• ETRI Journal
• /
• v.36 no.5
• /
• pp.714-720
• /
• 2014

In this paper, we propose the use of data-driven probabilistic utterance-level decision logic to improve Weighted Finite State Transducer (WFST)-based endpoint detection. In general, endpoint detection is dealt with using two cascaded decision processes. The first process is frame-level speech/non-speech classification based on statistical hypothesis testing, and the second process is a heuristic-knowledge-based utterance-level speech boundary decision. To handle these two processes within a unified framework, we propose a WFST-based approach. However, a WFST-based approach has the same limitations as conventional approaches in that the utterance-level decision is based on heuristic knowledge and the decision parameters are tuned sequentially. Therefore, to obtain decision knowledge from a speech corpus and optimize the parameters at the same time, we propose the use of data-driven probabilistic utterance-level decision logic. The proposed method reduces the average detection failure rate by about 14% for various noisy-speech corpora collected for an endpoint detection evaluation.

• Proceedings of the KIEE Conference
• /
• summer
• /
• pp.645-647
• /
• 2004

According to NERC definition, Available Transfer Capability (ATC) is a measure of the transfer capability remaining in the physical transmission network for the future commercial activity To calculate ATC, accurate and defensible TTC, CBM and TRM should be calculated in advance. This paper proposes a method to quantify time varying ATC based on probabilistic approach. The uncertainties of power system and market are considered as complex random variables. TRM with the desired probabilistic margin is calculated based on PLF analysis, and CBM is evaluated using LOLE of the system. Suggested ATC quantification method is verified using IEEE RTS with 72 bus. The proposed method shows efficiency and flexibility for the quantification of ATC.

• Proceedings of the Korean Information Science Society Conference
• /
• 2012.06b
• /
• pp.426-428
• /
• 2012

In order to identify proteins that are present in biological samples, these samples are separated and analyzed under the sequential procedure as follows: protein purification and digestion, peptide fragmentation by tandem mass spectrometry (MS/MS) which breaks peptides into fragments, peptide identification, and protein identification. One of the widely used methods for protein identification is based on probabilistic approaches such as ProteinProphet and BaysPro. However, they do not consider the difference in peptide identification probabilities according to their length. Here, we propose a probabilistic graphical model-based approach to protein identification from MS/MS data considering peptide identification probabilities, number of sibling peptides, and peptide length. We compared our approach with ProteinProphet using a yeast MS/MS dataset. As a result, our model identified 27 more proteins than ProteinProphet at 1% of FDR (false discovery rate), confirming the importance of peptide length information in protein identification.

• International Journal of Railway
• /
• v.6 no.2
• /
• pp.45-52
• /
• 2013

Railway construction needs vast soil investigation for its infrastructure foundation designs along the planned railway path to identify the design parameters for stability and serviceability checks. The soil investigation data are usually classified and grouped to decide design input parameters per each construction section and budget estimates. Deterministic design method which most civil engineer and practitioner are familiar with has a clear limitation in construction/maintenance budget control, and occasionally produced overdesigned or unsafe design problems. Instead of using a batch type analysis with predetermined input parameters, data population collected from site soil investigation and design load condition can be statistically estimated for the mean and variance to present the feature of data distribution and optimized with a best fitting probability function. Probabilistic approach using entire feature of design input data enables to predict the worst, best and most probable cases based on identified ranges of soil and load data, which will help railway designer select construction method to save the time and cost. This paper introduces two Monte Carlo simulations actually applied on estimation of retaining wall external stability and long term settlement of organic soil in soil investigation area for a recent high speed railway project.

• Proceedings of the KSME Conference
• /
• 2004.04a
• /
• pp.54-59
• /
• 2004

In this paper probabilistic fracture mechanics(PFM) approach is employed to evaluate the integrity of CANDU Zr-2.5Nb pressure tubes. Modified failure assessment diagram(Jr-FAD), plastic collapse, and critical crack length(CCL) approach are used for evaluating failure probability of the tubes. Jr-FAD was extended from the Kr-FAD because fracture of pressure tubes occurs in brittle manner due to hydrogen embrittlement of material by deuterium fluence. For developing the probabilistic integrity evaluation module, AECL procedures and fracture toughness parameters of EPRI were used.

