-
Rail break and derailment prediction using Probabilistic Graphical Modelling
Authors:
Rebecca M. C. Taylor,
Johan A. du Preez
Abstract:
Rail breaks are one of the most common causes of derailments internationally. This is no different for the South African Iron Ore line. Many rail breaks occur as a heavy-haul train passes over a crack, large defect or defective weld. In such cases, it is usually too late for the train to slow down in time to prevent a de-railment. Knowing the risk of a rail break occurring associated with a train…
▽ More
Rail breaks are one of the most common causes of derailments internationally. This is no different for the South African Iron Ore line. Many rail breaks occur as a heavy-haul train passes over a crack, large defect or defective weld. In such cases, it is usually too late for the train to slow down in time to prevent a de-railment. Knowing the risk of a rail break occurring associated with a train passing over a section of rail allows for better implementation of maintenance initiatives and mitigating measures. In this paper the Ore Line's specific challenges are discussed and the currently available data that can be used to create a rail break risk prediction model is reviewed. The development of a basic rail break risk prediction model for the Ore Line is then presented. Finally the insight gained from the model is demonstrated by means of discussing various scenarios of various rail break risk. In future work, we are planning on extending this basic model to allow input from live monitoring systems such as the ultrasonic broken rail detection system.
△ Less
Submitted 25 August, 2022;
originally announced August 2022.
-
SimLDA: A tool for topic model evaluation
Authors:
Rebecca M. C. Taylor,
Johan A. du Preez
Abstract:
Variational Bayes (VB) applied to latent Dirichlet allocation (LDA) has become the most popular algorithm for aspect modeling. While sufficiently successful in text topic extraction from large corpora, VB is less successful in identifying aspects in the presence of limited data. We present a novel variational message passing algorithm as applied to Latent Dirichlet Allocation (LDA) and compare it…
▽ More
Variational Bayes (VB) applied to latent Dirichlet allocation (LDA) has become the most popular algorithm for aspect modeling. While sufficiently successful in text topic extraction from large corpora, VB is less successful in identifying aspects in the presence of limited data. We present a novel variational message passing algorithm as applied to Latent Dirichlet Allocation (LDA) and compare it with the gold standard VB and collapsed Gibbs sampling. In situations where marginalisation leads to non-conjugate messages, we use ideas from sampling to derive approximate update equations. In cases where conjugacy holds, Loopy Belief update (LBU) (also known as Lauritzen-Spiegelhalter) is used. Our algorithm, ALBU (approximate LBU), has strong similarities with Variational Message Passing (VMP) (which is the message passing variant of VB). To compare the performance of the algorithms in the presence of limited data, we use data sets consisting of tweets and news groups. Using coherence measures we show that ALBU learns latent distributions more accurately than does VB, especially for smaller data sets.
△ Less
Submitted 19 August, 2022;
originally announced August 2022.
-
Constant Curvature Curve Tube Codes for Low-Latency Analog Error Correction
Authors:
Anders M. Buvarp,
Robert M. Taylor Jr.,
Kumar Vijay Mishra,
Lamine M. Mili,
Amir I. Zaghloul
Abstract:
Recent research in ultra-reliable and low latency communications (URLLC) for future wireless systems has spurred interest in short block-length codes. In this context, we analyze arbitrary harmonic bandwidth (BW) expansions for a class of high-dimension constant curvature curve codes for analog error correction of independent continuous-alphabet uniform sources. In particular, we employ the circum…
▽ More
Recent research in ultra-reliable and low latency communications (URLLC) for future wireless systems has spurred interest in short block-length codes. In this context, we analyze arbitrary harmonic bandwidth (BW) expansions for a class of high-dimension constant curvature curve codes for analog error correction of independent continuous-alphabet uniform sources. In particular, we employ the circumradius function from knot theory to prescribe insulating tubes about the centerline of constant curvature curves. We then use tube packing density within a hypersphere to optimize the curve parameters. The resulting constant curvature curve tube (C3T) codes possess the smallest possible latency, i.e., block-length is unity under BW expansion map**. Further, the codes perform within $5$ dB signal-to-distortion ratio of the optimal performance theoretically achievable at a signal-to-noise ratio (SNR) $< -5$ dB for BW expansion factor $n \leq 10$. Furthermore, we propose a neural-network-based method to decode C3T codes. We show that, at low SNR, the neural-network-based C3T decoder outperforms the maximum likelihood and minimum mean-squared error decoders for all $n$. The best possible digital codes require two to three orders of magnitude higher latency compared to C3T codes, thereby demonstrating the latter's utility for URLLC.
