Towards Comprehensive Innovation Landscape: Technology Retrieval Meets Large Language Models

Hien Thu Pham, Minh Hieu Nguyen, Hiep Minh Ha, Ngoc Quang Hung Le, Jun Jo

Abstract: In modern dynamic business environment, understanding the technologies companies employ is vital for creating business relationships, identifying market opportunities, and shaping strategic decisions. Traditional technology mapping methods, which rely on keyword-based approaches, face limitations in processing large, diverse datasets and often struggle to detect emerging technologies. To address these challenges, we introduce a novel framework called STARS (Semantic Technology and Retrieval System). STARS leverages Large Language Models (LLMs) and Sentence-BERT to extract relevant technologies from unstructured data, generate comprehensive company profiles, and rank technologies based on their relevance to each company operations. By integrating entity extraction with Chain-of-Thought prompting, and employing semantic ranking, STARS effectively maps companies’ technological portfolios. Our experimental results demonstrate that STARS significantly improves retrieval precision, offering a robust and scalable solution for mapping technologies across industries.

External IDs:dblp:conf/adc/PhamNHLJ24