-
Adaptive Thresholding Heuristic for KPI Anomaly Detection
Authors:
Ebenezer R. H. P. Isaac,
Akshat Sharma
Abstract:
A plethora of outlier detectors have been explored in the time series domain, however, in a business sense, not all outliers are anomalies of interest. Existing anomaly detection solutions are confined to certain outlier detectors limiting their applicability to broader anomaly detection use cases. Network KPIs (Key Performance Indicators) tend to exhibit stochastic behaviour producing statistical…
▽ More
A plethora of outlier detectors have been explored in the time series domain, however, in a business sense, not all outliers are anomalies of interest. Existing anomaly detection solutions are confined to certain outlier detectors limiting their applicability to broader anomaly detection use cases. Network KPIs (Key Performance Indicators) tend to exhibit stochastic behaviour producing statistical outliers, most of which do not adversely affect business operations. Thus, a heuristic is required to capture the business definition of an anomaly for time series KPI. This article proposes an Adaptive Thresholding Heuristic (ATH) to dynamically adjust the detection threshold based on the local properties of the data distribution and adapt to changes in time series patterns. The heuristic derives the threshold based on the expected periodicity and the observed proportion of anomalies minimizing false positives and addressing concept drift. ATH can be used in conjunction with any underlying seasonality decomposition method and an outlier detector that yields an outlier score. This method has been tested on EON1-Cell-U, a labeled KPI anomaly dataset produced by Ericsson, to validate our hypothesis. Experimental results show that ATH is computationally efficient making it scalable for near real time anomaly detection and flexible with multiple forecasters and outlier detectors.
△ Less
Submitted 21 August, 2023;
originally announced August 2023.
-
QBSD: Quartile-Based Seasonality Decomposition for Cost-Effective Time Series Forecasting
Authors:
Ebenezer RHP Isaac,
Bulbul Singh
Abstract:
In the telecom domain, precise forecasting of time series patterns, such as cell key performance indicators (KPIs), plays a pivotal role in enhancing service quality and operational efficiency. State-of-the-art forecasting approaches prioritize forecasting accuracy at the expense of computational performance, rendering them less suitable for data-intensive applications encompassing systems with a…
▽ More
In the telecom domain, precise forecasting of time series patterns, such as cell key performance indicators (KPIs), plays a pivotal role in enhancing service quality and operational efficiency. State-of-the-art forecasting approaches prioritize forecasting accuracy at the expense of computational performance, rendering them less suitable for data-intensive applications encompassing systems with a multitude of time series variables. To address this issue, we introduce QBSD, a live forecasting approach tailored to optimize the trade-off between accuracy and computational complexity. We have evaluated the performance of QBSD against state-of-the-art forecasting approaches on publicly available datasets. We have also extended this investigation to our curated network KPI dataset, now publicly accessible, to showcase the effect of dynamic operating ranges that varies with time. The results demonstrate that the proposed method excels in runtime efficiency compared to the leading algorithms available while maintaining competitive forecast accuracy.
△ Less
Submitted 16 August, 2023; v1 submitted 9 June, 2023;
originally announced June 2023.
-
AI Product Security: A Primer for Developers
Authors:
Ebenezer R. H. P. Isaac,
Jim Reno
Abstract:
Not too long ago, AI security used to mean the research and practice of how AI can empower cybersecurity, that is, AI for security. Ever since Ian Goodfellow and his team popularized adversarial attacks on machine learning, security for AI became an important concern and also part of AI security. It is imperative to understand the threats to machine learning products and avoid common pitfalls in A…
▽ More
Not too long ago, AI security used to mean the research and practice of how AI can empower cybersecurity, that is, AI for security. Ever since Ian Goodfellow and his team popularized adversarial attacks on machine learning, security for AI became an important concern and also part of AI security. It is imperative to understand the threats to machine learning products and avoid common pitfalls in AI product development. This article is addressed to developers, designers, managers and researchers of AI software products.
△ Less
Submitted 18 April, 2023;
originally announced April 2023.
-
Robust Analytics for Video-Based Gait Biometrics
Authors:
Ebenezer R. H. P. Isaac
Abstract:
Gait analysis is the study of the systematic methods that assess and quantify animal locomotion. Gait finds a unique importance among the many state-of-the-art biometric systems since it does not require the subject's cooperation to the extent required by other modalities. Hence by nature, it is an unobtrusive biometric.
