Enhance Presents Research Papers at 2 International AI Conferences

## Commerce AI Startup Enhance Dominates Global AI Academic Conferences, Presenting Innovative Solutions Enhance, a specialized commerce AI agent startup, has demonstrated its technological prowess by publishing two groundbreaking research papers at two of the most prestigious international academic conferences in the field of artificial intelligence. These papers present innovative approaches that solve the problem of cultural bias in LLMs (Large Language Models) and dramatically improve information retrieval performance, attracting significant attention from academia. Enhance presented its paper 'ValuesRAG: Enhancing Cultural Alignment through Retrieval-Augmented Contextual Learning,' co-authored with Peking University and Tsinghua University, at the '8th AAAI/ACM Conference on AI, Ethics, and Society (AAAI-AIES 2025),' a prestigious conference in the AI field. This paper proposes a unique framework that summarizes individuals' cultural values and demographic information, and by combining Retrieval-Augmented Generation (RAG) with context-based generation, enables LLM responses to encompass diverse cultural backgrounds. This demonstrates Enhance's deep insights into the important topic of AI's ethical responsibility and respect for diversity. Furthermore, at the 'SIGKDD 2025-AI Agent for Information Retrieval (Agent4IR) workshop,' the world's largest data mining conference, Enhance introduced 'AMD (Agent-Mediated Dialogic Inquiry),' a multi-agent-based framework for improving information retrieval performance. This paper, titled 'A Novel Query Expansion Technique Utilizing Conversational Structures of AI Agents,' through a three-stage conversational structure of a Socratic questioning agent, a conversational response agent, and a reflective feedback agent, provides more precise and rich answers to initial search queries. Each agent uses an LLM to perform question generation, response, and review, an innovative achievement that significantly improves search accuracy and diversity compared to existing methods. This research achievement, led by Enhance's AI researcher Won-deok Seo as the first author, is expected to lead to the technological advancement of Enhance's core services, CommerceOS and AIAP (AI Agent Platform). CEO Seung-hyun Lee stated, "Through these academic achievements, we have proven that Enhance's AI agent technology can expand beyond commerce into broad areas such as information retrieval and the resolution of cultural bias." He added, "Through continuous research with leading overseas universities, including MIT and Yale, we aim to lead next-generation AI agent technology and strengthen the technological competitiveness of Korean companies in the global market." Established in 2021, Enhance is an AI agent solution company based on LAM (Large Action Model). While existing LLMs focused on language generation, LAM possesses differentiated technological capabilities that autonomously perform actual tasks and contribute to revenue growth. Its potential was already recognized when it was selected for Palantir's 'Startup Fellowship' last May.