Advanced search in
Research products
arrow_drop_down
Searching FieldsTerms
Any field
arrow_drop_down
includes
arrow_drop_down
Include:
871 Research products, page 1 of 88

  • Publications
  • Other research products
  • 2018-2022
  • Preprint
  • AE
  • English

10
arrow_drop_down
Relevance
arrow_drop_down
  • Publication . Preprint . Article . 2019
    Open Access English
    Authors: 
    Wheatcroft, Edward; Wynn, Henry; Dent, Chris J.; Smith, Jim Q.; Copeland, Claire L.; Ralph, Daniel; Zachary, Stan;

    Scenario Analysis is a risk assessment tool that aims to evaluate the impact of a small number of distinct plausible future scenarios. In this paper, we provide an overview of important aspects of Scenario Analysis including when it is appropriate, the design of scenarios, uncertainty and encouraging creativity. Each of these issues is discussed in the context of climate, energy and legal scenarios.

  • Open Access English
    Authors: 
    Othman Benomar; M. J. Goupil; Kevin Belkacem; T. Appourchaux; Martin Bo Nielsen; M. Bazot; Laurent Gizon; Shravan M. Hanasoge; Katepalli R. Sreenivasan; B. Marchand;
    Publisher: HAL CCSD
    Country: France

    Oscillation properties are usually measured by fitting symmetric Lorentzian profiles to the power spectra of Sun-like stars. However the line profiles of solar oscillations have been observed to be asymmetrical for the Sun. The physical origin of this line asymmetry is not fully understood, although it should depend on the depth dependence of the source of wave excitation (convective turbulence) and details of the observable (velocity or intensity). For oscillations of the Sun, it has been shown that neglecting the asymmetry leads to systematic errors in the frequency determination. This could subsequently affects the results of seismic inferences of the solar internal structure. Using light curves from the {\it Kepler} spacecraft we have measured mode asymmetries in 43 stars. We confirm that neglecting the asymmetry leads to systematic errors that can exceed the $1\sigma$ confidence intervals for seismic observations longer than one year. Therefore, the application of an asymmetric Lorentzian profile is to be favoured to improve the accuracy of the internal stellar structure and stellar fundamental parameters. We also show that the asymmetry changes sign between cool Sun-like stars and hotter stars. This provides the best constraints to date on the location of the excitation sources across the Hertzsprung-Russel diagram. Comment: 8 pages, 7 Figures, 1 Table, Accepted to ApJ

  • Open Access English
    Authors: 
    Khaled Ai Thelaya; Marco Agus; Jens Schneider;

    In this paper, we present a novel data structure, called the Mixture Graph. This data structure allows us to compress, render, and query segmentation histograms. Such histograms arise when building a mipmap of a volume containing segmentation IDs. Each voxel in the histogram mipmap contains a convex combination (mixture) of segmentation IDs. Each mixture represents the distribution of IDs in the respective voxel's children. Our method factorizes these mixtures into a series of linear interpolations between exactly two segmentation IDs. The result is represented as a directed acyclic graph (DAG) whose nodes are topologically ordered. Pruning replicate nodes in the tree followed by compression allows us to store the resulting data structure efficiently. During rendering, transfer functions are propagated from sources (leafs) through the DAG to allow for efficient, pre-filtered rendering at interactive frame rates. Assembly of histogram contributions across the footprint of a given volume allows us to efficiently query partial histograms, achieving up to 178$\times$ speed-up over na$\mathrm{\"{i}}$ve parallelized range queries. Additionally, we apply the Mixture Graph to compute correctly pre-filtered volume lighting and to interactively explore segments based on shape, geometry, and orientation using multi-dimensional transfer functions. Comment: To appear in IEEE Transacations on Visualization and Computer Graphics (IEEE Vis 2020)

  • Publication . Conference object . Preprint . Article . 2020
    Open Access English

