• Journal of Intelligence and Information Systems
• /
• v.29 no.3
• /
• pp.145-165
• /
• 2023

Recommendation system plays a significant role on relieving difficulties of selecting information among rapidly increasing amount of information caused by the development of the Internet and on efficiently displaying information that fits individual personal interest. In particular, without the help of recommendation system, E-commerce and OTT companies cannot overcome the long-tail phenomenon, a phenomenon in which only popular products are consumed, as the number of products and contents are rapidly increasing. Therefore, the research on recommendation systems is being actively conducted to overcome the phenomenon and to provide information or contents that are aligned with users' individual interests, in order to induce customers to consume various products or contents. Usually, collaborative filtering which utilizes users' historical behavioral data shows better performance than contents-based filtering which utilizes users' preferred contents. However, collaborative filtering can suffer from cold-start problem which occurs when there is lack of users' historical behavioral data. In this paper, hybrid music recommendation system, which can solve cold-start problem, is proposed based on the playlist data of Melon music streaming service that is given by Kakao Arena for music playlist continuation competition. The goal of this research is to use music tracks, that are included in the playlists, and metadata of music tracks and playlists in order to predict other music tracks when the half or whole of the tracks are masked. Therefore, two different recommendation procedures were conducted depending on the two different situations. When music tracks are included in the playlist, LightFM is used in order to utilize the music track list of the playlists and metadata of each music tracks. Then, the result of Item2Vec model, which uses vector embeddings of music tracks, tags and titles for recommendation, is combined with the result of LightFM model to create final recommendation list. When there are no music tracks available in the playlists but only playlists' tags and titles are available, recommendation was made by finding similar playlists based on playlists vectors which was made by the aggregation of FastText pre-trained embedding vectors of tags and titles of each playlists. As a result, not only cold-start problem can be resolved, but also achieved better performance than ALS, BPR and Item2Vec by using the metadata of both music tracks and playlists. In addition, it was found that the LightFM model, which uses only artist information as an item feature, shows the best performance compared to other LightFM models which use other item features of music tracks.

• The Journal of the Korea Contents Association
• /
• v.20 no.10
• /
• pp.89-99
• /
• 2020

In this study, we tried to produce moving-image works through sentiment analysis of music. First, Google natural language API was used for the sentiment analysis of lyrics, then the result was applied to the image visualization rules. In prior engineering researches, text-based sentiment analysis has been conducted to understand users' emotions and attitudes by analyzing users' comments and reviews in social media. In this study, the data was used as a material for the creation of artworks so that it could be used for aesthetic expressions. From the machine's point of view, emotions are substituted with numbers, so there is a limit to normalization and standardization. Therefore, we tried to overcome these limitations by linking the results of sentiment analysis of lyrics data with the rules of formative elements in visual arts. This study aims to transform existing traditional art works such as literature, music, painting, and dance to a new form of arts based on the viewpoint of the machine, while reflecting the current era in which artificial intelligence even attempts to create artworks that are advanced mental products of human beings. In addition, it is expected that it will be expanded to an educational platform that facilitates creative activities, psychological analysis, and communication for people with developmental disabilities who have difficulty expressing emotions.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.17 no.4
• /
• pp.361-372
• /
• 2016

The purpose of this study was to develop a STEAM-based science education program for children and to verify its effectiveness. An S-STEAM-based science education program for young children was developed through careful analysis of prior research on science education for young children and S-STEAM. The participants were 29 four-year-old children from daycare centers located in Seoul (an experimental group of 14 and comparative group of 15). The S-STEAM program was applied to the experimental group, while the control group went through a general science education course provided by the government. TTCT of Creative Thinking (TTCT: Figures A and B) was used as a research tool, and a multiple intelligence test tool was applied to teachers of the groups. Afterwards, analysis of covariance was implemented to find the S-STEAM program's effects. First, the results showed positive effects on overall creativity, as well as in fluency, originality, abstractness, elaboration, and openness components of creativity. Second, the results showed positive effects on overall multiple intelligences and its components of linguistic, musical, spatial, logical/mathematical, physical exercise, interpersonal, and naturalist intelligence.

• Journal of Convergence for Information Technology
• /
• v.9 no.5
• /
• pp.34-39
• /
• 2019

One of the words frequently mentioned in our society today is the smart machine. Smart machines are machines that contain smart or intelligent functions. These smart machines have recently been applied in our home environment. These are phenomena that occur as a result of smart home. In a smart home environment, smart speakers have moved away from traditional music playback functions and are now increasingly serving as interfaces to control devices, the various components of a smart home. In this study, the technology trends of domestic and foreign smart speaker market are examined, problems of current products are analyzed, and necessary core technologies are described. In the domestic smart speaker market, SKT and KT are leading the related industries, while major IT companies such as Amazon, Google and Apple are focusing on launching related products and technology development.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2017.01a
• /
• pp.109-110
• /
• 2017

현재 국내외에 거실을 잡기위한 노력들이 전방위적으로 이뤄지고 있다. 거실에서 쉽게 쇼핑을 하고, 커튼이나 조명을 켜고, 음악을 듣고, 피자를 시켜 먹으며, 외부 약속을 위해 택시를 부르는, 글로벌 기업들의 '스마트 홈'의 기능으로 인공지능과 음성인식을 통한 산물이 되었다. 또한 이러한 데이터를 중심으로 빅데이터의 보고가 되어간다. 때문에 구글, 아마존, MS, 삼성과 우리나라의 SKT와 LG 등의 기업들이 이러한 기술기반으로 접근하는 현황을 파악하고, 기술에 대한 적정성을 제안할 필요가 있다.

