Filters








577 Hits in 3.7 sec

Anomaly Detection via Graphical Lasso [article]

Haitao Liu, Randy C. Paffenroth, Jian Zou, Chong Zhou
2018 arXiv   pre-print
We accomplish this decomposition by adding an additional ℓ_1 penalty to classic Graphical Lasso, and name it "Robust Graphical Lasso (Rglasso)".  ...  Accordingly, anomaly detection is important both for analyzing the anomalies themselves and for cleaning the data for further analysis of its ambient structure.  ...  Conclusion and Future Work In this paper, we propose a Robust Graphical Lasso (Rglasso) to detect sparse latent effects via Graphical Lasso.  ... 
arXiv:1811.04277v1 fatcat:na64xwv6pfbwdg64atjc7b3rrm

Contrastive Structured Anomaly Detection for Gaussian Graphical Models [article]

Abhinav Maurya, Mark Cheung
2016 arXiv   pre-print
Finding changepoints in the structural evolution of a GGM is therefore essential to detecting anomalies in the underlying system modeled by the GGM.  ...  In order to detect structural anomalies in a GGM, we consider the problem of estimating changes in the precision matrix of the corresponding Gaussian distribution.  ...  Another approach to anomaly detection would be to treat anomaly detection as a two-class modeling problem using joint graphical lasso (Danaher et al., 2014; Yuan & Lin, 2007; Hoefling, 2010) .  ... 
arXiv:1605.00355v1 fatcat:eazoybs4jzectdtcuyh74poibm

Attack Detection Approach by Packet Analysis Using Online Learning with Kernel Method and Correlation Change Method

Ayahiko Niimi, Koki Takahata
2020 International Journal of Intelligent Computing Research  
In addition, we propose an anomaly detection method to detect collapsed correlation via an attack on a network by structural change detection, where HTTP-DNS and syn-ack pairs are used as attributes.  ...  Thus, an anomaly detection and misuse detection based on machine learning and statistical methods for network monitoring is used as countermeasures against cyber-attacks.  ...  Anomaly detection is performed by the method of calculating structure change detection by two kinds of algorithm of graphic lasso and density ratio estimation.  ... 
doi:10.20533/ijicr.2042.4655.2020.0125 fatcat:nfn3denjzngdjnhrwbc4u6ybay

Granger Causality for Time-Series Anomaly Detection

Huida Qiu, Yan Liu, Niranjan A. Subrahmanya, Weichang Li
2012 2012 IEEE 12th International Conference on Data Mining  
Our goal is to efficiently compute a robust "correlation anomaly" score for each variable via Granger graphical models that can provide insights on the possible reasons of anomalies.  ...  In this paper, we propose Granger graphical models as an effective and scalable approach for anomaly detection whose results can be readily interpreted.  ...  via Granger graphical models.  ... 
doi:10.1109/icdm.2012.73 dblp:conf/icdm/QiuLSL12 fatcat:qf2xrcisezdwvgxrjuaftwnucm

Fast anomaly detection in SmartGrids via sparse approximation theory

Marco Levorato, Urbashi Mitra
2012 2012 IEEE 7th Sensor Array and Multichannel Signal Processing Workshop (SAM)  
Detecting anomalies in the behavior of the system requires a large number of observations and is unpractical.  ...  The critical observation behind the proposed framework in that these systems induce an underlying sparse structure which enables dimension reduction via compressed sensing-like schemes.  ...  We first show an example of model detection via Sparse Group LASSO.  ... 
doi:10.1109/sam.2012.6250561 dblp:conf/ieeesam/LevoratoM12 fatcat:uu6ntytgcbfv3ashrvib6b3dri

Social Media Anomaly Detection

Yan Liu, Sanjay Chawla
2017 Proceedings of the Tenth ACM International Conference on Web Search and Data Mining - WSDM '17  
Granger Graphical Models for Anomaly Detection Use Granger-lasso on training data: learn the coefficient βi (a) for each variable xi using lasso regression; Use constrained regression on the test data  ...  π p ), Y p,q ∝ Bernouli(B Zp→,Zp←q ), X p ∝ Multinomial(β Rp ) High computational cost Loose connection of MMSB and LDA components via the shared group membership Group Latent Anomaly Detection (GLAD  ...  We will use Kernel Mean Embedding (KME) to form Higher-Order Statistics Static Graph-based Group Anomaly Detection Graph-based group anomaly detection techniques seek to jointly utilize these observations  ... 
doi:10.1145/3018661.3022757 dblp:conf/wsdm/LiuC17 fatcat:5g3nnq33unbyhnumg4vc4dq5oa

New Insights on Subseasonal Arctic-Midlatitude Causal Connections from a Regularized Regression Model

Marie C. McGraw, Elizabeth A. Barnes
2019 Journal of Climate  
In particular, the regularized regression model results support recent work that indicates that the observed high pressure anomalies over Eurasia drive a significant response in the Arctic on submonthly  ...  The midlatitude circulation, however, Granger causes a nonzero Arctic temperature response via enhancement of the existing stationary wave pattern at lag day 5 (Fig. 2a) , and Z 500 anomalies over Alaska  ...  warm Arctic temperature anomalies), while blue arrows indicate a negative LASSO coefficient (warm Arctic temperature anomalies drive low geopotential height anomalies, and low height anomalies drive warm  ... 
doi:10.1175/jcli-d-19-0142.1 fatcat:xz5dfwmnfzds3nlyo6krxn2kci