    This research aims at identifying the unknown emotion using speaker cues. In this study, we identify the unknown emotion using a two-stage framework. The first stage focuses on identifying the speaker who uttered the unknown emotion, while the next stage focuses on identifying the unknown emotion uttered by the recognized speaker in the prior stage. This proposed framework has been evaluated on an Arabic Emirati-accented speech database uttered by fifteen speakers per gender. Mel-Frequency Cepstral Coefficients (MFCCs) have been used as the extracted features and Hidden Markov Model (HMM) has been utilized as the classifier in this work. Our findings demonstrate that emotion recognition accuracy based on the two-stage framework is greater than that based on the one-stage approach and the state-of-the-art classifiers and models such as Gaussian Mixture Model (GMM), Support Vector Machine (SVM), and Vector Quantization (VQ). The average emotion recognition accuracy based on the two-stage approach is 67.5%, while the accuracy reaches to 61.4%, 63.3%, 64.5%, and 61.5%, based on the one-stage approach, GMM, SVM, and VQ, respectively. The achieved results based on the two-stage framework are very close to those attained in subjective assessment by human listeners. 5 pages

  • Publication . Conference object . Preprint . Article . 2019
    Open Access English
    Authors: 
    Breton Minnehan; Andreas Savakis;

    We propose a data-driven approach for deep convolutional neural network compression that achieves high accuracy with high throughput and low memory requirements. Current network compression methods either find a low-rank factorization of the features that requires more memory, or select only a subset of features by pruning entire filter channels. We propose the Cascaded Projection (CaP) compression method that projects the output and input filter channels of successive layers to a unified low dimensional space based on a low-rank projection. We optimize the projection to minimize classification loss and the difference between the next layer's features in the compressed and uncompressed networks. To solve this non-convex optimization problem we propose a new optimization method of a proxy matrix using backpropagation and Stochastic Gradient Descent (SGD) with geometric constraints. Our cascaded projection approach leads to improvements in all critical areas of network compression: high accuracy, low memory consumption, low parameter count and high processing speed. The proposed CaP method demonstrates state-of-the-art results compressing VGG16 and ResNet networks with over 4x reduction in the number of computations and excellent performance in top-5 accuracy on the ImageNet dataset before and after fine-tuning.

  • Open Access English
    Authors: 
    Randa Asa'd; Paul Goudfrooij;

    We investigate the precision of the ages and metallicities of 21,000 mock simple stellar populations (SSPs) determined through full-spectrum fitting. The mock SSPs cover an age range of 6.8 $<$ log (age/yr) $<$ 10.2, for three wavelength ranges in the optical regime, using both Padova and MIST isochrone models. Random noise is added to the model spectra to achieve S/N ratios between 10 to 100 per wavelength pixel. We find that for S/N $\geq$ 50, this technique can yield ages of SSPs to an overall precision of $\Delta\,\mbox{log(age/yr)} \sim 0.1$ for ages in the ranges 7.0 $\leq$ log (age/yr) $\leq$ 8.3 and 8.9 $\leq$ log (age/yr) $\leq$ 9.4. For the age ranges of 8.3 $\leq$ log (age/yr) $\leq$ 8.9 and log (age/yr) $\geq$ 9.5, which have significant flux contributions from asymptotic giant branch (AGB) and red giant branch (RGB) stars, respectively, the age uncertainty rises to about $\pm 0.3$ dex. The precision of age and metallicity estimation using this method depends significantly on the S/N and the wavelength range used in the fitting. We quantify the systematic differences in age predicted by the MIST and Padova isochrone models, due to their different assumptions about stellar physics in various important (i.e., luminous) phases of stellar evolution, which needs to be taken in consideration when comparing ages of star clusters obtained using these popular models. Knowing the strengths and limitations of this technique is crucial in interpreting the results obtained for real star clusters and for deciding the optimal instrument setup before performing the observations. Comment: 44 pages, 20 of which contain text and 17 figures; accepted for publication in MNRAS. The 24 tables on pages 21-44 will be made available as supplementary online material at the MNRAS site

  • Open Access English
    Authors: 
    Prince Romeo Mensah;

    We study a singular limit of a scaled compressible Navier--Stokes--Coriolis system driven by both a deterministic and stochastic forcing terms in three dimensions. If the Mach number is comparable to the Froude number with both proportional to say $\varepsilon\ll 1$, whereas the Rossby number scales like $\varepsilon^m$ for $m>1$ large, then we show that any family of weak martingale solution to the $3$-D randomly forced rotating compressible equation (under the influence of a deterministic centrifugal force) converges in probability, as $\varepsilon\rightarrow0$, to the $2$-D incompressible Navier--Stokes system with a corresponding random forcing term. 40 pages

  • Publication . Article . Preprint . 2020
    Open Access English
    Authors: 
    Sangchul Oh; Jung Jun Park; Hyunchul Nha;
    Publisher: MDPI AG

