IDSL Lab

The Industrial Data Science Lab (IDSL) at Hanyang University conducts research at the intersection of data science and various industries, with a strong emphasis on Natural Language Processing (NLP) and Large Language Models (LLM). Our lab explores data-driven methodologies to address real-world industrial challenges, leveraging LLMs to enhance text understanding, knowledge extraction, and automated decision-making.

Large Language Models

We conduct a wide range of research related to natural language processing, including benchmark dataset creation, algorithm research, and applications. In particular, we have successfully conducted studies on deep learning-based intent classification and generation models for improving reasoning and comprehension performance on complex queries, as well as the development of human-like dialogue systems. By advancing natural language understanding and generation technologies using LLMs, we are achieving tangible results across various application domains.

Industrial Data Analytics (Industrial NLP)

To solve real-world industrial problems, we design and implement AI solutions that reflect the specific characteristics of industrial settings based on diverse industry data. Our focus is particularly on solving industrial challenges using NLP, and we have developed systems applicable to sectors such as finance and healthcare, which involve sensitive data. We have successfully developed solutions such as anomaly detection based on heterogeneous data, the Police Agency’s PolBot, and an ESG information support framework. By building datasets that meet industry demands and ensuring the safe utilization of data, we have produced numerous real-world industrial applications.

Financial & ESG

We conduct AI solution research specialized in the financial and ESG domains, developing practical solutions for enterprises. By utilizing NLP techniques, we analyze ESG factors and use this analysis to build frameworks for corporate performance management, production process optimization, and ESG information support for businesses. This enables companies to effectively utilize ESG-related data and respond strategically. Additionally, we have developed a framework optimized for insurance documents, enhancing the understanding of insurance-related materials and improving user convenience.

GRL