△ Less
Submitted 2 August, 2023; v1 submitted 24 May, 2022;
originally announced May 2022.
-
A derivation of variational message passing (VMP) for latent Dirichlet allocation (LDA)
Authors:
Rebecca M. C. Taylor,
Dirko Coetsee,
Johan A. du Preez
Abstract:
Latent Dirichlet Allocation (LDA) is a probabilistic model used to uncover latent topics in a corpus of documents. Inference is often performed using variational Bayes (VB) algorithms, which calculate a lower bound to the posterior distribution over the parameters. Deriving the variational update equations for new models requires considerable manual effort; variational message passing (VMP) has em…
▽ More
Latent Dirichlet Allocation (LDA) is a probabilistic model used to uncover latent topics in a corpus of documents. Inference is often performed using variational Bayes (VB) algorithms, which calculate a lower bound to the posterior distribution over the parameters. Deriving the variational update equations for new models requires considerable manual effort; variational message passing (VMP) has emerged as a "black-box" tool to expedite the process of variational inference. But applying VMP in practice still presents subtle challenges, and the existing literature does not contain the steps that are necessary to implement VMP for the standard smoothed LDA model, nor are available black-box probabilistic graphical modelling software able to do the word-topic updates necessary to implement LDA. In this paper, we therefore present a detailed derivation of the VMP update equations for LDA. We see this as a first step to enabling other researchers to calculate the VMP updates for similar graphical models.
△ Less
Submitted 25 August, 2022; v1 submitted 2 November, 2021;
originally announced November 2021.
-
ALBU: An approximate Loopy Belief message passing algorithm for LDA to improve performance on small data sets
Authors:
Rebecca M. C. Taylor,
Johan A. du Preez
Abstract:
Variational Bayes (VB) applied to latent Dirichlet allocation (LDA) has become the most popular algorithm for aspect modeling. While sufficiently successful in text topic extraction from large corpora, VB is less successful in identifying aspects in the presence of limited data. We present a novel variational message passing algorithm as applied to Latent Dirichlet Allocation (LDA) and compare it…
▽ More
Variational Bayes (VB) applied to latent Dirichlet allocation (LDA) has become the most popular algorithm for aspect modeling. While sufficiently successful in text topic extraction from large corpora, VB is less successful in identifying aspects in the presence of limited data. We present a novel variational message passing algorithm as applied to Latent Dirichlet Allocation (LDA) and compare it with the gold standard VB and collapsed Gibbs sampling. In situations where marginalisation leads to non-conjugate messages, we use ideas from sampling to derive approximate update equations. In cases where conjugacy holds, Loopy Belief update (LBU) (also known as Lauritzen-Spiegelhalter) is used. Our algorithm, ALBU (approximate LBU), has strong similarities with Variational Message Passing (VMP) (which is the message passing variant of VB). To compare the performance of the algorithms in the presence of limited data, we use data sets consisting of tweets and news groups. Additionally, to perform more fine grained evaluations and comparisons, we use simulations that enable comparisons with the ground truth via Kullback-Leibler divergence (KLD). Using coherence measures for the text corpora and KLD with the simulations we show that ALBU learns latent distributions more accurately than does VB, especially for smaller data sets.
△ Less
Submitted 19 August, 2022; v1 submitted 1 October, 2021;
originally announced October 2021.