    We investigate the quantum thermodynamics of two quantum systems, a two-level system and a four-level quantum photocell, each driven by photon pulses as a quantum heat engine. We set these systems to be in thermal contact only with a cold reservoir while the heat (energy) source, conventionally given from a hot thermal reservoir, is supplied by a sequence of photon pulses. The dynamics of each system is governed by a coherent interaction due to photon pulses in terms of the Jaynes-Cummings Hamiltonian together with the system-bath interaction described by the Lindblad master equation. We calculate the thermodynamic quantities for the two-level system and the quantum photocell including the change in system energy, power delivered by photon pulses, power output to an external load, heat dissipated to a cold bath, and entropy production. We thereby demonstrate how a quantum photocell in the cold bath can operate as a continuum quantum heat engine with the sequence of photon pulses continuously applied. We specifically introduce the power efficiency of the quantum photocell in terms of the ratio of output power delivered to an external load with current and voltage to the input power delivered by the photon pulse. Our study indicates a possibility that a quantum system driven by external fields can act as an efficient quantum heat engine under non-equilibrium thermodynamics. Comment: 10 pages, 8 figures, submitted

  • Open Access English
    Authors: 
    Yanping Chen; Renbin Yan; Claudia Maraston; Daniel Thomas; Guy S. Stringfellow; Dmitry Bizyaev; Joseph D. Gelfand; Timothy C. Beers; José G. Fernández-Trincado; Daniel Lazarz; +3 more
    Country: United Kingdom

    We report the stellar atmospheric parameters for 7503 spectra contained in the first release of the MaNGA stellar library (MaStar) in SDSS DR15. The first release of MaStar contains 8646 spectra measured from 3321 unique stars, each covering the wavelength range 3622 \AA\ to 10354 \AA\ with a resolving power of $R \sim$ 1800. In this work, we first determined the basic stellar parameters: effective temperature ($\rm T_{eff}$), surface gravity ($\log g$), and metallicity ($\rm[Fe/H]$), which best fit the data using an empirical interpolator based on the Medium-resolution Isaac Newton Telescope library of empirical spectra (MILES), as implemented by the University of Lyon Spectroscopic analysis Software (Koleva et al. 2008, ULySS) package. While we analyzed all 8646 spectra from the first release of MaStar, since MaStar has a wider parameter-space coverage than MILES, not all of these fits are robust. In addition, not all parameter regions covered by MILES yield robust results, likely due to the non-uniform coverage of the parameter space by MILES. We tested the robustness of the method using the MILES spectra itself and identified a proxy based on the local density of the training set. With this proxy, we identified 7503 MaStar spectra with robust fitting results. They cover the range from 3179K to 20,517K in effective temperature ($\rm T_{eff}$), from 0.40 to 5.0 in surface gravity ($\log g$), and from $-$2.49 to $+$0.73 in metallicity ($\rm[Fe/H]$). Comment: 14 pages, 13 figures

  • Publication . Article . Preprint . Conference object . 2018
    Open Access English
    Authors: 
    Barton, A.; Imani, M.; Ming, J.; Matthew Wright;

    The Tor anonymity system provides online privacy for millions of users, but it is slower than typical web browsing. To improve Tor performance, we propose PredicTor, a path selection technique that uses a Random Forest classifier trained on recent measurements of Tor to predict the performance of a proposed path. If the path is predicted to be fast, then the client builds a circuit using those relays. We implemented PredicTor in the Tor source code and show through live Tor experiments and Shadow simulations that PredicTor improves Tor network performance by 11% to 23% compared to Vanilla Tor and by 7% to 13% compared to the previous state-of-the-art scheme. Our experiments show that PredicTor is the first path selection algorithm to dynamically avoid highly congested nodes during times of high congestion and avoid long-distance paths during times of low congestion. We evaluate the anonymity of PredicTor using standard entropy-based and time-to-first-compromise metrics, but these cannot capture the possibility of leakage due to the use of location in path selection. To better address this, we propose a new anonymity metric called CLASI: Client Autonomous System Inference. CLASI is the first anonymity metric in Tor that measures an adversary's ability to infer client Autonomous Systems (ASes) by fingerprinting circuits at the network, country, and relay level. We find that CLASI shows anonymity loss for location-aware path selection algorithms, where entropy-based metrics show little to no loss of anonymity. Additionally, CLASI indicates that PredicTor has similar sender AS leakage compared to the current Tor path selection algorithm due to PredicTor building circuits that are independent of client location.

