• Journal of the Korea Society of Computer and Information
• /
• v.29 no.5
• /
• pp.31-38
• /
• 2024

In this paper, we propose an ASAP(AI-driven Service for Advertisement Production) application that provides a generative AI-based automatic advertising image production service. This application utilizes GPT-3.5 Turbo Instruct to generate suitable background mood and promotional copy based on user-entered keywords. It utilizes OpenAI's DALL·E 3 model and Stability AI's SDXL model to generate background images and text images based on these inputs. Furthermore, OCR technology is employed to improve the accuracy of text images, and all generated outputs are synthesized to create the final advertisement. Additionally, using the PILLOW and OpenCV libraries, text boxes are implemented to insert details such as phone numbers and business hours at the edges of promotional materials. This application offers small business owners who face difficulties in advertising production a simple and cost-effective solution.

• Journal of the Korea Computer Graphics Society
• /
• v.30 no.3
• /
• pp.149-158
• /
• 2024

Photo booths have traditionally provided a fun and easy way to capture and print photos to cherish memories. These booths allow individuals to capture their desired poses and props, sharing memories with friends and family. To enable diverse expressions, generative AI-powered photo booths have emerged. However, existing AI photo booths face challenges such as difficulty in taking group photos, inability to accurately reflect user's poses, and the challenge of applying different concepts to individual subjects. To tackle these issues, we present CINEMAPIC, a photo booth system that allows users to freely choose poses, positions, and concepts for their photos. The system workflow includes three main steps: pre-processing, generation, and post-processing to apply individualized concepts. To produce high-quality group photos, the system generates a transparent image for each character and enhances the backdrop-composited image through a small number of denoising steps. The workflow is accelerated by applying an optimized diffusion model and GPU parallelization. The system was implemented as a prototype, and its effectiveness was validated through a user study and a large-scale pilot operation involving approximately 400 users. The results showed a significant preference for the proposed system over existing methods, confirming its potential for real-world photo booth applications. The proposed CINEMAPIC photo booth is expected to lead the way in a more creative and differentiated market, with potential for widespread application in various fields.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.2
• /
• pp.397-402
• /
• 2024

Speech recognition and motion recognition technologies are applied and used in various smart devices, but they are composed of simple command recognition forms and are used as simple functions. Apart from simple functions for recognition data, professional command execution capabilities are required based on data learned in various fields. Research is being conducted on a system platform that provides optimal data to users using Generative AI, which is currently competing around the world, and can interact through voice recognition and motion recognition. The main technical processes designed for this study were designed using technologies such as voice and motion recognition functions, application of AI technology, and two-way communication. In this paper, two-way communication between a device and a user can be achieved by various input methods through voice recognition and motion recognition technology applied with AI technology.

• Journal of Platform Technology
• /
• v.11 no.6
• /
• pp.34-46
• /
• 2023

The design of non-pneumatic tires, which are created by filling the space between the wheel and the tread with elastomeric compounds or polygonal spokes, has become an important research topic in the automotive and aerospace industries. In this study, a system was designed for the design of non-pneumatic tires through the implementation of a generative adversarial network. We specifically examined factors that could impact the design, including the type of non-pneumatic tire, its intended usage environment, manufacturing techniques, distinctions from pneumatic tires, and how spoke design affects load distribution. Using OpenCV, various shapes and spoke configurations were generated as images, and a GAN model was trained on the projected GANs to generate shapes and spokes for non-pneumatic tire designs. The designed non-pneumatic tires were labeled as available or not, and a Vision Transformer image classification AI model was trained on these labels for classification purposes. Evaluation of the classification model show convergence to a near-zero loss and a 99% accuracy rate confirming the generation of non-pneumatic tire designs.

• Transactions on Semiconductor Engineering
• /
• v.1 no.1
• /
• pp.23-30
• /
• 2023

With the recent development of Generative AI technology by IT giants, the size of the transformer model is increasing exponentially over trillion won. In order to continuously enable these AI services, it is essential to reduce the weight of the model. In this paper, we find a hardware-friendly structured pruning pattern and propose a lightweight method of the transformer model. Since compression proceeds by utilizing the characteristics of the model algorithm, the size of the model can be reduced and performance can be maintained as much as possible. Experiments show that the structured pruning proposed when pruning GPT-2 and BERT language models shows almost similar performance to fine-grained pruning even in highly sparse regions. This approach reduces model parameters by 80% and allows hardware acceleration in structured form with 0.003% accuracy loss compared to fine-tuned pruning.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.75-79
• /
• 2023

