• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.1
• /
• pp.19-24
• /
• 2023

In traditional fish farming way, the workers have to observe all of the pools every time and every day to feed at the right timing. This method causes tremendous stress on workers and wastes time. To solve this problem, we implemented an automatic detection system for feeding time using deep neural network. The detection system consists of two steps: classification of the presence or absence of feed and checking DO (Dissolved Oxygen) of the pool. For the classification, the pretrained ResNet18 model and transfer learning with custom dataset are used. DO is obtained from the DO sensor in the pool through HTTP in real time. For better accuracy, the next step, checking DO proceeds when the result of the classification is absence of feed several times in a row. DO is checked if it is higher than a DO reference value that is set by the workers. These actions are performed automatically in the UI programs developed with LabVIEW.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.280-283
• /
• 2020

논문 초록은 논문의 내용을 요약해 제시함으로써 독자들의 연구결과물에 대한 빠른 검색과 이해를 도모한다. 초록의 구성은 대부분 전형적인 경우가 많기 때문에, 초록의 구조를 자동 분석하여 색인해두면 유사구조 초록을 검색하거나 생성하는 등의 연구효율화에 기여할 수 있다. 허세훈 외 (2019)는 초록 자동구조화를 위한 말뭉치 SPA2019 및 기계학습기반의 자동구조화 방법을 제시하였다. 본 연구는, 기존 SPA2019 의 구조화 오류를 바로잡고, SPA2019 에서 추출한 1,346 개의 초록데이터와 2,385 개의 초록데이터를 추가한 SPA2020 말뭉치를 새로이 소개한다. 또한, 다양한 선학습 기반 트랜스포머들을 활용하여 초록 자동구조화를 수행하였으며, 그 결과 BERT-0.86%, RoBERTa-0.86%, ALBERT-0.84%, XLNet-0.86%, DistilBERT-0.85% 등의 자동구조화 성능을 보임을 확인하였다.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.135-140
• /
• 2020

사전 학습된 Transformer 기반 언어 모델은 자연어처리 시스템에 적용되었을 시 광범위한 사례에서 큰 폭의 성능 향상을 보여준다. 여기서 사전 학습에 사용되는 언어 모델링 태스크는 비지도 학습에 속하는 기술이기 때문에 상대적으로 데이터의 확보가 쉬운 편이다. 하지만 몇 종의 주류 언어를 제외한 대부분 언어는 활용할 수 있는 언어 자원 자체가 희소하며, 따라서 이러한 사전 학습 기술의 혜택도 누리기 어렵다. 본 연구에서는 이와 같은 상황에서 발생할 수 있는 자원 희소성 문제를 극복하기 위해 이종 언어 간 전이 학습을 이용하는 방법을 제안한다. 본 방법은 언어 자원이 풍부한 언어에서 학습된 Transformer 기반 언어 모델에서 얻은 파라미터 중 재활용 가능한 부분을 이용하여 목표 언어의 모델을 초기화한 후 학습을 진행한다. 또한, 기존 언어와 목표 언어의 차이를 학습하는 역할을 하는 적응층들을 추가하여 이종 언어 간 전이 학습을 돕는다. 제안된 방법을 언어 자원이 희귀한 상황에 대하여 실험해본 결과, 전이 학습을 사용하지 않은 기준 모델 대비 perplexity와 단어 예측의 정확도가 큰 폭으로 향상됨을 확인하였다.

• Annual Conference of KIPS
• /
• 2021.11a
• /
• pp.749-752
• /
• 2021

오늘날 인공지능(Artificial Intelligence, AI) 분야에서 가장 구현하기 어려운 분야 중 하나는 추론이다. 근래 추론 분야에서 영상과 언어가 결합한 다중 모드(Multi-modal) 환경에서 영상 기반의 질의 응답(Visual Question Answering, VQA) 과업에 대한 AI 모델이 발표됐다. 얼마 지나지 않아 VQA 모델의 성능을 개선한 GVQA(Grounded Visual Question Answering) 모델도 발표됐다. 하지만 아직 GVQA 모델도 완벽한 성능을 내진 못한다. 본 논문에서는 GVQA 모델의 성능 개선을 위해 VCC(Visual Concept Classifier) 모델을 ViT-G(Vision Transformer-Giant)/14로 변경하고, ACP(Answer Cluster Predictor) 모델을 GPT(Generative Pretrained Transformer)-3으로 변경한다. 이와 같은 방법들은 성능을 개선하는 데 큰 도움이 될 수 있다고 사료된다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.222-224
• /
• 2022

Deep learning-based computer vision applications in the field of object segmentation take a transfer learning method using hyperparameters and models pretrained and distributed by STOA techniques to improve performance. Custom datasets used in this process require a lot of resources, such as time and labeling, in labeling tasks to generate Ground Truth information. In this paper, we present an automatic labeling construction method using object segmentation techniques so that resources such as time and labeling can be used less to build custom datasets used in deep learning neural networks.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.711-715
• /
• 2023

본 연구는 대부분의 언어 모델이 사용하고 있는 서브워드 토큰화 과정을 거치지 않고, 바이트 단위의 인코딩을 그대로 다룰 수 있는 토큰-프리 사전학습 언어모델에 대한 것이다. 토큰-프리 언어모델은 명시적인 미등록어 토큰이 존재하지 않고, 전 처리 과정이 단순하며 다양한 언어 및 표현 체계에 대응할 수 있는 장점이 있다. 하지만 관련 연구가 미흡, 서브워드 모델에 대비해 학습이 어렵고 낮은 성능이 보고되어 왔다. 본 연구에서는 한국어를 중심으로 토큰-프리 언어 이해-생성 모델을 사전 학습 후, 서브워드 기반 모델과 비교하여 가능성을 살펴본다. 또한, 토큰 프리 언어모델에서 지적되는 과도한 연산량을 감소시킬 수 있는 그래디언트 기반 서브워드 토크나이저를 적용, 처리 속도를 학습 2.7배, 추론 1.46배 개선하였다.

