Article
An Automated Summarization Framework for Complex Financial and Legal Texts
AI-powered framework designed to summarize complex legal and financial documents using modern Natural Language Processing techniques. The system combines two advanced models, COLBERT and T5, to improve both the relevance and readability of generated summaries. COLBERT is used to understand the contextual meaning of the document and rank important sentences effectively, while T5 generates concise and meaningful abstractive summaries from the selected content. Unlike traditional extractive methods, this approach focuses on capturing the true intent of the text and expressing it in a simplified manner. The model was trained using a dataset consisting of legal and financial documents along with their summaries. Due to computational limitations in Google Collab, a smaller dataset was used, containing 2,000 training samples and 300 testing samples. Despite these constraints, the system achieved a Rouge-1 score of 30%, indicating promising performance. The results highlight that even with limited resources, transformer-based models can produce useful summaries for domain-specific texts. This framework offers a scalable and cost-effective solution, with potential for significant improvement when trained on larger datasets & powerful hardware.
Full Text Attachment