This thesis discusses both hard and soft biometric characteristics of gait.…
▽ More
Gait analysis is the study of the systematic methods that assess and quantify animal locomotion. Gait finds a unique importance among the many state-of-the-art biometric systems since it does not require the subject's cooperation to the extent required by other modalities. Hence by nature, it is an unobtrusive biometric.
This thesis discusses both hard and soft biometric characteristics of gait. It shows how to identify gender based on gait alone through the Posed-Based Voting scheme. It then describes improving gait recognition accuracy using Genetic Template Segmentation. Members of a wide population can be authenticated using Multiperson Signature Map**. Finally, the map** can be improved in a smaller population using Bayesian Thresholding. All methods proposed in this thesis have outperformed their existing state of the art with adequate experimentation and results.
△ Less
Submitted 12 November, 2021;
originally announced November 2021.
-
Trait of Gait: A Survey on Gait Biometrics
Authors:
Ebenezer R. H. P. Isaac,
Susan Elias,
Srinivasan Rajagopalan,
K. S. Easwarakumar
Abstract:
Gait analysis is the study of the systematic methods that assess and quantify animal locomotion. The research on gait analysis has considerably evolved through time. It was an ancient art, and it still finds its application today in modern science and medicine. This paper describes how one's gait can be used as a biometric. It shall diversely cover salient research done within the field and explai…
▽ More
Gait analysis is the study of the systematic methods that assess and quantify animal locomotion. The research on gait analysis has considerably evolved through time. It was an ancient art, and it still finds its application today in modern science and medicine. This paper describes how one's gait can be used as a biometric. It shall diversely cover salient research done within the field and explain the nuances and advances in each type of gait analysis. The prominent methods of gait recognition from the early era to the state of the art are covered. This survey also reviews the various gait datasets. The overall aim of this study is to provide a concise roadmap for anyone who wishes to do research in the field of gait biometrics.
△ Less
Submitted 26 March, 2019;
originally announced March 2019.
-
Cryogenic positioning and alignment with micrometer precision
Authors:
Corinne E. Isaac,
Elizabeth A. Curley,
Pamela T. Nasr,
Hoang L. Nguyen,
John A. Marohn
Abstract:
Aligning a microcantilever to an area of interest on a sample is a critical step in many scanning probe microscopy experiments, particularly those carried out on devices and rare, precious samples. We report a series of protocols that rapidly and reproducibly align a high-compliance microcantilever to a $< 10 \: μ\mathrm{m}$ sample feature under high vacuum and at cryogenic temperatures. The first…
▽ More
Aligning a microcantilever to an area of interest on a sample is a critical step in many scanning probe microscopy experiments, particularly those carried out on devices and rare, precious samples. We report a series of protocols that rapidly and reproducibly align a high-compliance microcantilever to a $< 10 \: μ\mathrm{m}$ sample feature under high vacuum and at cryogenic temperatures. The first set of protocols, applicable to a cantilever oscillating parallel to the sample surface, involve monitoring the cantilever resonance frequency while laterally scanning the tip to map the sample substrate through electrostatic interactions of the substrate with the cantilever. We demonstrate that when operating a cantilever a few micrometers from the sample surface, large shifts in the cantilever resonance frequency are present near the edges of a voltage-biased sample electrode. Surprisingly, these "edge-finder" frequency shifts are retained when the electrode is coated with a polymer film and a $\sim 10 \: \mathrm{nm}$ thick metallic ground plane. The second series of methods, applicable to any scanning probe microscopy experiment, integrate a single-optical fiber to image line scans of the sample surface. The microscope modifications required for these methods are straightforward to implement, provide reliable micrometer-scale positioning, and decrease the experimental setup time from days to hours in a vacuum, cryogenic magnetic resonance force microscope.
△ Less
Submitted 8 January, 2018; v1 submitted 3 October, 2017;
originally announced October 2017.
-
View-invariant Gait Recognition through Genetic Template Segmentation
Authors:
Ebenezer Isaac,
Susan Elias,
Srinivasan Rajagopalan,
K. S. Easwarakumar
Abstract:
Template-based model-free approach provides by far the most successful solution to the gait recognition problem in literature. Recent work discusses how isolating the head and leg portion of the template increase the performance of a gait recognition system making it robust against covariates like clothing and carrying conditions. However, most involve a manual definition of the boundaries. The me…
▽ More
Template-based model-free approach provides by far the most successful solution to the gait recognition problem in literature. Recent work discusses how isolating the head and leg portion of the template increase the performance of a gait recognition system making it robust against covariates like clothing and carrying conditions. However, most involve a manual definition of the boundaries. The method we propose, the genetic template segmentation (GTS), employs the genetic algorithm to automate the boundary selection process. This method was tested on the GEI, GEnI and AEI templates. GEI seems to exhibit the best result when segmented with our approach. Experimental results depict that our approach significantly outperforms the existing implementations of view-invariant gait recognition.
