• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.8
• /
• pp.371-380
• /
• 2023

Relation extraction is to extract relationships between named entities from text. Traditionally, relation extraction methods only extract relations between predetermined subject and object entities. However, in end-to-end relation extraction, all possible relations must be extracted by considering the positions of the subject and object for each pair of entities, and so this method uses time and resources inefficiently. To alleviate this problem, this paper proposes a method that sets directions based on the positions of the subject and object, and extracts relations according to the directions. The proposed method utilizes existing relation extraction data to generate direction labels indicating the direction in which the subject points to the object in the sentence, adds entity position tokens and entity type to sentences to predict the directions using a pre-trained language model (KLUE-RoBERTa-base, RoBERTa-base), and generates representations of subject and object entities through probabilistic crossover operation. Then, we make use of these representations to extract relations. Experimental results show that the proposed model performs about 3 ~ 4%p better than a method for predicting integrated labels. In addition, when learning Korean and English data using the proposed model, the performance was 1.7%p higher in English than in Korean due to the number of data and language disorder and the values of the parameters that produce the best performance were different. By excluding the number of directional cases, the proposed model can reduce the waste of resources in end-to-end relation extraction.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.10a
• /
• pp.420-426
• /
• 1998

Evolutionary algorithms are probabilistic optimization algorithms based on the model of natural evolution. Recently the efforts to improve the performance of evolutionary algorithms have been made extensively. In this paper, we introduce the research for improving the convergence rate and search faculty of evolution algorithms by using reinforcement learning. After providing an introduction to evolution algorithms and reinforcement learning, we present adaptive genetic algorithms, reinforcement genetic programming, and reinforcement evolution strategies which are combined with reinforcement learning. Adaptive genetic algorithms generate mutation probabilities of each locus by interacting with the environment according to reinforcement learning. Reinforcement genetic programming executes crossover and mutation operations based on reinforcement and inhibition mechanism of reinforcement learning. Reinforcement evolution strategies use the variances of fitness occurred by mutation to make the reinforcement signals which estimate and control the step length.

• Proceedings of the Korea Society for Simulation Conference
• /
• 1997.04a
• /
• pp.93-99
• /
• 1997

Most of the past studies on FMS scheduling problems may be classified into two classes, namely off-line scheduling and on-line scheduling approach. The off-line scheduling methods are used mostly for FMS planning purposes and may not be useful real time control of FMSs, because it generates solutions only after a relatively long period of time. The on-line scheduling methods are used extensively for dynamic real-time control of FMSs although the performance of on-line scheduling algorithms tends vary dramatically depending on various configurations of FMS. Current study is about finding a better on-line scheduling rules for FMS operations. In this study, we propose a method to create priority functions that can be used in setting relative priorities among jobs or machines in on-line scheduling. The priority functions reflect the configuration of FMS and the user-defined objective functions. The priority functions are generated from diverse dispatching rules which may be considered a special priority functions by themselves, and used to determine the order of processing and transporting parts. Overall system of our work consists of two modules, the Priority Function Evolution Module (PFEM) and the FMS Simulation Module (FMSSM). The PFEM generates new priority functions using input variables from a terminal set and primitive functions from a function set by genetic programming. And the FMSSM evaluates each priority function by a simulation methodology. Based on these evaluated values, the PFEM creates new priority functions by using crossover, mutation operation and probabilistic selection. These processes are iteratively applied until the termination criteria are satisfied. We considered various configurations and objective functions of FMSs in our study, and we seek a workable solution rather than an optimum or near optimum solution in scheduling FMS operations in real time. To verify the viability of our approach, experimental results of our model on real FMS are included.