• Korean Journal of Radiology
• /
• v.24 no.10
• /
• pp.952-959
• /
• 2023

Large language models (LLMs) such as ChatGPT have garnered considerable interest for their potential to aid non-native English-speaking researchers. These models can function as personal, round-the-clock English tutors, akin to how Prometheus in Greek mythology bestowed fire upon humans for their advancement. LLMs can be particularly helpful for non-native researchers in writing the Introduction and Discussion sections of manuscripts, where they often encounter challenges. However, using LLMs to generate text for research manuscripts entails concerns such as hallucination, plagiarism, and privacy issues; to mitigate these risks, authors should verify the accuracy of generated content, employ text similarity detectors, and avoid inputting sensitive information into their prompts. Consequently, it may be more prudent to utilize LLMs for editing and refining text rather than generating large portions of text. Journal policies concerning the use of LLMs vary, but transparency in disclosing artificial intelligence tool usage is emphasized. This paper aims to summarize how LLMs can lower the barrier to academic writing in English, enabling researchers to concentrate on domain-specific research, provided they are used responsibly and cautiously.

• ETRI Journal
• /
• v.46 no.4
• /
• pp.671-682
• /
• 2024

Violence can be committed anywhere, even in crowded places. It is hence necessary to monitor human activities for public safety. Surveillance cameras can monitor surrounding activities but require human assistance to continuously monitor every incident. Automatic violence detection is needed for early warning and fast response. However, such automation is still challenging because of low video resolution and blind spots. This paper uses ResNet50v2 and the gated recurrent unit (GRU) algorithm to detect violence in the Movies, Hockey, and Crowd video datasets. Spatial features were extracted from each frame sequence of the video using a pretrained model from ResNet50V2, which was then classified using the optimal trained model on the GRU architecture. The experimental results were then compared with wavelet feature extraction methods and classification models, such as the convolutional neural network and long short-term memory. The results show that the proposed combination of ResNet50V2 and GRU is robust and delivers the best performance in terms of accuracy, recall, precision, and F1-score. The use of ResNet50V2 for feature extraction can improve model performance.

• Wind and Structures
• /
• v.37 no.5
• /
• pp.383-397
• /
• 2023

In the Structural Health Monitoring (SHM) system of civil engineering, data missing inevitably occurs during the data acquisition and transmission process, which brings great difficulties to data analysis and poses challenges to structural health monitoring. In this paper, Convolution Neural Network (CNN) is used to recover the nonstationary wind speed data missing randomly at sampling points. Given the technical constraints and financial implications, field monitoring data samples are often insufficient to train a deep learning model for the task at hand. Thus, simulation combined transfer learning strategy is proposed to address issues of overfitting and instability of the deep learning model caused by the paucity of training samples. According to a portion of target data samples, a substantial quantity of simulated data consistent with the characteristics of target data can be obtained by nonstationary wind-field simulation and are subsequently deployed for training an auxiliary CNN model. Afterwards, parameters of the pretrained auxiliary model are transferred to the target model as initial parameters, greatly enhancing training efficiency for the target task. Simulation synergy strategy effectively promotes the accuracy and stability of the target model to a great extent. Finally, the structural dynamic response analysis verifies the efficiency of the simulation synergy strategy.

• Imaging Science in Dentistry
• /
• v.50 no.2
• /
• pp.169-174
• /
• 2020

Purpose: Periodontal disease causes tooth loss and is associated with cardiovascular diseases, diabetes, and rheumatoid arthritis. The present study proposes using a deep learning-based object detection method to identify periodontally compromised teeth on digital panoramic radiographs. A faster regional convolutional neural network (faster R-CNN) which is a state-of-the-art deep detection network, was adapted from the natural image domain using a small annotated clinical data- set. Materials and Methods: In total, 100 digital panoramic radiographs of periodontally compromised patients were retrospectively collected from our hospital's information system and augmented. The periodontally compromised teeth found in each image were annotated by experts in periodontology to obtain the ground truth. The Keras library, which is written in Python, was used to train and test the model on a single NVidia 1080Ti GPU. The faster R-CNN model used a pretrained ResNet architecture. Results: The average precision rate of 0.81 demonstrated that there was a significant region of overlap between the predicted regions and the ground truth. The average recall rate of 0.80 showed that the periodontally compromised teeth regions generated by the detection method excluded healthiest teeth areas. In addition, the model achieved a sensitivity of 0.84, a specificity of 0.88 and an F-measure of 0.81. Conclusion: The faster R-CNN trained on a limited amount of labeled imaging data performed satisfactorily in detecting periodontally compromised teeth. The application of a faster R-CNN to assist in the detection of periodontally compromised teeth may reduce diagnostic effort by saving assessment time and allowing automated screening documentation.