Nowadays, the volume of textual data being generated is unprecedented. From social media posts, news articles, and academic papers to customer reviews, emails, and business documents, the sheer quantity of text data is growing exponentially. Traditional methods of analyzing this vast amount of data often fall short in terms of scalability, accuracy, and efficiency.



In this context, Generative AI (GenAI) is revolutionizing the field of Natural Language Processing (NLP) by enabling the creation of highly sophisticated Large Language Models (LLMs) that can generate, understand, and manipulate human language. GenAI models like GPT-4, Gemini 1.5 pro, Mistral Large and BERT are at the forefront of these advancements from chatbots to automated content creation. This workshop aims to provide participants with a deep understanding of LLMs, its applications in NLP, and the ethical considerations involved. GenAI models are designed to handle and process enormous datasets, making them ideal for textual document analysis. These models leverage advanced machine learning techniques to understand, interpret, and generate human-like text, allowing for more nuanced and comprehensive analysis. By using LLMs, we can uncover insights and patterns that would be impossible to detect using conventional methods.

This workshop is designed to provide a comprehensive understanding of how LLMs can be leveraged for textual document analysis. Participants will gain hands-on experience and theoretical knowledge about the applications, capabilities, and limitations of GenAI models in the context of analyzing textual data. The workshop will cover various techniques and tools, practical implementation, and the latest advancements in the field.

Novelty for this edition:

After the success of GENAIDOC 2024 at FLLM’24, in this edition of GENAIDOC will further explore a fast-emerging yet underformalized area at the intersection of large language models (LLMs), document understanding, and multimodal AI. The proposed extension focuses on the design of effective prompting strategies for extracting, interpreting, and aligning textual and visual information from real-world documents such as scanned PDFs, structured forms, and tabular data.

The new focus will delve into the design of prompting strategies that enable effective extraction, interpretation, and alignment of textual and visual elements from scanned documents, PDFs, structured forms, and tabular data. This includes analyzing how prompts influence the performance of LLMs when processing OCR-based content, determining the best approaches for handling visual structures such as tables and layout elements, and studying how textual and visual modalities can be integrated or separated during inference. The topic will also cover the development of standardized prompt templates adapted to industrial contexts such as invoice and document processing. This direction aims to combine technical depth with real-world applicability, enhancing both academic and practical contributions to the field.

The GENAIDOC workshop aims to bring together an area for experts from industry, science, and academia to exchange ideas and discuss ongoing research in natural language processing and GenAI for textual document analysis.