• The Korean Society of Law and Medicine
• /
• v.24 no.4
• /
• pp.67-101
• /
• 2023

The utilization of generative AI in the medical field is also being rapidly researched. Access to vast data sets reduces the time and energy spent in selecting information. However, as the effort put into content creation decreases, there is a greater likelihood of associated issues arising. For example, with generative AI, users must discern the accuracy of results themselves, as these AIs learn from data within a set period and generate outcomes. While the answers may appear plausible, their sources are often unclear, making it challenging to determine their veracity. Additionally, the possibility of presenting results from a biased or distorted perspective cannot be discounted at present on ethical grounds. Despite these concerns, the field of generative AI is continually advancing, with an increasing number of users leveraging it in various sectors, including biomedical and life sciences. This raises important legal considerations regarding who bears responsibility and to what extent for any damages caused by these high-performance AI algorithms. A general overview of issues with generative AI includes those discussed above, but another perspective arises from its fundamental nature as a large-scale language model ('LLM') AI. There is a civil law concern regarding "the memorization of training data within artificial neural networks and its subsequent reproduction". Medical data, by nature, often reflects personal characteristics of patients, potentially leading to issues such as the regeneration of personal information. The extensive application of generative AI in scenarios beyond traditional AI brings forth the possibility of legal challenges that cannot be ignored. Upon examining the technical characteristics of generative AI and focusing on legal issues, especially concerning the protection of personal information, it's evident that current laws regarding personal information protection, particularly in the context of health and medical data utilization, are inadequate. These laws provide processes for anonymizing and de-identification, specific personal information but fall short when generative AI is applied as software in medical devices. To address the functionalities of generative AI in clinical software, a reevaluation and adjustment of existing laws for the protection of personal information are imperative.

• 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.

• 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.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.936-937
• /
• 2023

본 논문에서는 LLM(Large Language Model) 모델의 fine-tuning 을 통한, 기초 수리 서술형 문항 풀이용 모델 및 Dall-E2 등 이미지 생성형 모델을 활용한 따른 영어 퀴즈풀이용 이미지 생성형 모델을 생성하여, 한국어 기반 LLM 자체 모델 학습 및 교육용 이미지 생성에 대한 방법을 고찰하였다.

• Korean Journal of Cognitive Science
• /
• v.34 no.3
• /
• pp.197-226
• /
• 2023

Generative Text Summarization is one of the Natural Language Processing tasks. It generates a short abbreviated summary while preserving the content of the long text. ROUGE is a widely used lexical-overlap based metric for text summarization models in generative summarization benchmarks. Although it shows very high performance, the studies report that 30% of the generated summary and the text are still inconsistent. This paper proposes a methodology for evaluating the performance of the summary model without using the correct summary. AggreFACT is a human-annotated dataset that classifies the types of errors in neural text summarization models. Among all the test candidates, the two cases, generation summary, and when errors occurred throughout the summary showed the highest correlation results. We observed that the proposed evaluation score showed a high correlation with models finetuned with BART and PEGASUS, which is pretrained with a large-scale Transformer structure.

• Journal of Information Processing Systems
• /
• v.18 no.5
• /
• pp.650-664
• /
• 2022

Mobile applications can be easily downloaded and installed via markets. However, malware and malicious applications containing unwanted advertisements exist in these application markets. Therefore, smartphone users install applications with reference to the application review to avoid such malicious applications. An application review typically comprises contents for evaluation; however, a false review with a specific purpose can be included. Such false reviews are known as fake reviews, and they can be generated using artificial intelligence (AI)-based text-generating models. Recently, AI-based text-generating models have been developed rapidly and demonstrate high-quality generated texts. Herein, we analyze the features of fake reviews generated from Generative Pre-Training-2 (GPT-2), an AI-based text-generating model and create a model to detect those fake reviews. First, we collect a real human-written application review from Kaggle. Subsequently, we identify features of the fake review using natural language processing and statistical analysis. Next, we generate fake review detection models using five types of machine-learning models trained using identified features. In terms of the performances of the fake review detection models, we achieved average F1-scores of 0.738, 0.723, and 0.730 for the fake review, real review, and overall classifications, respectively.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.11
• /
• pp.481-492
• /
• 2023

