DC25_PAPER_TRACK14_LargeLanguageModelEnabledEngineering_Kumar_V3.pdf

1、 Information Classification:General Large Language Model Enabled Engineering Code Generation Using Novel Data Processing and Augmentation Flow Akhilesh Kumar,Ansys Norman Chang,Ansys Yu-Chen Lin,Ansys Information Classification:General Wenliang Zhang,Ansys Muhammad Zakir,Ansys Rucha Apte,Nvidia ruch

2、aaumich.edu Haiyang He,Ansys Jyh-Shing Roger Jang,National Taiwan University jangcsie.ntu.edu.tw Information Classification:General Abstract This work describes a new methodology to augment the capabilities of Large Language Models(LLMs)for generating domain-specific engineering application code as

3、follows:(i)leveraging LLM-based data splitting and data renovation techniques to refine the semantic representation within the embedding space;(ii)proposing an effective method for refactoring existing scripts,enabling the generation of new and high-quality scripts with the aid of LLMs;(iii)developi

4、ng the Implicit Knowledge Expansion and Contemplation(IKEC)Prompt technique;and(iv)showcasing the efficacy of our data pre-processing approach through a case study using engineering simulation software RedHawk-SC.Our contributions collectively advance the Retrieval-Augmented Generation(RAG)framework

5、,enabling more relevant and precise information retrieval for the downstream reasoning and planning LLM agent.An arena-style evaluation by 28 domain experts and 182 votes confirms the significant effectiveness of our methods.Notably,our approach achieves up to 1.43 times the improvement in code gene

6、ration for MapReduce applications compared to the Chain-of-Thought(CoT)technique.Authors Biographies Akhilesh Kumar Akhilesh Kumar is a Senior Principal R&D Engineer at Ansys leading the AI/ML and Generative AI solutions for the Semiconductor BU products.He has deep experience in developing advanced