Advanced search in
Research products
arrow_drop_down
Searching FieldsTerms
Any field
arrow_drop_down
includes
arrow_drop_down
Include:
871 Research products, page 1 of 88
  • Publication . Preprint . Article . 2019
    Open Access English
    Authors: 
    Wheatcroft, Edward; Wynn, Henry; Dent, Chris J.; Smith, Jim Q.; Copeland, Claire L.; Ralph, Daniel; Zachary, Stan;

    Scenario Analysis is a risk assessment tool that aims to evaluate the impact of a small number of distinct plausible future scenarios. In this paper, we provide an overview of important aspects of Scenario Analysis including when it is appropriate, the design of scenarios, uncertainty and encouraging creativity. Each of these issues is discussed in the context of climate, energy and legal scenarios.

  • Open Access English
    Authors: 
    Othman Benomar; M. J. Goupil; Kevin Belkacem; T. Appourchaux; Martin Bo Nielsen; M. Bazot; Laurent Gizon; Shravan M. Hanasoge; Katepalli R. Sreenivasan; B. Marchand;
    Publisher: HAL CCSD
    Country: France

    Oscillation properties are usually measured by fitting symmetric Lorentzian profiles to the power spectra of Sun-like stars. However the line profiles of solar oscillations have been observed to be asymmetrical for the Sun. The physical origin of this line asymmetry is not fully understood, although it should depend on the depth dependence of the source of wave excitation (convective turbulence) and details of the observable (velocity or intensity). For oscillations of the Sun, it has been shown that neglecting the asymmetry leads to systematic errors in the frequency determination. This could subsequently affects the results of seismic inferences of the solar internal structure. Using light curves from the {\it Kepler} spacecraft we have measured mode asymmetries in 43 stars. We confirm that neglecting the asymmetry leads to systematic errors that can exceed the $1\sigma$ confidence intervals for seismic observations longer than one year. Therefore, the application of an asymmetric Lorentzian profile is to be favoured to improve the accuracy of the internal stellar structure and stellar fundamental parameters. We also show that the asymmetry changes sign between cool Sun-like stars and hotter stars. This provides the best constraints to date on the location of the excitation sources across the Hertzsprung-Russel diagram. Comment: 8 pages, 7 Figures, 1 Table, Accepted to ApJ

  • Open Access English
    Authors: 
    Khaled Ai Thelaya; Marco Agus; Jens Schneider;

    In this paper, we present a novel data structure, called the Mixture Graph. This data structure allows us to compress, render, and query segmentation histograms. Such histograms arise when building a mipmap of a volume containing segmentation IDs. Each voxel in the histogram mipmap contains a convex combination (mixture) of segmentation IDs. Each mixture represents the distribution of IDs in the respective voxel's children. Our method factorizes these mixtures into a series of linear interpolations between exactly two segmentation IDs. The result is represented as a directed acyclic graph (DAG) whose nodes are topologically ordered. Pruning replicate nodes in the tree followed by compression allows us to store the resulting data structure efficiently. During rendering, transfer functions are propagated from sources (leafs) through the DAG to allow for efficient, pre-filtered rendering at interactive frame rates. Assembly of histogram contributions across the footprint of a given volume allows us to efficiently query partial histograms, achieving up to 178$\times$ speed-up over na$\mathrm{\"{i}}$ve parallelized range queries. Additionally, we apply the Mixture Graph to compute correctly pre-filtered volume lighting and to interactively explore segments based on shape, geometry, and orientation using multi-dimensional transfer functions. Comment: To appear in IEEE Transacations on Visualization and Computer Graphics (IEEE Vis 2020)

  • Publication . Conference object . Preprint . Article . 2020
    Open Access English

    This research aims at identifying the unknown emotion using speaker cues. In this study, we identify the unknown emotion using a two-stage framework. The first stage focuses on identifying the speaker who uttered the unknown emotion, while the next stage focuses on identifying the unknown emotion uttered by the recognized speaker in the prior stage. This proposed framework has been evaluated on an Arabic Emirati-accented speech database uttered by fifteen speakers per gender. Mel-Frequency Cepstral Coefficients (MFCCs) have been used as the extracted features and Hidden Markov Model (HMM) has been utilized as the classifier in this work. Our findings demonstrate that emotion recognition accuracy based on the two-stage framework is greater than that based on the one-stage approach and the state-of-the-art classifiers and models such as Gaussian Mixture Model (GMM), Support Vector Machine (SVM), and Vector Quantization (VQ). The average emotion recognition accuracy based on the two-stage approach is 67.5%, while the accuracy reaches to 61.4%, 63.3%, 64.5%, and 61.5%, based on the one-stage approach, GMM, SVM, and VQ, respectively. The achieved results based on the two-stage framework are very close to those attained in subjective assessment by human listeners. 5 pages

