• Journal of Information Processing Systems
• /
• v.16 no.6
• /
• pp.1407-1423
• /
• 2020

A graph is a data structure consisting of nodes and edges between these nodes. Graph embedding is to generate a low dimensional vector for a given graph that best represents the characteristics of the graph. Recently, there have been studies on graph embedding, especially using deep learning techniques. However, until now, most deep learning-based graph embedding techniques have focused on unweighted graphs. Therefore, in this paper, we propose a graph embedding technique for weighted graphs based on long short-term memory (LSTM) autoencoders. Given weighted graphs, we traverse each graph to extract node-weight sequences from the graph. Each node-weight sequence represents a path in the graph consisting of nodes and the weights between these nodes. We then train an LSTM autoencoder on the extracted node-weight sequences and encode each nodeweight sequence into a fixed-length vector using the trained LSTM autoencoder. Finally, for each graph, we collect the encoding vectors obtained from the graph and combine them to generate the final embedding vector for the graph. These embedding vectors can be used to classify weighted graphs or to search for similar weighted graphs. The experiments on synthetic and real datasets show that the proposed method is effective in measuring the similarity between weighted graphs.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.34C no.10
• /
• pp.98-105
• /
• 1997

In this work, we proesent a bidirectional approximate reasoning method and fuzzy inference network for interval valued decision making systems. For this, we propose a new type of similarity measure between two fuzzy vectors based on the Ordered Weighted Averaging (OWA) operator. Since the proposed similarity measure has a structure to give the extreme values by choosing a suitable weighting vector of the OWA operator, it can render an interval valued similarity value. From this property, we derive a bidirectional approximate reasoning method based on the similarity measure and show its fuzzy inference network implementation for the decision making systems requiring the interval valued decisions.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.3 no.3
• /
• pp.39-44
• /
• 2010

In this paper, we addressed a face transform scheme with age-progressing based on vector representation. Proposed approach utilized a vector modeling as well as morphing so as to improve not only a reliability but also a consistency. For the more, some elements of texture change owing to the face shape are defined and some parameters with respect to the internal and external environments are also considered. To testify the proposed approach, estimation of similarity is performed with qualitative manner by using experimental output, and finally resulted in satisfactory for face shape transformation aged from sixty to fourteen.

• Speech Sciences
• /
• v.13 no.4
• /
• pp.177-186
• /
• 2006

In an effort to enhance the quality of feature vector classification and thereby reduce the recognition error rate of the speaker-independent speech recognition, we employ the Mahalanobis distance in the calculation of the similarity measure between feature vectors. It is assumed that the metric matrix of the Mahalanobis distance be diagonal for the sake of cost reduction in memory and time of calculation. We propose that the diagonal elements be given in terms of the variations of the feature vector components. Geometrically, this prescription tends to redistribute the set of data in the shape of a hypersphere in the feature vector space. The idea is applied to the speech recognition by hidden Markov model with fuzzy vector quantization. The result shows that the recognition is improved by an appropriate choice of the relevant adjustable parameter. The Viterbi score difference of the two winners in the recognition test shows that the general behavior is in accord with that of the recognition error rate.

• The Journal of the Convergence on Culture Technology
• /
• v.7 no.3
• /
• pp.591-598
• /
• 2021

It's possible to get the word-vector by the statistical SVD or deep-learning CBOW and LSTM methods and theses ones learn the contexts of forward/backward words or the sequence of following words. It's used to analyze the poems by Ki Hyung-do with similar words recommended by the word-vector showing the core images of the poetry. It seems at first sight that the words don't go well with the images but they express the similar style described by the reference words once you look close the contexts of the specific poems. The word-vector can analogize the words having the same relations with the ones between the representative words for the core images of the poems. Therefore you can analyze the poems in depth and in variety with the similarity and analogy operations by the word-vector estimated with the statistical SVD or deep-learning CBOW and LSTM methods.

• The Journal of the Acoustical Society of Korea
• /
• v.43 no.2
• /
• pp.200-206
• /
• 2024

Computing music similarity is an indispensable component in developing music search service. This paper proposes a relevance weight of each chromagram vector for cover song identification in computing a music similarity function in order to boost identification accuracy. We derive a music similarity function using the relevance weight based on the probabilistic relevance model, where higher relevance weights are assigned to less frequently-occurring discriminant chromagram vectors while lower weights to more frequently-occurring ones. Experimental results performed on two cover music datasets show that the proposed music similarity improves the cover song identification performance.

• Journal of the Korean Data and Information Science Society
• /
• v.14 no.4
• /
• pp.853-861
• /
• 2003

An improved K-means document clustering method has been presented, where a concept vector is manipulated for each cluster on the basis of cosine similarity of text documents. The concept vectors are unit vectors that have been normalized on the n-dimensional sphere. Because the standard K-means method is sensitive to initial starting condition, our improvement focused on starting condition for estimating the modes of a distribution. The improved K-means clustering algorithm has been applied to a set of text documents, called Classic3, to test and prove efficiency and correctness of clustering result, and showed 7% improvements in its worst case.

• The KIPS Transactions:PartB
• /
• v.19B no.4
• /
• pp.243-248
• /
• 2012

In the information retrieval systems like vector model implementation and document clustering, document similarity calculation takes a great part on the overall performance of the system. In this paper, GPU parallelism has been explored to enhance the processing speed of document similarity calculation in a CUDA framework. The proposed method increased the similarity calculation speed almost 15 times better compared to the typical CPU-based framework. It is 5.2 and 3.4 times better than the methods by using CUBLAS and Thrust, respectively.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2010.10a
• /
• pp.543-543
• /
• 2010

The objective of this research is estimating the location of interested sound source by using the similarity between a beamformer output in time domain and the candidate signal. The waveform of beamformer output at the location of sound source is similar with the waveform emitted by that source. To estimate the location of sound source by using this feature, we define quantified similarity between candidate signal and beamformer output. The candidate signal describes the signal which is generated by interested source. In this paper, similarity is defined by four methods. The two methods use time vector comparison, and the other two methods use time-frequency map or linear prediction coefficients. To figure out the results and performance of localization by using similarities, we demonstrate two conditions. The one is when two pure tone sources exist and the other condition is when several bird sounds exist. As a consequence, inner product with two time-vectors and structural similarity with spectrograms can estimate the locations of interest sound source.

• The Journal of the Acoustical Society of Korea
• /
• v.30 no.3
• /
• pp.136-141
• /
• 2011

In this paper, a method for measuring rhythmic similarity between two sound signals for auditory memory game is proposed. The proposed method analyzes energy fluctuation, the temporal duration of energy peak, the timbre of two signals, and detects beat positions for each signal. Then, it determines the rhythm vector after compensating a difference in tempo and the number of beats between two signals. Finally, a method for rhythmic similarity measurement is defined as a function of the dissimilarity between two rhythm vectors and a difference in the number of beats. The rhythmic similarity measured by the proposed method and that by the subjective listening test are compared, and the correlation of 0.86 between two results is achieved.