• Korean Journal of Cognitive Science
• /
• v.6 no.1
• /
• pp.5-30
• /
• 1995

Fodor and his followers argue that connectionist model does not have compositional structure,or else it is merely an implementation of a classicist model.'Classical constityency'is used importantly to show connectionist model does not have compositional structure.However,in contrast Fodor and his followers,I maintain that it is possible for connectionist model to have compositional structure without attributing classical constituency to cognitive model,and the very classical consitiuency makes the connectionist model nonimplementational.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.41 no.4
• /
• pp.331-338
• /
• 1992

The application of Connectionist expert system using neural network to fault diagnosis of power system is presented and compared with rule-based expert system. Also, the merits of Connectionist model using neural network is presented. In this paper, the neural network for fault diagnosis is hierarchically composed by 3 neural network classes. The whole power system is divided into subsystems, the neural networks (Class II) which take charge of each subsystem and the neural network (Class III) which connects subsystems are composed. Every section of power system is classified into one of the typical sections which can be applied with same diagnosis rules, as line-section, bus-section, transformer-section. For each typical section, only one neural network (Class I) is composed. As the proposed model has hierarchical structure, the great reduction of learning structure is achieved. With parallel distributed processing, we show the possibility of on-line fault diagnosis.

• The Journal of Society for e-Business Studies
• /
• v.7 no.1
• /
• pp.87-106
• /
• 2002

The behavioral and dynamic implications of an ERP implementation/installation are, to say the least, not well understood. Getting the switches set to enable the ERP software to go live is becoming straightforward. The really difficult part is understanding all of the dynamic interactions that accrue as a consequence. Dynamic causal and connectionist models are employed to facilitate an understanding of the dynamics and to enable control of the information-enhanced processes to take place. The connectionist model ran be analyzing (behind the scenes) the information accesses and transfers and coming If some conclusions about strong linkages that are getting established and what the behavioral implications of those new linkages and information accesses we. Ultimately, the connectionist model will come to an understanding of the dynamic, behavioral implications of the larger ERP implementation/installation per se. The underlying connectionist model will determine information transfers and workflow. Once a map of these two infrastructures is determined by the model, it becomes a relatively easy job for an analyst to suggest improvements in both. Connectionist models start with analog object structures and then use learning to produce mechanisms for managerial problem diagnoses. These mechanisms are neural models with multiple-layer structures that support continuous input/output. Based on earlier work performed and published by the author[10][11], a Connectionist ReasOning and LEarning System(CROLES) is developed that mimics the real-world reasoning infrastructure. Coupled with an explanation subsystem, this system can provide explanations as to why a particular reasoning structure behaved the way it did. Such a system operates in the backgmund, observing what is happening as every information access, every information response coming from each and every intelligent node (whether natural or artificial) operating within the ERP infrastructure is recorded and encoded. The CROLES is also able to transfer all workflows and map these onto the decision-making nodes of the organization.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1996.10a
• /
• pp.243-248
• /
• 1996

In this paper, a general neural-network-based connectionist model, called Fuzzy Neural Network(FNN), is proposed for the realization of a fuzzy logic control system. The proposed FNN is a feedforward multi-layered network which integrates the basic elements and functions of a traditional fuzzy logic controller into a connectionist structure which has distributed learning abilities. Such FNN can be constructed from training examples by learning rule, and the connectionist structure can be trained to develop fuzzy logic rules and find optimal input/output membership functions. Computer simulation examples will be presented to illustrate the performance and applicability of the proposed FNN, and their associated learning algorithms.

• Journal of Digital Contents Society
• /
• v.18 no.1
• /
• pp.93-99
• /
• 2017

A hybrid approach using Long Short Term Memory (LSTM) Recurrent Neural Network (RNN) has showed great improvement in speech recognition accuracy. For training acoustic model based on hybrid approach, it requires forced alignment of HMM state sequence from Gaussian Mixture Model (GMM)-Hidden Markov Model (HMM). However, high computation time for training GMM-HMM is required. This paper proposes an end-to-end approach for LSTM RNN-based Korean speech recognition to improve learning speed. A Connectionist Temporal Classification (CTC) algorithm is proposed to implement this approach. The proposed method showed almost equal performance in recognition rate, while the learning speed is 1.27 times faster.