  • Publication . Conference object . Preprint . Article . 2019
    Open Access English
    Authors: 
    Breton Minnehan; Andreas Savakis;

    We propose a data-driven approach for deep convolutional neural network compression that achieves high accuracy with high throughput and low memory requirements. Current network compression methods either find a low-rank factorization of the features that requires more memory, or select only a subset of features by pruning entire filter channels. We propose the Cascaded Projection (CaP) compression method that projects the output and input filter channels of successive layers to a unified low dimensional space based on a low-rank projection. We optimize the projection to minimize classification loss and the difference between the next layer's features in the compressed and uncompressed networks. To solve this non-convex optimization problem we propose a new optimization method of a proxy matrix using backpropagation and Stochastic Gradient Descent (SGD) with geometric constraints. Our cascaded projection approach leads to improvements in all critical areas of network compression: high accuracy, low memory consumption, low parameter count and high processing speed. The proposed CaP method demonstrates state-of-the-art results compressing VGG16 and ResNet networks with over 4x reduction in the number of computations and excellent performance in top-5 accuracy on the ImageNet dataset before and after fine-tuning.

  • Open Access English
    Authors: 
    Randa Asa'd; Paul Goudfrooij;

    We investigate the precision of the ages and metallicities of 21,000 mock simple stellar populations (SSPs) determined through full-spectrum fitting. The mock SSPs cover an age range of 6.8 $<$ log (age/yr) $<$ 10.2, for three wavelength ranges in the optical regime, using both Padova and MIST isochrone models. Random noise is added to the model spectra to achieve S/N ratios between 10 to 100 per wavelength pixel. We find that for S/N $\geq$ 50, this technique can yield ages of SSPs to an overall precision of $\Delta\,\mbox{log(age/yr)} \sim 0.1$ for ages in the ranges 7.0 $\leq$ log (age/yr) $\leq$ 8.3 and 8.9 $\leq$ log (age/yr) $\leq$ 9.4. For the age ranges of 8.3 $\leq$ log (age/yr) $\leq$ 8.9 and log (age/yr) $\geq$ 9.5, which have significant flux contributions from asymptotic giant branch (AGB) and red giant branch (RGB) stars, respectively, the age uncertainty rises to about $\pm 0.3$ dex. The precision of age and metallicity estimation using this method depends significantly on the S/N and the wavelength range used in the fitting. We quantify the systematic differences in age predicted by the MIST and Padova isochrone models, due to their different assumptions about stellar physics in various important (i.e., luminous) phases of stellar evolution, which needs to be taken in consideration when comparing ages of star clusters obtained using these popular models. Knowing the strengths and limitations of this technique is crucial in interpreting the results obtained for real star clusters and for deciding the optimal instrument setup before performing the observations. Comment: 44 pages, 20 of which contain text and 17 figures; accepted for publication in MNRAS. The 24 tables on pages 21-44 will be made available as supplementary online material at the MNRAS site

  • Open Access English
    Authors: 
    Prince Romeo Mensah;

    We study a singular limit of a scaled compressible Navier--Stokes--Coriolis system driven by both a deterministic and stochastic forcing terms in three dimensions. If the Mach number is comparable to the Froude number with both proportional to say $\varepsilon\ll 1$, whereas the Rossby number scales like $\varepsilon^m$ for $m>1$ large, then we show that any family of weak martingale solution to the $3$-D randomly forced rotating compressible equation (under the influence of a deterministic centrifugal force) converges in probability, as $\varepsilon\rightarrow0$, to the $2$-D incompressible Navier--Stokes system with a corresponding random forcing term. 40 pages

  • Publication . Article . Preprint . 2020
    Open Access English
    Authors: 
    Sangchul Oh; Jung Jun Park; Hyunchul Nha;
    Publisher: MDPI AG

