• Bulletin of the Korean Chemical Society
• /
• 제20권3호
• /
• pp.285-290
• /
• 1999

In the present paper, we report a molecular dynamics (MD) simulation study for the structure and dynamics of H+ ions in non-rigid dehydrated H12-A zeolite framework at 298.15 K, using the same method we used in our previous studies of rigid and non-rigid zeolite-A frameworks. It is found that two different structures appear, depending on the choice of the Lennard-Jones parameter (σ) for the H+ ion, as is also observed in the study of rigid dehydrated H12-A zeolite framework, but the ranges of σ are different for the two structures. It is also found that some of the H+ ions exchanged their sites without changing the number of H+ ions at each site. The agreement between experimental and calculated structural parameters for non-rigid dehydrated H12-A zeolite is generally quite good. The calculated IR spectrum by Fourier transform of the total dipole moment auto-correlation function shows two major peaks, one around 2700 cm-1 and the other around 7000 cm-1. The former appears in the calculated IR spectra of non-rigid zeolite-A framework only system and the latter remains unexplained except, perhaps, as an indication of a new formation of a vibrational mode of the framework due to the adsorption of the H+ ions.

• International journal of advanced smart convergence
• /
• 제12권1호
• /
• pp.53-58
• /
• 2023

With the fast development of artificial intelligence day by day, users are demanding explanations about the results of algorithms and want to know what parameters influence the results. In this paper, we propose a model for bankruptcy prediction with interpretability using the SHAP framework. SHAP (SHAPley Additive exPlanations) is framework that gives a visualized result that can be used for explanation and interpretation of machine learning models. As a result, we can describe which features are important for the result of our deep learning model. SHAP framework Force plot result gives us top features which are mainly reflecting overall model score. Even though Fully Connected Neural Networks are a "black box" model, Shapley values help us to alleviate the "black box" problem. FCNNs perform well with complex dataset with more than 60 financial ratios. Combined with SHAP framework, we create an effective model with understandable interpretation. Bankruptcy is a rare event, then we avoid imbalanced dataset problem with the help of SMOTE. SMOTE is one of the oversampling technique that resulting synthetic samples are generated for the minority class. It uses K-nearest neighbors algorithm for line connecting method in order to producing examples. We expect our model results assist financial analysts who are interested in forecasting bankruptcy prediction of companies in detail.

• 로봇학회논문지
• /
• 제19권2호
• /
• pp.203-212
• /
• 2024

Simultaneous motion is essential in the activities of daily living (ADL). For motion intention recognition, surface electromyogram (sEMG) and corresponding motion label is necessary. However, this process is time-consuming and it may increase the burden of the user. Therefore, we propose a simultaneous motion recognition framework using data augmentation based on muscle activation model. The model consists of multiple point sources to be optimized while the number of point sources and their initial parameters are automatically determined. From the experimental results, it is shown that the framework has generated the data which are similar to the real one. This aspect is quantified with the following two metrics: structural similarity index measure (SSIM) and mean squared error (MSE). Furthermore, with k-nearest neighbor (k-NN) or support vector machine (SVM), the classification accuracy is also enhanced with the proposed framework. From these results, it can be concluded that the generalization property of the training data is enhanced and the classification accuracy is increased accordingly. We expect that this framework reduces the burden of the user from the excessive and time-consuming data acquisition.

• 한국멀티미디어학회논문지
• /
• 제24권10호
• /
• pp.1403-1413
• /
• 2021

In this paper, we investigate automated data validation techniques for artificial intelligence training, and also disclose open-source frameworks, such as Google's TensorFlow Data Validation (TFDV), that support automated data validation in the AI model development process. We also introduce an experimental study using public data sets to demonstrate the effectiveness of the open-source data validation framework. In particular, we presents experimental results of the data validation functions for schema testing and discuss the limitations of the current open-source frameworks for semantic data. Last, we introduce the latest studies for the semantic data validation using machine learning techniques.

