• The Korean Journal of Applied Statistics
• /
• v.34 no.6
• /
• pp.863-874
• /
• 2021

In this paper, we propose a post-selection inference procedure for the fused lasso signal approximator (FLSA). The FLSA finds underlying sparse piecewise constant mean structure by applying total variation (TV) semi-norm as a penalty term. However, it is widely known that this convex relaxation can cause asymptotic inconsistency in change points detection. As a result, there can remain false change points even though we try to find the best subset of change points via a tuning procedure. To remove these false change points, we propose a post-selection inference for the FLSA. The proposed procedure applies a permutation test based on CUSUM statistic. Our post-selection inference procedure is an extension of the permutation test of Antoch and Hušková (2001) which deals with single change point problems, to multiple change points detection problems in combination with the FLSA. Numerical study results show that the proposed procedure is better than naïve z-tests and tests based on the limiting distribution of CUSUM statistics.

• Phonetics and Speech Sciences
• /
• v.13 no.4
• /
• pp.47-53
• /
• 2021

Autoregressive Text-to-Speech (TTS) models suffer from inference instability and slow inference speed. Inference instability occurs when a poorly predicted sample at time step t affects all the subsequent predictions. Slow inference speed arises from a model structure that forces the predicted samples from time steps 1 to t-1 to predict the sample at time step t. In this study, an end-to-end non-autoregressive fast text-to-speech model is suggested as a solution to these problems. The results of this study show that this model's Mean Opinion Score (MOS) is close to that of Tacotron 2 - WaveNet, while this model's inference speed and stability are higher than those of Tacotron 2 - WaveNet. Further, this study aims to offer insight into the improvement of non-autoregressive models.

• Journal of the Korea Concrete Institute
• /
• v.17 no.1 s.85
• /
• pp.157-166
• /
• 2005

This article involves architecting prototype-fuzzy expert system for designing the nominal cover thickness by means of fuzzy inference for quantitatively representing the environment affecting factor to reinforced concrete in chloride-induced corrosion environment. In this work, nominal cover thickness to reinforcement in concrete was determined by the sum of minimum cover thickness and tolerance to that defined from skill level, constructability and the significance of member. Several variables defining the quality of concrete and environment affecting factor (EAF) including relative humidity, temperature, cyclic wet and dry, and the distance from coast were treated as fuzzy variables. To qualify EAF the environment conditions of cycle degree of wet-dry, relative humidity, distance from coast and temperature were used as input variables. To determine the nominal cover thickness a qualified EAF, concrete grade, and water-cement ratio were used. The membership functions of each fuzzy variable were generated from the engineering knowledge and intuition based on some references as well as some international codes of practice.

• ETRI Journal
• /
• v.41 no.6
• /
• pp.771-781
• /
• 2019

We propose the VS³-NET model to solve the task of question answering questions with machine-reading comprehension that searches for an appropriate answer in a given context. VS³-NET is a model that trains latent variables for each question using variational inferences based on a model of a simple recurrent unit-based sentences and self-matching networks. The types of questions vary, and the answers depend on the type of question. To perform efficient inference and learning, we introduce neural question-type models to approximate the prior and posterior distributions of the latent variables, and we use these approximated distributions to optimize a reparameterized variational lower bound. The context given in machine-reading comprehension usually comprises several sentences, leading to performance degradation caused by context length. Therefore, we model a hierarchical structure using sentence encoding, in which as the context becomes longer, the performance degrades. Experimental results show that the proposed VS³-NET model has an exact-match score of 76.8% and an F1 score of 84.5% on the SQuAD test set.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.15 no.2
• /
• pp.131-138
• /
• 2016

In order to improve the performance of an extended Kalman filter, a simplified indirect inference method (SIIM) fuzzy logic system (FLS) is proposed. The proposed FLS is composed of two fuzzy input variables, four fuzzy rules and one fuzzy output. Two normalized fuzzy input variables are the variance between the trace of a prior and a posterior covariance matrix, and the residual error of a Kalman algorithm. One fuzzy output variable is the weighting factor to adjust for the Kalman gain. There is no need to decide the number and the membership function of input variables, because we employ the normalized monotone increasing/decreasing function. The single parameter to be determined is the magnitude of a universe of discourse in the output variable. The structure of the proposed FLS is simple and easy to apply to various nonlinear state estimation problems. The simulation results show that the proposed FLS has strong adaptability to estimate the states of the incoming/outgoing moving objects, and outperforms the conventional extended Kalman filter algorithm by providing solutions that are more accurate.

