• The Journal of Korea Robotics Society
• /
• v.19 no.2
• /
• pp.196-202
• /
• 2024

The ability to understand given environments and plan a sequence of actions leading to goal state is crucial for personal service robots. With recent advancements in deep learning, numerous studies have proposed methods for state representation in planning. However, previous works lack explicit information about relationships between objects when the state observation is converted to a single visual embedding containing all state information. In this paper, we introduce graph-based state representation that incorporates both object and relationship features. To leverage these advantages in addressing the task planning problem, we propose a Graph Neural Network (GNN)-based subgoal prediction model. This model can extract rich information about object and their interconnected relationships from given state graph. Moreover, a search-based algorithm is integrated with pre-trained subgoal prediction model and state transition module to explore diverse states and find proper sequence of subgoals. The proposed method is trained with synthetic task dataset collected in simulation environment, demonstrating a higher success rate with fewer additional searches compared to baseline methods.

• Smart Media Journal
• /
• v.12 no.1
• /
• pp.9-16
• /
• 2023

Recently, a convolutional neural network (CNN) based system is being developed to overcome the limitations of human resources in the apple quality classification of farmhouse. However, since convolutional neural networks receive only images of the same size, preprocessing such as sampling may be required, and in the case of oversampling, information loss of the original image such as image quality degradation and blurring occurs. In this paper, in order to minimize the above problem, to generate a image patch based graph of an original image and propose a random walk-based positional encoding method to apply the graph transformer model. The above method continuously learns the position embedding information of patches which don't have a positional information based on the random walk algorithm, and finds the optimal graph structure by aggregating useful node information through the self-attention technique of graph transformer model. Therefore, it is robust and shows good performance even in a new graph structure of random node order and an arbitrary graph structure according to the location of an object in an image. As a result, when experimented with 5 apple quality datasets, the learning accuracy was higher than other GNN models by a minimum of 1.3% to a maximum of 4.7%, and the number of parameters was 3.59M, which was about 15% less than the 23.52M of the ResNet18 model. Therefore, it shows fast reasoning speed according to the reduction of the amount of computation and proves the effect.

• Journal of Microbiology and Biotechnology
• /
• v.26 no.11
• /
• pp.1972-1982
• /
• 2016

Influenza A virus is a single-stranded RNA virus with a genome of negative polarity. Owing to the antigenic diversity and cross concrete shift, an immense number of novel strains have developed astronomically over the years. The present work deals with the codon utilization partialness among five different influenza A viruses isolated from human hosts. All the subtypes showed the homogeneous pattern of nucleotide utilization with a little variation in their utilization frequencies. A lower bias in codon utilization was observed in all the subtypes as reflected by higher magnitudes of an efficacious number of codons. Dinucleotide analysis showed very low CpG utilization and a high predilection of A/T-ending codons. The H5N1 subtype showed noticeable deviation from the rest. Codon pair context analysis showed remarkable depletion of NNC-GNN and NNT-ANN contexts. The findings alluded towards GC-compositional partialness playing a vital role, which is reflected in the consequential positive correlation between the GC contents at different codon positions. Untangling the codon utilization profile would significantly contribute to identifying novel drug targets that will pacify the search for antivirals against this virus.

• Journal of Auto-vehicle Safety Association
• /
• v.13 no.4
• /
• pp.7-13
• /
• 2021

This paper presents an vulnerable road user (VRU) classification and tracking algorithm using vision and LiDAR sensor fusion method for urban autonomous driving. The classification and tracking for vulnerable road users such as pedestrian, bicycle, and motorcycle are essential for autonomous driving in complex urban environments. In this paper, a real-time object image detection algorithm called Yolo and object tracking algorithm from LiDAR point cloud are fused in the high level. The proposed algorithm consists of four parts. First, the object bounding boxes on the pixel coordinate, which is obtained from YOLO, are transformed into the local coordinate of subject vehicle using the homography matrix. Second, a LiDAR point cloud is clustered based on Euclidean distance and the clusters are associated using GNN. In addition, the states of clusters including position, heading angle, velocity and acceleration information are estimated using geometric model free approach (GMFA) in real-time. Finally, the each LiDAR track is matched with a vision track using angle information of transformed vision track and assigned a classification id. The proposed fusion algorithm is evaluated via real vehicle test in the urban environment.

• International Journal of High-Rise Buildings
• /
• v.12 no.3
• /
• pp.203-208
• /
• 2023

The implementation of artificial intelligence (AI) design for tall building structures is an essential solution for addressing critical challenges in the current structural design industry. Generative AI technology is a crucial technical aid because it can acquire knowledge of design principles from multiple sources, such as architectural and structural design data, empirical knowledge, and mechanical principles. This paper presents a set of AI design techniques for building structures based on two types of generative AI: generative adversarial networks and graph neural networks. Specifically, these techniques effectively master the design of vertical and horizontal component layouts as well as the cross-sectional size of components in reinforced concrete shear walls and frame structures of tall buildings. Consequently, these approaches enable the development of high-quality and high-efficiency AI designs for building structures.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.9
• /
• pp.39-48
• /
• 2024

In this study, we propose a method to augment the provided reasoning paths to improve the answer performance and explanatory power of KGQA. In the proposed method, we utilize LLMs and GNNs to retrieve reasoning paths related to the question from the knowledge graph and evaluate reasoning paths. Then, we retrieve the external information related to the question and then converted into triples to answer the question and explain the reason. Our method evaluates the reasoning path by checking inference results and semantically by itself. In addition, we find related texts to the question based on their similarity and converting them into triples of knowledge graph. We evaluated the performance of the proposed method using the WebQuestion Semantic Parsing dataset, and found that it provides correct answers with higher accuracy and more questions with explanations than the reasoning paths by the previous research.

• Journal of Intelligence and Information Systems
• /
• v.29 no.3
• /
• pp.419-437
• /
• 2023

As the number of thin filers in Korea surpasses 12 million, there is a growing interest in enhancing the accuracy of assessing their credit default risk to generate additional revenue. Specifically, researchers are actively pursuing the development of default prediction models using machine learning and deep learning algorithms, in contrast to traditional statistical default prediction methods, which struggle to capture nonlinearity. Among these efforts, Graph Neural Network (GNN) architecture is noteworthy for predicting default in situations with limited data on thin filers. This is due to their ability to incorporate network information between borrowers alongside conventional credit-related data. However, prior research employing graph neural networks has faced limitations in effectively handling diverse categorical variables present in credit information. In this study, we introduce the Transformer embedded Graph Convolutional Network (TeGCN), which aims to address these limitations and enable effective default prediction for thin filers. TeGCN combines the TabTransformer, capable of extracting contextual information from categorical variables, with the Graph Convolutional Network, which captures network information between borrowers. Our TeGCN model surpasses the baseline model's performance across both the general borrower dataset and the thin filer dataset. Specially, our model performs outstanding results in thin filer default prediction. This study achieves high default prediction accuracy by a model structure tailored to characteristics of credit information containing numerous categorical variables, especially in the context of thin filers with limited data. Our study can contribute to resolving the financial exclusion issues faced by thin filers and facilitate additional revenue within the financial industry.