Intelligent Attribution of Engineering Contract Claims and Quantification of Economic Losses based on Large Language Models

The prevailing need for advanced techniques for claim identification and measurement of financial losses arises from the increasing intricacy of engineering contracts as well as the regularity of disputes in large-scale initiatives. This work introduces a method based on large language models (LLMs) for the intelligent attribution of engineering contract claim events and quantification of the associated economic losses. Using existing large language models as a basis of our approach, we present a more holistic framework that involves natural language processing (NLP), machine learning algorithms, and domain knowledge to help analyze contract documents, ascertain potential claim events, and determine who will typically take the blame for such claims—all founded on context and patterns present in language. On the basis of the existing methods, the model innovatively puts forward the multi-layer attention mechanism and the deep reinforcement learning component for dynamically adjusting the loss and the fusion of mixed knowledge graph, and then improves the accuracy of claim classification and economic loss estimation. The experimental results show that the proposed model outperforms the traditional rule-based system and the previously mentioned machine learning models.