Papers by Young-tack Park
Blackboard Scheduler Control Knowledge for Recursive Heuristic Classification
Dynamic and explicit ordering of strategies is a key process in modeling knowledge-level problem-... more Dynamic and explicit ordering of strategies is a key process in modeling knowledge-level problem-solving behavior. This paper addressed the important problem of howl to make the scheduler more knowledge-intensive in a way that facilitates the acquisition, integration, and maintenance of the scheduler control knowledge. The solution a, pp.oach described in this paper involved formulating the scheduler task as a heuristic classification problem, and then implementing it as a classification expert system. By doing this, the wide spectrum of known methods of acquiring, refining, and maintaining the knowledge of a classification expert system are a, pp.icable to the scheduler control knowledge. One important innovation of this research is that of recursive heuristic classification : this paper demonstrates that it is possible to formulate and solve a key subcomponent of heuristic classification as heuristic classification problem. Another key innovation is the creation of a method of dynamic heuristic classification : the classification alternatives that are selected among are dynamically generated in real-time and then evidence is gathered for and aginst these alternatives. In contrast, the normal model of heuristic classification is that of structured selection between a set of preenumerated fixed alternatives.
Blackboard scheduler control knowledge for heuristic classification: Representation and inference
The scheduler is an key component of a blackboard system architecture. This thesis addressed the ... more The scheduler is an key component of a blackboard system architecture. This thesis addressed the important problem of how to make the blackboard scheduler more knowledge intensive in a way that facilitates the acquisition, integration, and maintenance of the blackboard scheduler knowledge. The solution approach described in this thesis involved formulating the blackboard scheduler task as a heuristic classification problem, and then implementing it as a classification expert system. By doing this, the wide spectrum of known methods of acquiring, refining, and maintaining the knowledge of a classification expert system are applicable to the blackboard scheduler knowledge.In this thesis, the MINERVA expert system shell was extended by the addition of a blackboard scheduler level. The problem solving cycle involves a deliberation phase, wherein all the heuristic classification strategies that are applicable are collected. This is followed by a scheduling phase wherein the classification expert system for scheduling automatically gathers evidence for and against each of the applicable strategic actions, thereby ranking them according to desirability. Finally, there is an action phase that executes the most highly ranked strategic task.One important innovation of this research is that of recursive heuristic classification: this thesis demonstrates that it is possible to formulate and solve a key subcomponent of heuristic classification as a heuristic classification problem. Another key innovation is the creation of a method of dynamic heuristic classification: the classification alternatives that are selected among are dynamically generated in real-time and then evidence is gathered for and against these alternatives. In contrast, the normal model of heuristic classification is that of structured selection between a set of preenumerated fixed alternatives.U of I OnlyETDs are only available to UIUC Users without author permissio
The Implementation of the Personalized Emotional Character Agent
The Kips Transactions:partb, 2001
Recently, character agents are used as a user-friendly interface. In this paper, we have studied ... more Recently, character agents are used as a user-friendly interface. In this paper, we have studied a generic framework for emotional character agents which are designed to infer emotions from diverse personalities, situations, user behaviors and to express them. The method of emotion inference is based on blackboard systems which are used to solve the problems in AI. Because it keeps independence between knowledge sources which are rules of emotions, a blackboard-based inference engine is easy to manage knowledge sources, Blackboard-based systems gave the system flexibility. So we can adapt the engine to various application systems. Each emotional agent monitors user behavior, learns user profile and infers user behavior. And it generates characters emotions according to the user profile. So, in case of same situations, the agent can generate different emotions according to users. We have studied to build an personalized emotional character agent which according to situations and user...
A path-based relation networks model for knowledge graph completion
Expert Systems with Applications, 2021
Abstract We consider the problem of learning and inference in a large-scale knowledge graph conta... more Abstract We consider the problem of learning and inference in a large-scale knowledge graph containing incomplete knowledge. We show that a simple neural network module for relational reasoning through the path extracted from the knowledge base can be used to reliably infer new facts for the missing link. In our work, we used path ranking algorithm to extract the relation path from knowledge graph and use it to build train data. In order to learn the characteristics of relation, a detour path between nodes was created as training data using the extracted relation path. Using this, we trained a model that can predict whether a given triple (Head entity, relation, tail entity) is valid or not. Experiments show that our model obtains better link prediction, relation prediction and triple classification results than previous state-of-the-art models on benchmark datasets WN18RR, FB15k-237, WN11 and FB13.
