Presentation Information
[5G3-OS-37b-06]Towards Effective Integration of Generative AI in Knowledge Intensive ConsultingEvaluation of Knowledge Graphs and Retrieval Augmented Generation for Domain Adaptation
〇Noriaki Hirokawa1 (1. KPMG Advisory Lighthouse)
Keywords:
Knowledge Graph,Retrieval-Augmented Generation,Domain Adaptation,Generative AI
Consulting firms accumulate and integrate specialized knowledge across industries and functional domains. However, because such knowledge is organized for consultants’ understanding and practical application, it tends to be densely aggregated, and no simple proportional relationship exists between its comprehensiveness and the number of tokens required to represent it.
Considering this characteristic, I examine the applicability of knowledge graphs and Retrieval-Augmented Generation for the effective integration of generative AI into consulting practices.
Empirical evaluations conducted across multiple domains revealed no issues specific to the Japanese language. Nevertheless, even with state-of-the-art methods, limitations remain regarding the accuracy of knowledge graphs and the interpretive capabilities of large language models.
Considering this characteristic, I examine the applicability of knowledge graphs and Retrieval-Augmented Generation for the effective integration of generative AI into consulting practices.
Empirical evaluations conducted across multiple domains revealed no issues specific to the Japanese language. Nevertheless, even with state-of-the-art methods, limitations remain regarding the accuracy of knowledge graphs and the interpretive capabilities of large language models.
Comment
To browse or post comments, you must log in.Log in