문서에서 테이블은 중요한 정보들을 축약하여 모아 놓은 정보 집합체라고 할 수 있다. 이러한 테이블을 대상으로 질의응답하는 테이블 질의응답 기술이 연구되고 있으며, 이 중 언어모델을 이용한 연구가 좋은 결과를 보이고 있다. 본 연구에서는 최근 주목받고 있는 생성형 언어모델 기술을 테이블 질의응답에 적용하여 언어모델과 프롬프트의 변경에 따른 결과를 살펴보고, 단답형 정답과 생성형 결과의 특성에 적합한 평가방법으로 측정해 보았다. 자체 개발한 EXAONE 1.7B 모델의 경우 KorWiki 데이터셋에 대해 적용하여 EM 92.49, F1 94.81의 결과를 얻었으며, 이를 통해 작은 크기의 모델을 파인튜닝하여 GPT-4와 같은 초거대 모델보다 좋은 성능을 보일 수 있음을 확인하였다.

• International journal of advanced smart convergence
• /
• v.12 no.4
• /
• pp.154-163
• /
• 2023

With the rapid development of AI technology, ChatGPT and other AI content creation tools are becoming common, and users are becoming curious and adopting them. These tools, unlike search engines, generate results based on user prompts, which puts them at risk of inaccuracy or plagiarism. This allows unethical users to create inappropriate content and poses greater educational and corporate data security concerns. AI content detection is needed and AI-generated text needs to be identified to address misinformation and trust issues. Along with the positive use of AI tools, monitoring and regulation of their ethical use is essential. When detecting content created by AI with an AI content detection tool, it can be used efficiently by using the appropriate tool depending on the usage environment and purpose. In this paper, we collect data on AI content detection tools and compare and analyze the functions and characteristics of AI content detection tools to help meet these needs.

• Journal of the Korea Convergence Society
• /
• v.13 no.1
• /
• pp.13-21
• /
• 2022

Recent deep learning-based research shows excellent performance in most natural language processing (NLP) fields with pre-trained language models. In particular, the auto-encoder-based language model proves its excellent performance and usefulness in various fields of Korean language understanding. However, the decoder-based Korean generative model even suffers from generating simple sentences. Also, there is few detailed research and data for the field of conversation where generative models are most commonly utilized. Therefore, this paper constructs multi-turn dialogue data for a Korean generative model. In addition, we compare and analyze the performance by improving the dialogue ability of the generative model through transfer learning. In addition, we propose a method of supplementing the insufficient dialogue generation ability of the model by extracting recommended response candidates from external knowledge information through a retrival model.

• Electronics and Telecommunications Trends
• /
• v.38 no.5
• /
• pp.1-11
• /
• 2023

The emergence of generative hyperscale artificial intelligence (AI) has enabled new services, such as image-generating AI and conversational AI based on large language models. Such services likely lead to the influx of numerous users, who cannot be handled using conventional AI models. Furthermore, the exponential increase in training data, computations, and high user demand of AI models has led to intensive hardware resource consumption, highlighting the need to develop domain-specific semiconductors for hyperscale AI. In this technical report, we describe development trends in technologies for hyperscale AI processors pursued by domestic and foreign semiconductor companies, such as NVIDIA, Graphcore, Tesla, Google, Meta, SAPEON, FuriosaAI, and Rebellions.

• International Journal of Computer Science & Network Security
• /
• v.23 no.4
• /
• pp.15-24
• /
• 2023

One of the objectives and aspirations of scientists and engineers ever since the development of computers has been to interact naturally with machines. Hence features of artificial intelligence (AI) like natural language processing and natural language generation were developed. The field of AI that is thought to be expanding the fastest is interactive conversational systems. Numerous businesses have created various Virtual Personal Assistants (VPAs) using these technologies, including Apple's Siri, Amazon's Alexa, and Google Assistant, among others. Even though many chatbots have been introduced through the years to diagnose or treat psychological disorders, we are yet to have a user-friendly chatbot available. A smart generative cognitive behavioral therapy with spoken dialogue systems support was then developed using a model Persona Perception (P2) bot with Generative Pre-trained Transformer-2 (GPT-2). The model was then implemented using modern technologies in VPAs like voice recognition, Natural Language Understanding (NLU), and text-to-speech. This system is a magnificent device to help with voice-based systems because it can have therapeutic discussions with the users utilizing text and vocal interactive user experience.