• Korean Journal of Cognitive Science
• /
• v.28 no.2
• /
• pp.111-131
• /
• 2017

The connectionist model is one approach to studying language processing from a computational perspective. And building a representation in the connectionist model study is just as important as making the structure of the model in that it determines the level of learning and performance of the model. The connectionist model has been constructed in two different ways: localist representation and distributed representation. However, the localist representation used in the previous studies had limitations in that the unit of the output layer having a rare target activation value is inactivated, and the past distributed representation has the limitation of difficulty in confirming the result by the opacity of the displayed information. This has been a limitation of the overall connection model study. In this paper, we present a new method to induce distributed representation with local representation using abstraction of information, which is a feature of restricted Boltzmann machine, with respect to the limitation of such representation of the past. As a result, our proposed method effectively solves the problem of conventional representation by using the method of information compression and inverse transformation of distributed representation into local representation.

• Korean Journal of Cognitive Science
• /
• v.10 no.4
• /
• pp.71-83
• /
• 1999

Conventional symbolic inference systems lack flexibility because they do not well reflect flexible semantic structure of knowledge and use symbolic logic for their basic inference mechanism. For solving this problem. we have recently proposed the 'Connectionist Semantic Network(CSN)' as a model for flexible knowledge representation and inference based on neural networks. The CSN is capable of carrying out both approximate reasoning and commonsense reasoning based on similarity and association. However. we have difficulties in representing general and structured high-level knowledge and variable binding using the connectionist framework of the CSN. In this paper. we propose a hybrid system called SymCSN(Symbolic CSN) that combines a symbolic module for representing general and structured high-level knowledge and a connectionist module for representing and learning low-level semantic structure Simulation results show that the SymCSN is a plausible model for human-like flexible knowledge representation and inference.

• Korean Journal of Cognitive Science
• /
• v.2 no.2
• /
• pp.487-513
• /
• 1990

The present paper reviews experimental finding such as kanji stroop effect, kana superiority effect in naming task, kanji superiority effect in lexical devision task, and the different pattern of facilitatory priming effect in repetition priming task. Most of the experimental findings indicate that kana script and kanji script are processed independently and modularly. These indications are also consistent with the basic observations on Japanese dyslexics. A connectionist model named JIA(Japanese Interactive Activation)is proposed which is a revision of interactive activation model proposed by McClelland & Rumelhart(1981). The differences between the two models are as follows. Firstly, JIA has a kana module and kanji module at letter level. Secondly, JIA adopts script-specific interconnections between letter-level nodes and word-level nodes:word nodes receive larger activation from the script consistent letter-level nodes. JIA successfully explains all the experimental findings and many cases of Japanese dyslexia. A computer program which simulates JIA model was written and run.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.6
• /
• pp.1692-1705
• /
• 2024

In this paper, we present a method that integrates a Grammar Transducer as an external language model to enhance the accuracy of the pre-trained Korean End-to-end (E2E) Automatic Speech Recognition (ASR) model. The E2E ASR model utilizes the Connectionist Temporal Classification (CTC) loss function to derive hypothesis sentences from input audio. However, this method reveals a limitation inherent in the CTC approach, as it fails to capture language information from transcript data directly. To overcome this limitation, we propose a fusion approach that combines a clause-level n-gram language model, transformed into a Weighted Finite-State Transducer (WFST), with the E2E ASR model. This approach enhances the model's accuracy and allows for domain adaptation using just additional text data, avoiding the need for further intensive training of the extensive pre-trained ASR model. This is particularly advantageous for Korean, characterized as a low-resource language, which confronts a significant challenge due to limited resources of speech data and available ASR models. Initially, we validate the efficacy of training the n-gram model at the clause-level by contrasting its inference accuracy with that of the E2E ASR model when merged with language models trained on smaller lexical units. We then demonstrate that our approach achieves enhanced domain adaptation accuracy compared to Shallow Fusion, a previously devised method for merging an external language model with an E2E ASR model without necessitating additional training.

• The Korean Journal of Applied Statistics
• /
• v.32 no.2
• /
• pp.213-227
• /
• 2019

In this paper, we propose an end-to-end deep learning model combining Bayesian neural network with Korean speech recognition. In the past, Korean speech recognition was a complicated task due to the excessive parameters of many intermediate steps and needs for Korean expertise knowledge. Fortunately, Korean speech recognition becomes manageable with the aid of recent breakthroughs in "End-to-end" model. The end-to-end model decodes mel-frequency cepstral coefficients directly as text without any intermediate processes. Especially, Connectionist Temporal Classification loss and Attention based model are a kind of the end-to-end. In addition, we combine Bayesian neural network to implement the end-to-end model and obtain Monte Carlo estimates. Finally, we carry out our experiments on the "WorimalSam" online dictionary dataset. We obtain 4.58% Word Error Rate showing improved results compared to Google and Naver API.