Solving Non-deterministic Problem of Ontology Reasoning and Identifying Causes of Inconsistent Ontology using Negated Assumption-based Truth Maintenance System
Journal of KIISE:Software and Applications, 2009
In order to derive hidden information (concept subsumption, concept satisfiability and realizatio... more In order to derive hidden information (concept subsumption, concept satisfiability and realization) of OWL ontology, a number of OWL reasoners have been introduced. The most of these ontology reasoners were implemented using the tableau algorithm. However most reasoners simply report this information without providing a justification for any arbitrary entailment and unsatisfiable concept derived from OWL ontologies. The purpose of this paper is to investigate an optimized method for non-deterministic rule of the tableau algorithm and finding axioms to cause inconsistency in ontology. In this paper, therefore, we propose an optimized method for non-deterministic rule and finding axiom to cause inconsistency using NATMS. In the first place, we introduce Dependency Directed Backtracking to deal non-deterministic rule, a tableau-based decision procedure to find unsatisfiable axiom Furthermore we propose an improved method adapting NATMS.
Methods to Reduce Execution Time of Ontology Reasoners based on Tableaux Algorithm
Journal of KIISE:Software and Applications, 2009
As size of ontology has been increased more and more, the descriptions in the ontologies become m... more As size of ontology has been increased more and more, the descriptions in the ontologies become more complicated, Therefore finding and modifying unsatisfiable concepts is hard work in ontology construction process, Minerva is an ontology reasoner which detects unsatisfiable concepts automatically and infers subsumption relation between concepts in ontology, Most description logic based ontology reasoners (including Minerva) work using tableaux algorithm, Because tableaux algorithm is very costly, ontology reasoners need various optimization methods, In this paper, we propose optimizing methods to reduce execution time of tableaux algorithm based ontology reasoner. Proposed methods were applied to Minerva which was developed as preceding study result. In consequence the new version Minerva shows high performance.
Design and implementation of SCADA system to support scalability and openness
Journal of Institute of Control Robotics and Systems, 1999
The existing SCADA(Supervisory Control and Data Acquisition) system software is usually developed... more The existing SCADA(Supervisory Control and Data Acquisition) system software is usually developed to suitable for the specific hardware platforms. However, as per rapid improvement of computer performance and development of network technology, it is required to support scalability and inter-operability in existing different SCADA systems. In order to meet such requirements, in this paper, we propose a new type of SCADA testbed using Java for electric distribution applications. The system consists of three modules; development support tools, client and server modules. The basic architecture of the proposed SCADA system is similar to existing one, however, we improve the function of MTU and MMI interface to facilitate LAN and WAN environment. Also, the proposed system can deals with alarm and history data by using heterogeneous DBMS. Since the system is built in Java environment, the development cost is cheap and it can support sacalability and portability. Our experience can be utili...
A Dynamic Service Supporting Model for Semantic Web-based Situation Awareness Service
Journal of KIISE:Software and Applications, 2009
The technology of Semantic Web realizes the base technology for context-awareness that creates ne... more The technology of Semantic Web realizes the base technology for context-awareness that creates new services by dynamically and flexibly combining various resources (people, concepts, etc). According to the realization of ubiquitous computing technology, many researchers are currently working for the embodiment of web service. However, most studies of them bring about the only predefined results those are limited to the initial description by service designer. In this paper, we propose a new service supporting model to provide an automatic method for plan related tasks which achieve goal state from initial state. The inputs on an planner are intial and goal descriptions which are mapped to the current situation and to the user request respectively. The idea of the method is to infer context from world model by DL-based ontology reasoning using OWL domain ontology. The context guide services to be loaded into planner. Then, the planner searches and plans at least one service to satisf...
Medusa: An Extended DL-Reasoner for SWRL-enabled Ontologies
Journal of KIISE:Software and Applications, 2009
In order to derive hidden Information (concept subsumption, concept satisfiability and realizatio... more In order to derive hidden Information (concept subsumption, concept satisfiability and realization) of OWL ontologies, a number of OWL reasoners have been introduced. Most of the reasoners were implemented to be based on tableau algorithm. However this approach has certain limitation. This paper presents architecture for Medusa. The Medusa is an extended DL-reasoner for SWRL(Semantic Web Rule Language) reasoning under well-founded semantics with ontologies specified in Description Logic. Description logic based ontology reasoners theoretically explore knowledge representation and its reasoning in concept languages. However these logics are not equipped with rule-based reasoning mechanisms for assertional knowledge base; specifically, rule and facts in logic programming, or interaction of rules and facts with terminology. In order to deal with the enriched reasoning, The Medusa provides combining DL-knowledge base and rule based reasoner. The described prototype uses API[1] for contr...