• Annual Conference of KIPS
• /
• 2019.10a
• /
• pp.228-230
• /
• 2019

일상생활에서 빈번히 사용되는 주변 생활 제품은 기술의 급속한 발전과 더불어 지능화 정도가 가속화 되고 있는 추세에 있다. 일반적으로 LED 조명이나 실내 에어컨을 자동으로 조정하거나 자동청소로봇 등은 이미 일반화된 제품들이다. 그럼에도 불구하고 우리 생활에서 가장 필요한 용품인 거울에 대한 지능적인 제품은 비교적 고가이어서 소비자가 쉽게 접근하기 어려운 생활용품이 되고 있어서 지능화 제품의 보급이 더딘편이다. 따라서 본 논문에서는 Raspberry Pi 3B+ 를 기반으로 하여 음성제어가 가능한 스마트 미러를 설계하고 구현하였다. 이를 위하여 저렴한 raspberry pi의 WiFi를 통해 네트워크에 연결하도록 하여 미러가 시간, 날씨 및 뉴스 정보 기능을 자동으로 업데이트 할 수 있도록 하였고 기상 조건, 사전 시간 또는 음악 재생과 같은 음성 제어가 가능하기 위하도록 Google Asistant 음성 인터페이스를 적용하였다. 본 논문에서 제안한 제품이 실용화될 경우 저가이면서 고기능 사양을 제공하고 있어서 스마트 미러 보급에 많은 기여가 예산된다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.851-853
• /
• 2022

인공지능을 활용한 다양한 딥러닝 기술의 보급과 상용화로 오디오 음성 인식 분야에서도 음성 인식의 정확도를 높이기 위한 다양한 연구가 진행되고 있다. 최근 STT 를 위한 음성 인식 엔진은 딥러닝 기술을 기반으로 과거에 비해 높은 정확도를 보이고 있다. 하지만 예능 프로그램, 드라마, 스포츠 방송 등과 같이 비음성 신호와 음성 신호가 함께 녹음되는 오디오의 경우 음성 인식 정확도가 크게 낮아지는 문제가 발생한다. 이에 본 연구에서는 다양한 장르의 오디오를 음성과 음악을 분리하는 딥러닝 모델을 활용하여 음성 신호와 비음성 신호로 분리하는 방법을 제시하고, STT 결과를 분석하여 음성 인식의 정확도를 높이기 위한 연구 방향을 제시한다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.7
• /
• pp.945-951
• /
• 2007

Noise cancellation is essential for the devices which use speech as an interface. In real environments, speech quality and recognition rates are degraded by the auditive noises coming near the microphone. In this paper, we propose a noise cancellation algorithm using stereo microphones basically. The advantage of the use of multiple microphones is that the direction information of the target source could be applied. The proposed noise canceller is based on the Wiener filter. To estimate the filter, noise and target speech frequency responses should be known and they are estimated by the spectral classification in the frequency domain. The performance of the proposed algorithm is compared with that of the well-known Frost algorithm and the generalized sidelobe canceller (GSC) with an adaptation mode controller (AMC). As performance measures, the perceptual evaluation of speech quality (PESQ), which is the most widely used among various objective speech quality methods, and speech recognition rates are adopted.

• Journal of Intelligence and Information Systems
• /
• v.13 no.4
• /
• pp.65-78
• /
• 2007

This paper proposes a new method of weighted template matching for machine-printed numeral recognition. The proposed weighted template matching, which emphasizes the feature of a pattern using adaptive Hamming distance on local feature areas, improves the recognition rate while template matching processes an input image as one global feature. Template matching is vulnerable to random noises that generate ragged outlines of a pattern when it is binarized. This paper offers a method of chain code trimming in order to remove ragged outlines. The method corrects specific chain codes within the chain codes of the inner and the outer contour of a pattern. The experiment compares confusion matrices of both the template matching and the proposed weighted template matching with chain code trimming. The result shows that the proposed method improves fairly the recognition rate of the machine-printed numerals.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.21 no.3
• /
• pp.401-406
• /
• 2011

In order to solve a complex optimization problem more efficiently than traditional approaches, various meta-heuristic algorithms such as genetic algorithm, ant-colony algorithm, and harmony search algorithm have been extensively researched. Compared with other meta-heuristic algorithm, harmony search algorithm shows a better result to resolve the complex optimization issues. Harmony search algorithm is inspired by the improvision process of musician for most suitable harmony. In general, the performance of harmony search algorithm is determined by the value of harmony memory considering rate, and pitch adjust rate. In this paper, modified harmony search algorithm is proposed in order to derive best harmony. If the optimal solution of a specific problem can not be found for a certain period of time, a part of original harmony memory is updated as the selected suitable harmonies. Experimental results using test function demonstrate that the updated harmony memory can induce the approximation of reliable optimal solution in the short iteration, because of a few change of fitness.