-
SketchBio: A Scientist's 3D Interface for Molecular Modeling and Animation
Authors:
Shawn M. Waldon,
Peter M. Thompson,
Patrick J. Hahn,
Russell M. Taylor II
Abstract:
Background: Because of the difficulties involved in learning and using 3D modeling and rendering software, many scientists hire programmers or animators to create models and animations. This both slows the discovery process and provides opportunities for miscommunication. Working with multiple collaborators, we developed a set of design goals for a tool that would enable them to directly construct…
▽ More
Background: Because of the difficulties involved in learning and using 3D modeling and rendering software, many scientists hire programmers or animators to create models and animations. This both slows the discovery process and provides opportunities for miscommunication. Working with multiple collaborators, we developed a set of design goals for a tool that would enable them to directly construct models and animations. Results: We present SketchBio, a tool that incorporates state-of-the-art bimanual interaction and drop shadows to enable rapid construction of molecular structures and animations. It includes three novel features: crystal by example, pose-mode physics, and spring-based layout that accelerate operations common in the formation of molecular models. We present design decisions and their consequences, including cases where iterative design was required to produce effective approaches. Conclusions: The design decisions, novel features, and inclusion of state-of-the-art techniques enabled SketchBio to meet all of its design goals. These features and decisions can be incorporated into existing and new tools to improve their effectiveness
△ Less
Submitted 11 July, 2014;
originally announced July 2014.
-
Probing the QCD Critical Point with Relativistic Heavy-Ion Collisions
Authors:
Steffen A. Bass,
Hannah Petersen,
Cory Quammen,
Hal Canary,
Christopher G. Healey,
Russell M. Taylor II
Abstract:
We utilize an event-by-event relativistic hydrodynamic calculation performed at a number of different incident beam energies to investigate the creation of hot and dense QCD matter near the critical point. Using state-of-the-art analysis and visualization tools we demonstrate that each collision event probes QCD matter characterized by a wide range of temperatures and baryo-chemical potentials, ma…
▽ More
We utilize an event-by-event relativistic hydrodynamic calculation performed at a number of different incident beam energies to investigate the creation of hot and dense QCD matter near the critical point. Using state-of-the-art analysis and visualization tools we demonstrate that each collision event probes QCD matter characterized by a wide range of temperatures and baryo-chemical potentials, making a dynamical response of the system to the vicinity of the critical point very difficult to isolate above the background.
△ Less
Submitted 31 January, 2012;
originally announced February 2012.
-
Observation of Nanometer-Scale Rolling Motion Mediated by Commensurate Contact
Authors:
M. R. Falvo,
J. Steele,
A. Buldum,
J. D. Schall,
R. M. Taylor II,
J. P. Lu,
D. W. Brenner,
R. Superfine
Abstract:
We report, through experimental observations and computer simulations, that atomic lattice interlocking can determine whether an object rolls or slides on a surface. We have quantitatively manipulated carbon nanotubes (CNTs) on a variety of substrates with an atomic force microscope (AFM) and observe rolling to occur only on graphite. We measure the forces when the CNT is in-registry with the gr…
▽ More
We report, through experimental observations and computer simulations, that atomic lattice interlocking can determine whether an object rolls or slides on a surface. We have quantitatively manipulated carbon nanotubes (CNTs) on a variety of substrates with an atomic force microscope (AFM) and observe rolling to occur only on graphite. We measure the forces when the CNT is in-registry with the graphite lattice, and observe rolling only in this lock-in state. Atomistic computer simulations identify the energy barriers for sliding and rolling, elucidate atomic-scale features of slip-roll motion, and explain the details of the lateral force data in terms of the intrinsic faceting of multiwall CNTs.
△ Less
Submitted 5 April, 2000;
originally announced April 2000.
-
In Situ Resistance Measurements of Strained Carbon Nanotubes
Authors:
S. Paulson,
M. R. Falvo,
N. Snider,
A. Helser,
T. Hudson,
A. Seeger,
R. M. Taylor II,
R. Superfine,
S. Washburn
Abstract:
We investigate the response of multi-walled carbon nanotubes to mechanical strain applied with an Atomic Force Microscope (AFM) probe. We find that in some samples, changes in the contact resistance dominate the measured resistance change. In others, strain large enough to fracture the tube can be applied without a significant change in the contact resistance. In this case we observe that enough…
▽ More
We investigate the response of multi-walled carbon nanotubes to mechanical strain applied with an Atomic Force Microscope (AFM) probe. We find that in some samples, changes in the contact resistance dominate the measured resistance change. In others, strain large enough to fracture the tube can be applied without a significant change in the contact resistance. In this case we observe that enough force is applied to break the tube without any change in resistance until the tube fails. We have also manipulated the ends of the broken tube back in contact with each other, re-establishing a finite resistance. We observe that in this broken configuration the resistance of the sample is tunable to values 15-350 kW greater than prior to breaking.
△ Less
Submitted 20 May, 1999;
originally announced May 1999.