A Performance Analysis of Large ABox Reasoning in OWL-DL Reasoners
Journal of KIISE:Software and Applications, 2007
Reasoners using typical Tableaux algorithm such as RacerPro, Pellet have a problem in Tableaux al... more Reasoners using typical Tableaux algorithm such as RacerPro, Pellet have a problem in Tableaux algorithm large ABox reasoning. Researches to solve these Problems are dealt with Instance Store of University of Manchester which uses Tableaux algorithm based reasoner and DBMS and KAON2 of University of Karlsruhe using Disjunctive Datalog approach. An evaluation experiment for present reasoners is the experiment of TBox reasoning in most of Tableaux algorithm based one. The most of benchmarking tests in reasoning systems haven't done with ABox reasoning based Tableaux Algorithm but done with TBox reasoning based Tableaux Algorithm. Especially, rarely reported benchmarking tests in reasoners have been issued nowadays. Therefore, this thesis evaluates systems with theory of each reasoners for large ABox reasoning that becomes issues recently with typical reasoners. The large AoBx reasoning engine will be analyzed using Instance Store and KAON2 of Manchester University for large ABox p...
Knowledge Completion System using Neuro-Symbolic-based Rule Induction and Inference Engine
Journal of KIISE, 2021
Journal of KIISE, 2018
Vehicular Ad-hoc NETworks (VANETs) are designed to improve transportation efficiency such as to i... more Vehicular Ad-hoc NETworks (VANETs) are designed to improve transportation efficiency such as to increase safety and reduce traffic accidents. In addition, VANET is created to connect and exchange information between vehicles or between vehicle and infrastructure. For VANET, Medium Access Control (MAC) protocols, which provide an efficient broadcast service, are designed to efficiently and fairly share the wireless medium between vehicles and providers. Recently, the hybrid MAC protocol was designed to combine TDMA-and CSMA-based mechanisms into a single mechanism to improve the Quality of Service (QoS) and decrease the collision rate. In this paper, we propose an Efficient and Adaptable Hybrid Multi-channel Multi-hop MAC protocol in VANETs, called the EAHMAC protocol, which allows vehicles to not only occupy time slots but also to broadcast packets in a flexible way based on the two-hop neighbor's information. The simulation results show that our proposal outperforms the existing protocols in terms of access collision rate, packet delivery ratio, and throughput on the service channel.
Journal of KIISE, 2018
Low-Rank Representation (LRR) based methods are widely used in many practical applications, such ... more Low-Rank Representation (LRR) based methods are widely used in many practical applications, such as face clustering and object detection, because they can guarantee high prediction accuracy when used to constructing graphs in graph-based semi-supervised learning. However, in order to solve the LRR problem, it is necessary to perform singular value decomposition on the square matrix of the number of data points for each iteration of the algorithm; hence the calculation is inefficient. To solve this problem, we propose an improved and faster LRR method based on the recently published Fast LRR (FaLRR) and suggests ways to introduce and optimize additional constraints on the underlying optimization goals in order to address the fact that the FaLRR is fast but actually poor in classification problems. Our experiments confirm that the proposed method finds a better solution than LRR does. We also propose Fast MLRR (FaMLRR), which shows better results when the goal of minimizing is added.
Confidence Value based Large Scale OWL Horst Ontology Reasoning
Journal of KIISE, 2016
Robust route inference and representation for uncertain sensor data
Computers & Electrical Engineering, 2016
Display Omitted We propose DBN based route models and prediction of users' routes using robus... more Display Omitted We propose DBN based route models and prediction of users' routes using robust particle filtering.We use a DBN to infer the next locations or destinations based on the observed spatio-temporal data.The robust particle filter handles uncertainty and constraints to enhance accuracy and efficiency. This paper proposes a robust particle filter to deal with incomplete sensor data to predict the user's routes and represents users' movements using a dynamic Bayesian network model that patterns the user's spatiotemporal routine. The proposed particle filter includes robust particle generation to supplement any incorrect and incomplete sensor information, efficient switching/weight functions to reduce computation complexity while considering uncertainty, and resampling to enhance the accuracy of the particles by solving the degeneracy problem. The robust particle filter enhances the accuracy and efficiency with which a user's routes and destinations are determined.
