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.
Topics of interests:
This workshop invites submissions with high-quality works that are related, but are not limited, to the topics below:
Prompts to extract textual information
Prompts to extract visual information
Prompts to extract data from tables
Prompts for specific documents type
Prompts to classify documents
Text classification
Automatic document summarization
Automatic machine translation
Sentiment analysis
Text generation
Deep learning for NLP
Reinforcement Learning for NLP
Unsupervised Learning for NLP
Speaker identification
Speech recognition
Speech to Text
Text detection and recognition from images
Question Answering systems
Transfer Learning for NLP
Active Learning for NLP
Real-life and industrially relevant NLP applications
Email filtering
invoice information extraction
News generation
Meeting analysis
CVs analysis and classification
Submission:
Papers submitted for review should conform to IEEE specifications. Manuscript templates can be downloaded from IEEE website. The maximum length of papers is 8 pages. All the papers will go through the double-blind peer review process. Authors’ names and affiliations should not appear in the submitted paper. Authors’ prior work should be cited in the third person. Authors should also avoid revealing their identities and/or institutions in the text, figures, links, etc.
Authors should also ensure that their identity is not revealed indirectly by citing their previous work in the third person and omitting acknowledgments until the camera-ready version. Papers have to be submitted via the workshop's EasyChair submission page.
Please include in the paper title "Full paper: Title" or "Short paper: Title" to precise the contribution type. At least one author of each accepted paper must register for the workshop, in order to present the paper. For further instructions, please refer to the FLLM 2025 page.
Important dates:
Submission Deadline: August 31, 2025 September 21st, 2025
Decisions Announced: September 30, 2025 October 10th, 2025
Camera Ready Deadline: October 08, 2025 October 25th, 2025
Workshop: To be announced
Publication:
Accepted papers will be submitted to IEEEXplore for possible publication.
Workshop Chairs
Rim Hantach, Engie, France
Rafika Boutalbi, Aix-Marseille University, France
Karima Boutalbi, Cegedim Business Services, France