• Annual Conference of KIPS
• /
• 2023.05a
• /
• pp.442-444
• /
• 2023

그래프 임베딩 방법은 그래프 구조를 이용하여 그래프의 노드를 저차원 임베딩 공간에서 벡터로 매핑하여 각 노드를 벡터로 표현하는 것을 목표로 한다. 다양한 방법들이 제시되었지만 기존의 방법들은 그래프에서 노드 간의 유사성을 잘 보존할 수 없어 다양한 기계 학습에 대해 부정확한 벡터를 생성하였다. 이러한 문제를 해결하기 위해 노드 사이의 유사성을 이용한 방법이 제안되었다. 본 논문에서, 우리는 여섯 가지 실세계 데이터셋을 사용하여 세 가지 기계 학습 작업시 그래프 임베딩 방법들의 성능을 비교하여 유사성 기반의 그래프 임베딩 방법의 우수성을 확인했다.

• Annual Conference of KIPS
• /
• 2023.05a
• /
• pp.468-469
• /
• 2023

실세계 네트워크 데이터에서 노드들 간의 관계는 종종 친구/적 혹은 지지/반대와 같이 대조적인 부호를 갖는다. 이러한 네트워크를 분석하기 위해, 부호가 있는 네트워크 임베딩 (signed network embedding, 이하 SNE) 문제에 대한 관심이 급증하고 있다. 특히, 최근 들어 그래프 합성곱 네트워크 기술을 기반으로 하는 SNE 방법들에 대한 연구가 활발히 수행되어 오고 있다. 본 논문에서는, 부호가 있는 네트워크의 희소성 정도가 기존 SNE 방법들의 성능에 어떻게 영향을 미치는 지에 대해 분석하고자 한다. 4 개의 실세계 데이터 집합들을 이용한 실험을 통해, 우리는 기존 방법들의 부호 예측 정확도가 희소한 네트워크들에서는 상당히 감소하는 것을 확인하였다.

• Annual Conference of KIPS
• /
• 2023.11a
• /
• pp.1049-1050
• /
• 2023

빅데이터 시대에 추천시스템은 끊임없이 진화하고 있으며 그 중요성도 나날이 높아지고 있다. 그럼에도, 친구추천시스템은 여전히 사용자의 맥락 정보에만 의존하는 제한적인 모습을 보이고 있다. 이에 본 연구는 사용자의 대화에 드러난 관심사와 대화 만족도 등을 직접 사용자 임베딩에 활용한 Chat-Trip 모델을 제안하고, 실험을 통해 그 효용성을 입증하였다.

• 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.

• Annual Conference of KIPS
• /
• 2024.05a
• /
• pp.641-643
• /
• 2024

실세계에서는 두개 이상의 객체들이 서로 관계를 맺고있다. 단 두 객체 간의 관계만 표현하는 그래프와는 달리 여러 객체들 간의 관계를 표현하는 하이퍼그래프는 그룹 상호작용을 잘 표현할 수 있다. 이러한 강점으로 하이퍼그래프를 활용한 응용들이 많이 제안되고 있다. 하이퍼그래프 임베딩은 하이퍼그래프의 구조를 이용하여 노드를 저차원 벡터로 표현하는 방법이다. 이렇게 표현된 벡터들은 노드 분류, 커뮤니티 탐지, 링크예측 등 광범위한 응용에 활용된다. 하지만 하이퍼그래프는 그래프보다 희소성 문제가 훨씬 더 심해 데이터 셋의 희소성이 하이퍼그래프 임베딩 방법의 성능에 큰 영향을 미칠 수 있다. 따라서, 본 논문에서는 희소성에 따른 하이퍼그래프 임베딩 방법들의 성능을 분석하고자 한다. 우리는 8 개의 실세계 데이터셋을 이용한 실험을 통해 데이터가 희소할수록 하이퍼그래프 임베딩 방법들의 성능이 감소하는 것을 확인하였다.

• Journal of Intelligence and Information Systems
• /
• v.5 no.1
• /
• pp.103-123
• /
• 1999

Input filtering as a preprocessing method is so much crucial to get good performance in time series forecasting. There are a few preprocessing methods (i.e. ARMA outputs as time domain filters, and Fourier transform or wavelet transform as time-frequency domain filters) for handling time series. Specially, the time-frequency domain filters describe the fractal structure of financial markets better than the time domain filters due to theoretically additional frequency information. Therefore, we, first of all, try to describe and analyze specially some issues on the effectiveness of different filtering methods from viewpoint of the performance of a neural network based forecasting. And then we discuss about neural network model architecture issues, for example, what type of neural network learning architecture is selected for our time series forecasting, and what input size should be applied to a model. In this study an input selection problem is limited to a size selection of the lagged input variables. To solve this problem, we simulate on analyzing and comparing a few neural networks having different model architecture and also use an embedding dimension measure as chaotic time series analysis or nonlinear dynamic analysis to reduce the dimensionality (i.e. the size of time delayed input variables) of the models. Throughout our study, experiments for integration methods of joint time-frequency analysis and neural network techniques are applied to a case study of daily Korean won / U. S dollar exchange returns and finally we suggest an integration framework for future research from our experimental results.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.12
• /
• pp.3798-3814
• /
• 2022