The KIPS Transactions:PartB, 2009
Author name disambiguation is essential for improving performance of document indexing, retrieval... more Author name disambiguation is essential for improving performance of document indexing, retrieval, and web search. Author name disambiguation resolves the conflict when multiple authors share the same name label. This paper introduces a novel approach which exploits ontologies and WordNet-based category utility for author name disambiguation. Our method utilizes author knowledge in the form of populated ontology that uses various types of properties: titles, abstracts and co-authors of papers and authors' affiliation. Author ontology has been constructed in the artificial intelligence and semantic web areas semi-automatically using OWL API and heuristics. Author name disambiguation determines the correct author from various candidate authors in the populated author ontology. Candidate authors are evaluated using proposed WordNet-based category utility to resolve disambiguation. Category utility is a tradeoff between intra-class similarity and inter-class dissimilarity of author instances, where author instances are described in terms of attribute-value pairs. WordNet-based category utility has been proposed to exploit concept information in WordNet for semantic analysis for disambiguation. Experiments using the WordNet-based category utility increase the number of disambiguation by about 10% compared with that of category utility, and increase the overall amount of accuracy by around 98%.
Human Facial Expression Recognition Using Stepwise Linear Discriminant Analysis and Hidden Conditional Random Fields
IEEE Transactions on Image Processing, 2015
This paper introduces an accurate and robust facial expression recognition (FER) system. For feat... more This paper introduces an accurate and robust facial expression recognition (FER) system. For feature extraction, the proposed FER system employs stepwise linear discriminant analysis (SWLDA). SWLDA focuses on selecting the localized features from the expression frames using the partial F-test values, thereby reducing the within class variance and increasing the low between variance among different expression classes. For recognition, the hidden conditional random fields (HCRFs) model is utilized. HCRF is capable of approximating a complex distribution using a mixture of Gaussian density functions. To achieve optimum results, the system employs a hierarchical recognition strategy. Under these settings, expressions are divided into three categories based on parts of the face that contribute most toward an expression. During recognition, at the first level, SWLDA and HCRF are employed to recognize the expression category; whereas, at the second level, the label for the expression within the recognized category is determined using a separate set of SWLDA and HCRF, trained just for that category. In order to validate the system, four publicly available data sets were used, and a total of four experiments were performed. The weighted average recognition rate for the proposed FER approach was 96.37% across the four different data sets, which is a significant improvement in contrast to the existing FER methods.
Applied Intelligence, 2011
In this paper, a novel feature selection method based on the normalization of the well-known mutu... more In this paper, a novel feature selection method based on the normalization of the well-known mutual information measurement is presented. Our method is derived from an existing approach, the max-relevance and minredundancy (mRMR) approach. We, however, propose to normalize the mutual information used in the method so that the domination of the relevance or of the redundancy can be eliminated. We borrow some commonly used recognition models including Support Vector Machine (SVM), k-Nearest-Neighbor (kNN), and Linear Discriminant Analysis (LDA) to compare our algorithm with the original (mRMR) and a recently improved version of the mRMR, the Normalized Mutual Information Feature Selection (NMIFS) algorithm. To avoid data-specific statements, we conduct our classification experiments using various datasets from the UCI machine learning repository. The results confirm that our feature selection method is more robust than the others with regard to classification accuracy.
Journal of KIISE, 2019
Sports broadcasting requires understanding and reasoning of a current situation based on informat... more Sports broadcasting requires understanding and reasoning of a current situation based on information regarding sports scenes, players, and past knowledge. In this paper, we introduced how scene classifier, player detector, motion recognizer could be used to obtain information on sports images and understand current situations. We created three types of commentaries. One was from web data, another was from 13 scenes with scene classifier, and the other was generated by the position of the players, eight motions, and the ontology. Data from the KBO (Korea Baseball Organization League) games from April 1, 2018, to April 14, 2018, were directly labeled to learn the model.
순환신경망 기반의 사용자 의도 예측 모델
정보과학회논문지 (Journal of KIISE), Apr 1, 2018
Uploads
Papers by Young-tack Park