Social Media Anomaly Detection

Yan Liu, Sanjay Chawla
2015 Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining - KDD '15  
Granger Graphical Models for Anomaly Detection Use Granger-lasso on training data: learn the coefficientβ i (a) for each variable xi using lasso regression; Use constrained regression on the test data  ...  Group Anomaly Detection Definition Group anomaly or "collective anomaly" detection in social network aims to discover groups of participants that collectively behave anomalously Chandola et al. (2007)  ...  process Measure the number of retweets/shares over time [Bessi (2017)] Cluster based on activity Colluding users will interact with similar items are similar times [Cao et al. (2014)] Example 1: Detecting  ... 
doi:10.1145/2783258.2789990 dblp:conf/kdd/LiuC15 fatcat:olzczq23gbfgvikba5lbf5jlai

A Machine Learning-Based Framework for Building Application Failure Prediction Models

Alessandro Pellegrini, Pierangelo Di Sanzo, Dimiter R. Avresky
2015 2015 IEEE International Parallel and Distributed Processing Symposium Workshop  
Framework for building Failure Prediction Models (F 2 PM), a Machine Learning-based Framework to build models for predicting the Remaining Time to Failure (RTTF) of applications in the presence of software anomalies  ...  ACKNOWLEDGEMENTS The research presented in this paper has been supported by the European Union via the EC funded project PANACEA, contract number FP7 610764.  ...  A framework to automatically detect anomalies and track the performance of an application is proposed in [7] .  ... 
doi:10.1109/ipdpsw.2015.110 dblp:conf/ipps/PellegriniSA15 fatcat:isyozzyjm5el7f4chw27bms6aq

Clustered Fused Graphical Lasso

Yizhi Zhu, Oluwasanmi Koyejo
2018 Conference on Uncertainty in Artificial Intelligence  
We propose Clustered Fused Graphical Lasso (CFGL), a method using precomputed clustering information to improve the signal detectability as compared to typical Fused Graphical Lasso methods.  ...  While usual methods are designed to take advantage of temporal consistency to overcome noise, they conflict with the detectability of anomalies.  ...  The eye state was detected via a camera during the measurement and added later manually after analyzing the video frames.  ... 
dblp:conf/uai/ZhuK18 fatcat:ghruoikmzjeunczrhy4lqofqxm

Uncertainty Quantification and Resource-Demanding Computer Vision Applications of Deep Learning [article]

Julian Burghoff, Robin Chan, Hanno Gottschalk, Annika Muetze, Tobias Riedlinger, Matthias Rottmann, Marius Schubert
2022 arXiv   pre-print
Furthermore, the authors gratefully acknowledge financial support by the state Ministry of Economy, Innovation and Energy of Northrhine Westphalia (MWIDE) and the European Fund for Regional Development via  ...  We introduce a method to detect such unknown objects in semantic segmentation 33 , which is commonly known as the task of anomaly segmentation.  ...  The training of such models requires heavy computer resources, in particular general purpose graphic processing units (GPGPU) 14 .  ... 
arXiv:2205.14917v1 fatcat:mnsszj4wonacpbrkf4ls6tee5q

Network Inference via the Time-Varying Graphical Lasso [article]

David Hallac, Youngsuk Park, Stephen Boyd, Jure Leskovec
2017 arXiv   pre-print
In this paper, we introduce the time-varying graphical lasso (TVGL), a method of inferring time-varying networks from raw time series data.  ...  In order to spot trends, detect anomalies, and interpret the temporal dynamics of such data, it is essential to understand the relationships between the different entities and how these relationships evolve  ...  This work relates to recent advancements in both graphical models and convex optimization. Inferring static networks via the graphical lasso is a well-studied topic [2, 6, 7, 35] .  ... 
arXiv:1703.01958v2 fatcat:hlilnu62xrcgpgt2qgxyiuacve

Network Inference via the Time-Varying Graphical Lasso

David Hallac, Youngsuk Park, Stephen Boyd, Jure Leskovec
2017 Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining - KDD '17  
In this paper, we introduce the time-varying graphical lasso (TVGL), a method of inferring time-varying networks from raw time series data.  ...  In order to spot trends, detect anomalies, and interpret the temporal dynamics of such data, it is essential to understand the relationships between the different entities and how these relationships evolve  ...  Related Work: This work relates to recent advancements in both graphical models and convex optimization. Inferring static networks via the graphical lasso is a well-studied topic [2, 6, 7, 35] .  ... 
doi:10.1145/3097983.3098037 pmid:29770256 pmcid:PMC5951186 dblp:conf/kdd/HallacPBL17 fatcat:qlgzieq7bvgipoalexcbzfthxm

Learning to Estimate Driver Drowsiness from Car Acceleration Sensors Using Weakly Labeled Data

Takayuki Katsuki, Kun Zhao, Takayuki Yoshizumi
2020 ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
As the anomaly detection approaches, we used Lasso [21] and Graphical Lasso (Glasso) [10] , which are traditional but still state-of-the-art methods in anomaly detection for industrial sensors having  ...  Method AUC1 AUC2 Anomaly Lasso 0.41 0.51 detection Glasso 0.36 0.44 Logistic 0.69 0.34 Classification SVM 0.71 0.47 MLP 0.79 0.59 Proposed 0.82 0.69 By using the learned functionf  ... 
doi:10.1109/icassp40776.2020.9053100 dblp:conf/icassp/KatsukiZY20 fatcat:s5kb7nkptbhmzdc6o7d5bltkra

Estimating Dynamic Graphical Models from Multivariate Time-Series Data: Recent Methods and Results [chapter]

Alex J. Gibberd, James D. B. Nelson
2016 Lecture Notes in Computer Science  
The bulk of work in such graphical structure learning problems has focused in the stationary i.i.d setting.  ...  We consider the problem of estimating dynamic graphical models that describe the time-evolving conditional dependency structure between a set of data-streams.  ...  detection.  ... 
doi:10.1007/978-3-319-44412-3_8 fatcat:ocpmpacsnbemfnyxtbzu26no2i
« Previous Showing results 1 — 15 out of 577 results