• Structural Engineering and Mechanics
• /
• 제35권5호
• /
• pp.571-592
• /
• 2010

An energy-based fatigue life prediction framework was previously developed by the authors for prediction of axial, bending and shear fatigue life at various stress ratios. The framework for the prediction of fatigue life via energy analysis was based on a new constitutive law, which states the following: the amount of energy required to fracture a material is constant. In the first part of this study, energy expressions that construct the constitutive law are equated in the form of total strain energy and the distortion energy dissipated in a fatigue cycle. The resulting equation is further evaluated to acquire the equivalent stress per cycle using energy based methodologies. The equivalent stress expressions are developed both for biaxial and multiaxial fatigue loads and are used to predict the number of cycles to failure based on previously developed prediction criterion. The equivalent stress expressions developed in this study are further used in a new finite element procedure to predict the fatigue life for two and three dimensional structures. In the second part of this study, a new Quadrilateral fatigue finite element is developed through integration of constitutive law into minimum potential energy formulation. This new QUAD-4 element is capable of simulating biaxial fatigue problems. The final output of this finite element analysis both using equivalent stress approach and using the new QUAD-4 fatigue element, is in the form of number of cycles to failure for each element on a scale in ascending or descending order. Therefore, the new finite element framework can provide the number of cycles to failure at each location in gas turbine engine structural components. In order to obtain experimental data for comparison, an Al6061-T6 plate is tested using a previously developed vibration based testing framework. The finite element analysis is performed for Al6061-T6 aluminum and the results are compared with experimental results.

• Management Science and Financial Engineering
• /
• 제8권1호
• /
• pp.53-75
• /
• 2002

In this Paper, an attempt is made to establish a learning framework for robust planning and real-time execution control. Necessary definitions and concepts are clearly presented to describe real-time operational control in response to Plan disruptions. A general mathematical framework for disruption recovery is also laid out. Global disruption model is decomposed into suitable number of local disruption models. Execution Pattern is designed to capture local disruptions using decomposed-reverse neural mappings, and to further demonstrate how the decomposed-reverse mappings could be applied for solving disrubtion recovery problems. Two decomposed-reverse neural mappings, N-K-M and M-K-N are employed to produce transportation solutions in react-time. A potential extension is also discussed using the proposed mapping principle and other hybrid heuristics. Experimental results are provided to verify the proposed approach.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제7권4호
• /
• pp.819-833
• /
• 2013

Computing global features based on local features using a bio-inspired framework has shown promising performance. However, for some tough applications with large intra-class variances, a single local feature is inadequate to represent all the attributes of the images. To integrate the complementary abilities of multiple local features, in this paper we have extended the efficacy of the bio-inspired framework, HMAX, to adapt heterogeneous features for global feature extraction. Given multiple global features, we propose an approach, designated as parameterized metric learning, for high dimensional feature fusion. The fusion parameters are solved by maximizing the canonical correlation with respect to the parameters. Experimental results show that our method achieves significant improvements over the benchmark bio-inspired framework, HMAX, and other related methods on the Caltech dataset, under varying numbers of training samples and feature elements.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2006년 학술대회 논문집 정보 및 제어부문
• /
• pp.181-183
• /
• 2006

This paper proposes an intelligent decision framework for update of the environment model using GSPN(generalized stochastic petri nets). The GSPN has several advantages over direct use of the Markov Process. The modeling, analysis, and performance evaluation are conducted on the mathematical basis. By adopting the probabilistic approach, our decision framework helps the robot to decide the time to update the map. The robot navigates autonomously for a long time in dynamic environments. Experimental results show that the proposed scheme is useful for service robots which work semi-permanently and improves dependability of navigation in dynamic environments.

• 제어로봇시스템학회논문지
• /
• 제15권7호
• /
• pp.720-727
• /
• 2009

This paper proposes a new integrated navigation system for a mobile robot in indoor environments. This system consists of five frameworks which are classified by function. This architecture can make the navigation system scalable and flexible. The robot can recover from exceptional situations, such as environmental changes, failure of entering the narrow path, and path occupation by moving objects, using the exception recovery framework. The environmental change can be dealt with using the probabilistic approach, and the problems with the narrow path and path occupation are solved using the ray casting algorithm and the Bayesian update rule. The proposed navigation system was successfully applied to several robots and operated in various environments. Experimental results showed good performance in that the exception recovery framework significantly increased the success rate of navigation. The system architecture proposed in this paper can reduce the time for developing robot applications through its reusability and changeability.

• Journal of Ubiquitous Convergence Technology
• /
• 제1권1호
• /
• pp.18-22
• /
• 2007

Multi-Instance Learning(MIL) performs well to deal with inherently ambiguity of images in multimedia retrieval. In this paper, an effective framework for Contented-Based Image Retrieval(CBIR) with MIL techniques is proposed, the effective mechanism is based on the image segmentation employing improved Mean Shift algorithm, and processes the segmentation results utilizing mathematical morphology, where the goal is to detect the semantic concepts contained in the query. Every sub-image detected is represented as a multiple features vector which is regarded as an instance. Each image is produced to a bag comprised of a flexible number of instances. And we apply a few number of MIL algorithms in this framework to perform the retrieval. Extensive experimental results illustrate the excellent performance in comparison with the existing methods of CBIR with MIL.