages can readily be identified by observing very abrupt changes in the network traffic. In the early stage of an attack, however, these changes are hard to detect and difficult to distinguish from usual traffic fluctuations. Rapid response, a minimal false-alarm rate, and the capability to detect a wide spectrum of attacks are the crucial features of intrusion detection systems. In this paper, we develop efficient adaptive sequential and batch-sequential methods for an early detection of attacks that lead to changes in network traffic, such as denial-of-service attacks, worm-based attacks, portscanning, and man-in-the-middle attacks. These methods employ a statistical analysis of data from multiple layers of the network protocol to detect very subtle traffic changes. The algorithms are based on change-point detection theory and utilize a thresholding of test statistics to achieve a fixed rate of false alarms while allowing us to detect changes in statistical models as soon as possible. There are three attractive features of the proposed approach. First, the developed algorithms are self-learning, which enables them to adapt to various network loads and usage patterns. Secondly, they allow for the detection of attacks with a small average delay for a given false-alarm rate. Thirdly, they are computationally simple and thus can be implemented online. Theoretical frameworks for detection procedures are presented. We also give the results of the experimental study with the use of a network simulator testbed as well as real-life testing for TCP SYN flooding attacks.

Download

Conditional inference for incomplete permutation bootstraps in multiple linear regression

Related Authors

Uploads

Papers by Rudolf Blažek

Log In