This study explores how to build a Korean dataset to extract information from text using generative large language models. In modern society, mixed information circulates rapidly, and effectively categorizing and extracting it is crucial to the decision-making process. However, there is still a lack of Korean datasets for training. To overcome this, this study attempts to extract information using text-based zero-shot learning using a generative large language model to build a purposeful Korean dataset. In this study, the language model is instructed to output the desired result through prompt engineering in the form of "system"-"instruction"-"source input"-"output format", and the dataset is built by utilizing the in-context learning characteristics of the language model through input sentences. We validate our approach by comparing the generated dataset with the existing benchmark dataset, and achieve 25.47% higher performance compared to the KLUE-RoBERTa-large model for the relation information extraction task. The results of this study are expected to contribute to AI research by showing the feasibility of extracting knowledge elements from Korean text. Furthermore, this methodology can be utilized for various fields and purposes, and has potential for building various Korean datasets.

• Annual Conference on Human and Language Technology
• /
• 2022.10a
• /
• pp.88-91
• /
• 2022

디지털 휴먼, 민원 상담, ARS 등 칫챗의 활용과 수요가 증가함에 따라 칫챗의 성능 향상을 위한 다양한 연구가 진행되고 있다. 특히, 오토 인코더(Auto-encoder) 기반의 생성 모델(Generative Model)은 높은 성능을 보이며 지속적인 연구가 이루어지고 있으나, 이전 대화들에 대한 충분한 문맥 정보의 반영이 어렵고 문법적으로 부적절한 답변을 생성하는 문제가 있다. 이를 개선하기 위해 검색 기반의 생성 모델과 관련된 연구가 진행되고 있으나, 현재 시점의 문장이 유사해도 이전 문장들에 따라 의도와 답변이 달라지는 멀티턴 대화 특징을 반영하여 대화를 검색하는 연구가 부족하다. 본 논문에서는 이와 같은 멀티턴 대화의 특징이 고려된 검색 방법을 제안하고 검색된 레퍼런스(준정답 문장)를 멀티턴 대화와 함께 생성 모델의 입력으로 활용하여 학습시키는 방안을 제안한다. 제안 방안으로 학습된 발화 생성 모델은 기존 모델과 비교 평가를 수행하며 Rouge-1 스코어에서 13.11점, Rouge-2 스코어에서 10.09점 Rouge-L 스코어에서 13.2점 향상된 성능을 보였고 이를 통해 제안 방안의 우수성을 입증하였다.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.586-590
• /
• 2023

인터넷과 소셜미디어 사용량의 급증으로, 제품 리뷰, 온라인 피드백, 소셜 미디어 게시물 등을 통해 고객의 감정을 파악하는 것이 중요해졌다. 인공지능이 활용되어 고객이 제품이나 서비스의 어떤 부분에 만족하거나 불만을 가지는지를 분석하는 연구를 ABSA라고 하며 이미 해외에서는 이런 연구가 활발하게 이루어지는 반면, 국내에서는 상대적으로 부족한 상황이다. 이 연구에서는 ABSA의 두 개의 주요 작업인 ACD와 ASC에 대해 생성 모델 중 하나인 T5 모델을 사용하는 방법론을 제시한다. 이 방법론은 기존 판별 모델을 사용하는 것에 비해 시간과 성능 측면에서 크게 향상되었음을 보여준다.

• International journal of advanced smart convergence
• /
• v.13 no.1
• /
• pp.99-107
• /
• 2024

In AI image generation tools, most general users must use an effective prompt to craft queries or statements to elicit the desired response (image, result) from the AI model. But we are software engineers who focus on software processes. At the process's early stage, we use informal and formal requirement specifications. At this time, we adapt the natural language approach into requirement engineering and toon engineering. Most Generative AI tools do not produce the same image in the same query. The reason is that the same data asset is not used for the same query. To solve this problem, we intend to use informal requirement engineering and linguistics to create a toon. Therefore, we propose a sequence diagram and image generation mechanism by analyzing and applying key objects and attributes as an informal natural language requirement analysis. Identify morpheme and semantic roles by analyzing natural language through linguistic methods. Based on the analysis results, a sequence diagram and an image are generated through the diagram. We expect consistent image generation using the same image element asset through the proposed mechanism.