△ Less
Submitted 3 July, 2017; v1 submitted 15 May, 2017;
originally announced May 2017.
-
Dynamic nuclear polarization in a magnetic resonance force microscope experiment
Authors:
Corinne E. Isaac,
Christine M. Gleave,
Paméla T. Nasr,
Hoang L. Nguyen,
Elizabeth A. Curley,
Jonilyn L. Yoder,
Eric W. Moore,
Lei Chen,
John A. Marohn
Abstract:
We report achieving enhanced nuclear magnetization in a magnetic resonance force microscope experiment at 0.6 tesla and 4.2 kelvin using the dynamic nuclear polarization (DNP) effect. In our experiments a microwire coplanar waveguide delivered radiowaves to excite nuclear spins and microwaves to excite electron spins in a 250 nm thick nitroxide-doped polystyrene sample. Both electron and proton sp…
▽ More
We report achieving enhanced nuclear magnetization in a magnetic resonance force microscope experiment at 0.6 tesla and 4.2 kelvin using the dynamic nuclear polarization (DNP) effect. In our experiments a microwire coplanar waveguide delivered radiowaves to excite nuclear spins and microwaves to excite electron spins in a 250 nm thick nitroxide-doped polystyrene sample. Both electron and proton spin resonance were observed as a change in the mechanical resonance frequency of a nearby cantilever having a micron-scale nickel tip. NMR signal, not observable from Curie-law magnetization at 0.6 tesla, became observable when microwave irradiation was applied to saturate the electron spins. The resulting NMR signal's size, buildup time, dependence on microwave power, and dependence on irradiation frequency was consistent with a transfer of magnetization from electron spins to nuclear spins. Due to the presence of an inhomogenous magnetic field introduced by the cantilever's magnetic tip, the electron spins in the sample were saturated in a microwave-resonant slice 10's of nm thick. The spatial distribution of the nuclear polarization enhancement factor $ε$ was mapped by varying the frequency of the applied radiowaves. The observed enhancement factor was zero for spins in the center of the resonant slice, was $ε= +10$ to $+20$ for spins proximal to the magnet, and was $ε= -10$ to $-20$ for spins distal to the magnet. We show that this bipolar nuclear magnetization profile is consistent with cross-effect DNP in a $\sim \! 10^{5} \: \mathrm{T} \: \mathrm{m}^{-1}$ magnetic field gradient. Potential challenges associated with generating and using DNP-enhanced nuclear magnetization in a nanometer-resolution magnetic resonance imaging experiment are elucidated and discussed.
△ Less
Submitted 18 February, 2016; v1 submitted 26 January, 2016;
originally announced January 2016.
-
Modeling the dynamics of a tracer particle in an elastic active gel
Authors:
E. Ben Isaac,
É. Fodor,
P. Visco,
F. van Wijland,
N. S. Gov
Abstract:
The internal dynamics of active gels, both in artificial (in-vitro) model systems and inside the cytoskeleton of living cells, has been extensively studied by experiments of recent years. These dynamics are probed using tracer particles embedded in the network of biopolymers together with molecular motors, and distinct non-thermal behavior is observed. We present a theoretical model of the dynamic…
▽ More
The internal dynamics of active gels, both in artificial (in-vitro) model systems and inside the cytoskeleton of living cells, has been extensively studied by experiments of recent years. These dynamics are probed using tracer particles embedded in the network of biopolymers together with molecular motors, and distinct non-thermal behavior is observed. We present a theoretical model of the dynamics of a trapped active particle, which allows us to quantify the deviations from equilibrium behavior, using both analytic and numerical calculations. We map the different regimes of dynamics in this system, and highlight the different manifestations of activity: breakdown of the virial theorem and equipartition, different elasticity-dependent "effective temperatures" and distinct non-Gaussian distributions. Our results shed light on puzzling observations in active gel experiments, and provide physical interpretation of existing observations, as well as predictions for future studies.
△ Less
Submitted 3 July, 2015;
originally announced July 2015.
