Digital Library[ Search Result ]
A Path Fragment Management Structure for Fast Projection Candidate Selection of the Path Prediction Algorithm
Dongwon Jeong, Sukhoon Lee, Doo-Kwon Baik
This paper proposes an enhanced projection candidate selection algorithm to improve the performance of the existing path prediction algorithm. Various user path prediction algorithms have previously been developed, but those algorithms are inappropriate for a real-time and close user path prediction environment. To resolve this issue, a new prediction algorithm has been proposed, but several problems still remain. In particular, this algorithm should be enhanced to provide much faster processing performance. The major cause of the high processing time of the previous path prediction algorithm is the high time complexity of its projection candidate selection. Therefore, this paper proposes a new path fragment management structure and an improved projection candidate selection algorithm to improve the processing speed of the existing projection candidate selection algorithm. This paper also shows the effectiveness of the algorithm herein proposed through a comparative performance evaluation.
Robust Particle Filter Based Route Inference for Intelligent Personal Assistants on Smartphones
Much research has been conducted on location-based intelligent personal assistants that can understand a user"s intention by learning the user"s route model and then inferring the user"s destinations and routes using data of GPS and other sensors in a smartphone. The intelligence of the location-based personal assistant is contingent on the accuracy and efficiency of the real-time predictions of the user"s intended destinations and routes by processing movement information based on uncertain sensor data. We propose a robust particle filter based on Dynamic Bayesian Network model to infer the user"s routes. The proposed robust particle filter includes a particle generator to supplement the incorrect and incomplete sensor information, an efficient switching function and an weight function to reduce the computation complexity as well as a resampler to enhance the accuracy of the particles. The proposed method improves the accuracy and efficiency of determining a user"s routes and destinations.
A Pretrained Model-Based Approach to Improve Generalization Performance for ADMET Prediction of Drug Candidates
http://doi.org/10.5626/JOK.2025.52.7.601
Accurate prediction of Absorption, Distribution, Metabolism, Excretion, and Toxicity (ADMET) properties plays an important role in reducing clinical trial failure rates and lowering drug development costs. In this study, we propose a novel method to improve ADMET prediction performance for drug candidate compounds by integrating molecular embeddings from a graph transformer model with pretrained embeddings from a UniMol model. The proposed model can capture bond type information from molecular graph structures, generating chemically refined representations, while leveraging UniMol’s pretrained 3D embeddings to effectively learn spatial molecular characteristics. Through this, the model is designed to address the problem of data scarcity and enhance the generalization performance. In this study, we conducted prediction experiments on 10 ADMET properties. The experiment results demonstrated that our proposed model outperformed existing methods and that the prediction accuracy for ADMET properties could be improved by effectively integrating atomic bond information and 3D structures.
Time Series Data Imbalance Resolution Techniques for Early Prediction
Eungseon An, Taehyoung Kwon, Doguk Kim
http://doi.org/10.5626/JOK.2025.52.7.593
Time series forecasting is a critical task that involves analyzing observed time series data to predict future values. However, when dealing with imbalanced data, model performance can degrade, leading to biased predictions. Although recent studies have explored various deep learning techniques and data augmentation methods, many fail to address challenges posed by data imbalance and the intrinsic characteristics of time series data simultaneously, leaving underlying issues unresolved. This study proposed a novel approach that could leverage temporal patterns to generate synthetic samples and extend the scope of early prediction. By identifying key moments that could effectively distinguish between positive and negative classes, our method enhanced the ability to predict further into the future. The method proposed in this study demonstrated superior performance to existing methods and proved the feasibility of early prediction for longer time lags.
Analysis of QoQ GDP Prediction Performance Using Deep Learning Time Series Model
Yeonhee Lee, Youngmin Kim, Taewan You
http://doi.org/10.5626/JOK.2022.49.10.873
In this paper, we proposed an algorithm for predicting GDP growth rate using a deep learning time series model spotlighted recently. The proposed algorithm adopts an ensemble deep learning method to ensure stable prediction performance using a large number of economic time series data with low frequency. It also uses a gradual learning method to ensure adaptive performance even in business fluctuations. By demonstrating that the performance could be improved by using economic sector information in learning, the necessity of convergence with domain knowledge was confirmed and the importance of AI operation technology to provide adaptive predictive power was emphasized. Through performance comparison with traditional machine learning models for the COVID-19 period, we proved that deep learning could be a relatively reasonable predictive tool under rapid economic fluctuations. The deep learning-based adaptive AI algorithm presented in this paper is expected to be developed into a deep learning-based autonomous adaptive economic prediction system through combination with AI operation technology.