• Proceedings of the KIEE Conference
• /
• 1999.07g
• /
• pp.2849-2852
• /
• 1999

This paper describes the fuzzy PD controller using simplified indirect inference method. First, the fuzzy PD controller is derived from the conventional continuous time linear PD controller. Then the fuzzification, control-rule base, and defuzzification using SIIM in the design of the fuzzy controller are discussed in detail. The resulting controller is a discrete time fuzzy version of the conventional PD controller. which has the same linear structure. but are nonlinear functions of the input signals. The proposed controller enhances the self-tuning control capability. particularly when the process to be controlled is nonlinear. As the SIIM is applied, the fuzzy Inference results can be calculated with splitting fuzzy variables into each action component and are determined as the functional form of corresponding variables. So the Proposed method has the capability of the high speed inference and extending the fuzzy input variables easily. Computer simulation results have demonstrated the superior to the control Performance of the one Proposed by D. Misir et at. Final)y. we simulated the contour control of the X-Y tables with direct control strategies using the proposed fuzzy PD controller.

• Proceedings of the KIEE Conference
• /
• 1996.07b
• /
• pp.1104-1106
• /
• 1996

In this paper, a hierarchical fuzzy controller is proposed for the stabilization control of the inverted pendulum system. The design of controller for that system is difficult because of its complicated nonlinear mathematical model with unknown parameters. Conventional fuzzy control strategy based only on dynamics of pendulum made have failed to stabilize. However, proposed control strategies are to swing pendulum from natural stable up equilibrium point to an unstable equilibrium point and are to transport a cart from an arbitrary position toward a center of rail. Thus, the proposed fuzzy stabilization controller have a hierarchical fuzzy inference structure; that is, the lower level is for inference interface for the virtual equilibrium point and the higher level one for the position control of cart according to the firstly inferred virtual equilibrium point.

• Journal of KIISE
• /
• v.41 no.9
• /
• pp.674-685
• /
• 2014

The Knowledge service system needs to infer a new knowledge from indicated knowledge to provide its effective service. Most of the Knowledge service system is expressed in terms of ontology. The volume of knowledge information in a real world is getting massive, so effective technique for massive data of ontology is drawing attention. This paper is to provide the method to infer massive data-ontology to the extent of RDFS, based on cloud computing environment, and evaluate its capability. RDFS inference suggested in this paper is focused on both the method applying MapReduce based on RDFS meta table, and the method of single use of cloud computing memory without using MapReduce under distributed file computing environment. Therefore, this paper explains basically the inference system structure of each technique, the meta table set-up according to RDFS inference rule, and the algorithm of inference strategy. In order to evaluate suggested method in this paper, we perform experiment with LUBM set which is formal data to evaluate ontology inference and search speed. In case LUBM6000, the RDFS inference technique based on meta table had required 13.75 minutes(inferring 1,042 triples per second) to conduct total inference, whereas the method applying the cloud computing memory had needed 7.24 minutes(inferring 1,979 triples per second) showing its speed twice faster.

• The Journal of Society for e-Business Studies
• /
• v.18 no.2
• /
• pp.153-173
• /
• 2013

The ontology has been gaining increasing interests by recent arise of the semantic web and related technologies. The focus is mostly on inference query processing that requires high-level techniques for storage and searching ontologies efficiently, and it has been actively studied in the area of semantic-based searching. W3C's recommendation is to use RDFS and OWL for representing ontologies. However memory-based editors, inference engines, and triple storages all store ontology as a simple set of triplets. Naturally the performance is limited, especially when a large-scale ontology needs to be processed. A variety of researches on proposing algorithms for efficient inference query processing has been conducted, and many of them are based on using proven relational database technology. However, none of them had been successful in obtaining the complete set of inference results which reflects the five characteristics of the ontology properties. In this paper, we propose a new index structure called hyper cube index to efficiently process inference queries. Our approach is based on an intuition that an index can speed up the query processing when extensive inferencing is required.

• Journal of Electrical Engineering and Technology
• /
• v.12 no.6
• /
• pp.2388-2398
• /
• 2017

In this study, we propose a robust face recognition system to pose variations based on automatic pose estimation. Radial basis function neural network is applied as one of the functional components of the overall face recognition system. The proposed system consists of preprocessing and recognition modules to provide a solution to pose variation and high-dimensional pattern recognition problems. In the preprocessing part, principal component analysis (PCA) and 2-dimensional 2-directional PCA ($(2D)^2$ PCA) are applied. These functional modules are useful in reducing dimensionality of the feature space. The proposed RBFNNs architecture consists of three functional modules such as condition, conclusion and inference phase realized in terms of fuzzy "if-then" rules. In the condition phase of fuzzy rules, the input space is partitioned with the use of fuzzy clustering realized by the Fuzzy C-Means (FCM) algorithm. In conclusion phase of rules, the connections (weights) are realized through four types of polynomials such as constant, linear, quadratic and modified quadratic. The coefficients of the RBFNNs model are obtained by fuzzy inference method constituting the inference phase of fuzzy rules. The essential design parameters (such as the number of nodes, and fuzzification coefficient) of the networks are optimized with the aid of Particle Swarm Optimization (PSO). Experimental results completed on standard face database -Honda/UCSD, Cambridge Head pose, and IC&CI databases demonstrate the effectiveness and efficiency of face recognition system compared with other studies.