• Smart Structures and Systems
• /
• v.23 no.2
• /
• pp.185-194
• /
• 2019

The probabilistic tsunami risk assessment of large coastal areas is challenging because the inland propagation of a tsunami wave requires an accurate numerical model that takes into account the interaction between the ground, the infrastructures, and the wave itself. Classic mesh-based methods face many challenges in the propagation of a tsunami wave inland due to their ever-moving boundary conditions. In alternative, mesh-less based methods can be used, but they require too much computational power in the far-field. This study proposes a hybrid approach. A mesh-based method propagates the tsunami wave from the far-field to the near-field, where the influence of the sea floor is negligible, and a mesh-less based method, smooth particle hydrodynamic, propagates the wave onto the coast and inland, and takes into account the wave structure interaction. Nowadays, this can be done because the advent of general purpose GPUs made mesh-less methods computationally affordable. The method is used to simulate the inland propagation of the 2004 Indian Ocean tsunami off the coast of Indonesia.

• International conference on construction engineering and project management
• /
• 2013.01a
• /
• pp.488-493
• /
• 2013

Accurate initial cost estimates are essential to effective management of construction projects where many decisions are made in the course of project management by referencing the estimates. In practice, the initial estimates are frequently derived from historical actual cost data, for which standard distribution-based techniques are widely applied in the construction industry to account for risk associated with the estimates. This approach assumes the same probability distribution of estimate errors for any selected estimates. This assumption, however, is not always satisfied. In order to account for the probabilistic nature of estimate errors, an alternative method for measuring the risk associated with a selected initial estimate is developed by applying the Bayesian probability approach. An application example include demonstrates how the method is implemented. A hypothesis test is conducted to reveal the robustness of the Bayesian probability model. The method is envisioned to effectively complement cost estimating methods that are currently in use by providing benefits as follows: (1) it effectively accounts for the probabilistic nature of errors in estimates; (2) it is easy to implement by using historical estimates and actual costs that are readily available in most construction companies; and (3) it minimizes subjective judgment by using quantitative data only.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2006.05b
• /
• pp.181-184
• /
• 2006

Traditional durability analysis is not possible to provide a controlled durability and long-term performance of concrete structures. Recently, research works have shown that probabilistic approach based on the theory of structural reliability, would be very valuable for durability analysis. In this study, the probabilistic durability analysis based on a Monte Carlo Simulation was carried out using sample data selected from detailed field investigation. The probabilistic properties of some design variables, such as diffusion coefficients of concrete and surface chloride concentration, were newly determined using some experimental data. By applying a probabilistic durability analysis to an integral structural design, the durability performance of concrete structures would be remarkably improved.

• Journal of Mechanical Science and Technology
• /
• v.19 no.spc1
• /
• pp.350-356
• /
• 2005

A method of dynamic analysis of mechanical systems considering probabilistic properties is proposed in this paper. Probabilistic properties that result from manufacturing tolerances can be represented by means and standard deviations (or variances). The probabilistic characteristics of dynamic responses of constrained multi-body systems are obtained by two ways : the proposed analytical approach and the Monte Carlo simulation. The formerpaper, necessitates sensitivity information to calculate the standard deviations. In this a direct differentiation method is employed to find the sensitivities of constrained multi-body systems. To verify the accuracy of the proposed method, numerical examples are solved and the results obtained by using the proposed method are compared to those obtained by Monte Carlo simulation.

• Structural Engineering and Mechanics
• /
• v.11 no.4
• /
• pp.373-392
• /
• 2001

The main idea behind the paper is to present two alternative methods of homogenization of the heat conduction problem in composite materials, where the heat conductivity coefficients are assumed to be random variables. These two methods are the Monte-Carlo simulation (MCS) technique and the second order perturbation second probabilistic moment method, with its computational implementation known as the Stochastic Finite Element Method (SFEM). From the mathematical point of view, the deterministic homogenization method, being extended to probabilistic spaces, is based on the effective modules approach. Numerical results obtained in the paper allow to compare MCS against the SFEM and, on the other hand, to verify the sensitivity of effective heat conductivity probabilistic moments to the reinforcement ratio. These computational studies are provided in the range of up to fourth order probabilistic moments of effective conductivity coefficient and compared with probabilistic characteristics of the Voigt-Reuss bounds.