Social recommendation algorithm can alleviate data sparsity and cold start problems in recommendation system by integrated social information. Among them, matrix-based decomposition algorithms are the most widely used and studied. Such algorithms use dot product operations to calculate the similarity between users and items, which ignores user's potential preferences, reduces algorithms' recommendation accuracy. This deficiency can be avoided by a metric learning-based social recommendation algorithm, which learns the distance between user embedding vectors and item embedding vectors instead of vector dot-product operations. However, previous works provide no theoretical explanation for its plausibility. Moreover, most works focus on the indirect impact of social friends on user's preferences, ignoring the direct impact on user's rating preferences, which is the influence of user rating preferences. To solve these problems, this study proposes a user bias drift social recommendation algorithm based on metric learning (BDML). The main work of this paper is as follows: (1) the process of introducing metric learning in the social recommendation scenario is introduced in the form of equations, and explained the reason why metric learning can replace the click operation; (2) a new user bias is constructed to simultaneously model the impact of social relationships on user's ratings preferences and user's preferences; Experimental results on two datasets show that the BDML algorithm proposed in this study has better recommendation accuracy compared with other comparison algorithms, and will be able to guarantee the recommendation effect in a more sparse dataset.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.8
• /
• pp.53-58
• /
• 2024

Traffic accidents are not only a threat to human lives but also pose significant societal costs. Recently, research has been conducted to address the issue of traffic accidents by predicting the risk using deep learning technology and spatiotemporal information of roads. However, while traffic accidents are influenced not only by the spatiotemporal information of roads but also by human factors, research on the latter has been relatively less active. This paper analyzes driver groups and characteristics by applying clustering techniques to a traffic accident dataset and proposes and applies a method to calculate the Risk Level for each driver group and characteristic. In this process, the preprocessing technique suggested in this paper demonstrates a higher Silhouette Score of 0.255 compared to the commonly used One-Hot Embedding & Min-Max Scaling techniques, indicating its suitability as a preprocessing method.

• Journal of Mechanical Science and Technology
• /
• v.20 no.4
• /
• pp.455-466
• /
• 2006

This study was an attempt to control and monitor Computerized Numerical Controller (CNC) machines anywhere and anytime for the development of a ubiquitous machine (u-machine). With a Personal Digital Assistant (PDA) phone, the machine status and machining data of CNC machines can be monitored in wired and wireless environments, including the environments of IMT2000 and Wireless LAN. Moreover, CNC machines can be controlled anywhere and anytime. The concept of the anywhere-anytime controlling and monitoring of a manufacturing system was implemented in this study for the purpose of u-manufacturing and u-machines. In this concept, the communication between the CNC controller and the PDA phone was successfully performed anywhere and anytime for the real-time monitoring and control of CNC machines. In addition, the interface between the CNC controller and the developed application module was implemented by Object linking and embedding for Process Control (OPC) and shared CNC memory. For communication, the design of a server contents module within the target CNC was based on a TCP/IP. Furthermore, the client contents module within the PDA phone was designed with the aid of embedded c++ programming for mobile communication. For the interface, the monitoring data, such as the machine status, the machine running state, the name of the Numerical Control (NC) program, the alarm and the position of the stage axes, were acquired in real time from real machines with the aid of the OPC method and by sharing the CNC memory. The control data, such as the start, hold, emergency stop, reserved start and reserved stop, were also applied to the CNC domain of the real machine. CNC machines can therefore be controlled and monitored in real time, anywhere and anytime. Moreover, prompt notification from CNC machines to mobile phones, including cellular phones and PDA phones, can be automatically realized in emergencies.

• Journal of KIISE
• /
• v.44 no.1
• /
• pp.51-56
• /
• 2017

The research area for solving question answering (QA) problems using artificial intelligence models is in a methodological transition period, and one such architecture, the dynamic memory network (DMN), is drawing attention for two key attributes: its attention mechanism defined by neural network operations and its modular architecture imitating cognition processes during QA of human. In this paper, we increased accuracy of the inferred answers, by adapting an automatic data augmentation method for lacking amount of training data, and by improving the ability of time perception. The experimental results showed that in the 1K-bAbI tasks, the modified DMN achieves 89.21% accuracy and passes twelve tasks which is 13.58% higher with passing four more tasks, as compared with one implementation of DMN. Additionally, DMN's word embedding vectors form strong clusters after training. Moreover, the number of episodic passes and that of supporting facts shows direct correlation, which affects the performance significantly.