Knowledge Graph Embedding with Entity Type Constraints
Seunghwan Kong, Chanyoung Chung, Suheon Ju, Joyce Jiyoung Whang
http://doi.org/10.5626/JOK.2022.49.9.773
Knowledge graph embedding represents entities and relationships in the feature space by utilizing the structural properties of the graph. Most knowledge graph embedding models rely only on the structural information to generate embeddings. However, some real-world knowledge graphs include additional information such as entity types. In this paper, we propose a knowledge graph embedding model by designing a loss function that reflects not only the structure of a knowledge graph but also the entity-type information. In addition, from the observation that certain type constraints exist on triplets based on their relations, we present a negative sampling technique considering the type constraints. We create the SMC data set, a knowledge graph with entity-type restrictions to evaluate our model. Experimental results show that our model outperforms the other baseline models.
Quality Estimation of Machine Translation using Dual-Encoder Architecture
Dam Heo, Wonkee Lee, Jong-Hyeok Lee
http://doi.org/10.5626/JOK.2022.49.7.521
Quality estimation (QE) is the task of estimating the quality of given machine translations (MTs) without their reference translations. A recent research trend is to apply transfer learning to a pre-training model based on Transformer encoder with a parallel corpus in QE. In this paper, we proposed a dual-encoder architecture that learns a monolingual representation of each respective language in encoders. Thereafter, it learns a cross-lingual representation of each language in cross-attention networks. Thus, it overcomes the limitations of a single-encoder architecture in cross-lingual tasks, such as QE. We proved that the dual-encoder architecture is structurally more advantageous over the single-encoder architecture and furthermore, improved the performance and stability of the dual-encoder model in QE by applying the pre-trained language model to the dual-encoder model. Experiments were conducted on WMT20 QE data for En-De pair. As pre-trained models, our model employs English BERT (Bidirectional Encoder Representations from Transformers) and German BERT to each encoder and achieves the best performance.
Efficient Compositional Translation Embedding for Visual Relationship Detection
Yu-Jung Heo, Eun-Sol Kim, Woo Suk Choi, Kyoung-Woon On, Byoung-Tak Zhang
http://doi.org/10.5626/JOK.2022.49.7.544
Scene graphs are widely used to express high-order visual relationships between objects present in an image. To generate the scene graph automatically, we propose an algorithm that detects visual relationships between objects and predicts the relationship as a predicate. Inspired by the well-known knowledge graph embedding method TransR, we present the CompTransR algorithm that i) defines latent relational subspaces considering the compositional perspective of visual relationships and ii) encodes predicate representations by applying transitive constraints between the object representations in each subspace. Our proposed model not only reduces computational complexity but also outperformed previous state-of-the-art performance in predicate detection tasks in three benchmark datasets: VRD, VG200, and VrR-VG. We also showed that a scene graph could be applied to the image-caption retrieval task, which is one of the high-level visual reasoning tasks, and the scene graph generated by our model increased retrieval performance.
Knowledge Graph Embedding for Link Prediction using Node-Link Interaction-based Graph Attention Networks
http://doi.org/10.5626/JOK.2022.49.7.555
Knowledge graphs are structures that express knowledge in the real world in the form of nodes and links-based triple form. These knowledge graphs are incomplete and many embedding techniques have been studied to effectively represent nodes and links in low-dimensional vector spaces to find other missing relationships. Recently, many neural network-based knowledge graph link prediction methods have been studied. However existing models consider nodes and links independently when determining the importance of a triple to a node which makes it difficult to reflect the interaction between nodes and links. In this paper, we propose an embedding method that will be used to analyze the importance of triple units by simultaneously considering nodes and links using composition operators, and at the same time prove that the model outperforms other methods in knowledge graph link prediction.
Development of an Apartment Price Change Rate Prediction Model with Geographical Adjacency
http://doi.org/10.5626/JOK.2022.49.6.424
Recently, in the real estate market, decoupling in which housing prices fluctuate by the region has been escalating. This phenomenon implies that each region is composed of districts that are adjacent to one another. This thesis confirms that the prices of a district change in synchronization with that of the adjacent districts and proves that the fluctuations in apartment prices in the districts within Seoul are due to the neighbors. The rate of change in apartment prices, macroeconomic indicators, and private education indicators are used to test the hypothesis with a 3D (time, distance, and attribute) model, which is further deciphered using CNN. The model considers the situation of neighbors and is subdivided into the following three sub-models: consideration only for the target area (I), consideration for long-distance areas (II), and change in the number of neighbors (III). The metrics used are mean absolute error and mean directional accuracy. It was observed that the model with neighbors performed better than the persistence model and XGBoost. Furthermore, its sub-models showed good performance in the order of model III (with 3 neighbors), II, and I. This study clearly exhibits that the factor “neighbor” affects the rate of change in apartment prices.
Search
Journal of KIISE
- ISSN : 2383-630X(Print)
- ISSN : 2383-6296(Electronic)
- KCI Accredited Journal
Editorial Office
- Tel. +82-2-588-9240
- Fax. +82-2-521-1352
- E-mail. chwoo@kiise.or.kr