    We investigate the quantum thermodynamics of two quantum systems, a two-level system and a four-level quantum photocell, each driven by photon pulses as a quantum heat engine. We set these systems to be in thermal contact only with a cold reservoir while the heat (energy) source, conventionally given from a hot thermal reservoir, is supplied by a sequence of photon pulses. The dynamics of each system is governed by a coherent interaction due to photon pulses in terms of the Jaynes-Cummings Hamiltonian together with the system-bath interaction described by the Lindblad master equation. We calculate the thermodynamic quantities for the two-level system and the quantum photocell including the change in system energy, power delivered by photon pulses, power output to an external load, heat dissipated to a cold bath, and entropy production. We thereby demonstrate how a quantum photocell in the cold bath can operate as a continuum quantum heat engine with the sequence of photon pulses continuously applied. We specifically introduce the power efficiency of the quantum photocell in terms of the ratio of output power delivered to an external load with current and voltage to the input power delivered by the photon pulse. Our study indicates a possibility that a quantum system driven by external fields can act as an efficient quantum heat engine under non-equilibrium thermodynamics. Comment: 10 pages, 8 figures, submitted

  • Open Access English
    Authors: 
    Yanping Chen; Renbin Yan; Claudia Maraston; Daniel Thomas; Guy S. Stringfellow; Dmitry Bizyaev; Joseph D. Gelfand; Timothy C. Beers; José G. Fernández-Trincado; Daniel Lazarz; +3 more
    Country: United Kingdom

    We report the stellar atmospheric parameters for 7503 spectra contained in the first release of the MaNGA stellar library (MaStar) in SDSS DR15. The first release of MaStar contains 8646 spectra measured from 3321 unique stars, each covering the wavelength range 3622 \AA\ to 10354 \AA\ with a resolving power of $R \sim$ 1800. In this work, we first determined the basic stellar parameters: effective temperature ($\rm T_{eff}$), surface gravity ($\log g$), and metallicity ($\rm[Fe/H]$), which best fit the data using an empirical interpolator based on the Medium-resolution Isaac Newton Telescope library of empirical spectra (MILES), as implemented by the University of Lyon Spectroscopic analysis Software (Koleva et al. 2008, ULySS) package. While we analyzed all 8646 spectra from the first release of MaStar, since MaStar has a wider parameter-space coverage than MILES, not all of these fits are robust. In addition, not all parameter regions covered by MILES yield robust results, likely due to the non-uniform coverage of the parameter space by MILES. We tested the robustness of the method using the MILES spectra itself and identified a proxy based on the local density of the training set. With this proxy, we identified 7503 MaStar spectra with robust fitting results. They cover the range from 3179K to 20,517K in effective temperature ($\rm T_{eff}$), from 0.40 to 5.0 in surface gravity ($\log g$), and from $-$2.49 to $+$0.73 in metallicity ($\rm[Fe/H]$). Comment: 14 pages, 13 figures

  • Publication . Article . Preprint . Conference object . 2018
    Open Access English
    Authors: 
    Barton, A.; Imani, M.; Ming, J.; Matthew Wright;

    The Tor anonymity system provides online privacy for millions of users, but it is slower than typical web browsing. To improve Tor performance, we propose PredicTor, a path selection technique that uses a Random Forest classifier trained on recent measurements of Tor to predict the performance of a proposed path. If the path is predicted to be fast, then the client builds a circuit using those relays. We implemented PredicTor in the Tor source code and show through live Tor experiments and Shadow simulations that PredicTor improves Tor network performance by 11% to 23% compared to Vanilla Tor and by 7% to 13% compared to the previous state-of-the-art scheme. Our experiments show that PredicTor is the first path selection algorithm to dynamically avoid highly congested nodes during times of high congestion and avoid long-distance paths during times of low congestion. We evaluate the anonymity of PredicTor using standard entropy-based and time-to-first-compromise metrics, but these cannot capture the possibility of leakage due to the use of location in path selection. To better address this, we propose a new anonymity metric called CLASI: Client Autonomous System Inference. CLASI is the first anonymity metric in Tor that measures an adversary's ability to infer client Autonomous Systems (ASes) by fingerprinting circuits at the network, country, and relay level. We find that CLASI shows anonymity loss for location-aware path selection algorithms, where entropy-based metrics show little to no loss of anonymity. Additionally, CLASI indicates that PredicTor has similar sender AS leakage compared to the current Tor path selection algorithm due to PredicTor building circuits that are independent of client location.

Send a message
How can we help?
We usually respond in a few hours.