Retrieval-Augmented Generation Workshop

A thematic track at the International Conference on Computational Science

RAG Workshop is a part of the International Conference on Computational Sciencce (ICCS) taking place July 7-9, 2025 in Singapore. The theme of this year's conference is "Making Complex Systems tractable through Computational Science".

Important dates:

Paper submission: 31 January 2025

Notification to authors: 31 March 2025

Camera-ready papers: 18 April 2025

More info here: https://www.iccs-meeting.org/iccs2025/important-dates/

Contacts:

Aleksander Smywiński-Pohl, AGH University of Krakow, apohllo@agh.edu.pl

Magdalena Król, AGH University of Krakow, magdakrol@ahg.edu.pl

Workshop Description:

Advances in Retrieval-Augmented Generation (RAG) for Specialized Domains

Retrieval-Augmented Generation (RAG) has emerged as a critical and rapidly evolving theme in the fields of Artificial Intelligence (AI) and Natural Language Processing (NLP). RAG systems address a significant limitation of Large Language Models (LLMs): their inability to incorporate real-time updates and private knowledge efficiently, as their knowledge is static and embedded within the model’s weights, making updates costly and time-intensive.

By integrating traditional and modern information retrieval (IR) techniques with the generative capabilities of LLMs, RAG systems offer a dynamic solution that allows models to access and utilize the most up-to-date and domain-specific information. This combination makes RAG a promising approach for building systems that are not only powerful but also accurate, efficient, and adaptable to specialized contexts.

This workshop aims to bring together researchers and practitioners working on RAG systems, particularly those focused on complex and high-stakes domains such as law, biology, physics, and medicine, where accuracy and reliability are paramount. Our goal is to foster discussions, share novel approaches, and identify emerging challenges in the development and application of RAG systems.

Topics of Interest

We invite high-quality submissions addressing various aspects of RAG systems, including but not limited to:

1. Core Improvements and Architectures
• •  Innovations in information retrieval models.
• •  New architectures and frameworks for RAG systems.
• •  Enhanced techniques for vector representation and storage.
2. Efficiency and Scalability
• •  Methods to improve processing speed and reduce memory consumption in RAG pipelines.
• •  Quantization techniques for efficient vector storage and retrieval.
• •  Scalable solutions for large-scale domain-specific datasets.
3. Domain-Specific Applications
• •  Development of RAG systems tailored for specialized fields such as legal, biomedical, and scientific domains.
• •  Strategies for adapting LLMs to retrieval tasks in niche contexts.
4. Fusion and Retrieval Optimization
• •  Models for combining results from diverse retrieval systems.
• •  Novel fusion techniques to enhance relevance and accuracy.
5. Evaluation and Robustness
• •  Creation of new evaluation datasets and benchmarks.
• •  Metrics and methodologies for assessing RAG performance.
• •  Techniques for controlling hallucinations in generated outputs.
6. Broader Applications and Challenges
• •  Ethical considerations and biases in RAG systems.
• •  Cross-lingual or multilingual RAG applications.
• •  Use of RAG systems in low-resource or under-represented domains.

Who Should Attend?

This workshop is designed for researchers, industry professionals, and students interested in the intersection of retrieval systems and generative models. Whether your focus is on improving foundational technologies, developing novel applications, or tackling real-world challenges in specialized domains, we invite you to join us in advancing the field of RAG.

Papers

WCS (Workshops on Computational Science) papers (Full Paper, Short Paper) will be published in separate proceedings, that is, there will be a LNCS publication for the Main Conference, and a different LNCS publication for the Workshops. But both ICCS and WCS will be published by LNCS, and indexed in the same way.
The only difference between Full and Short papers is the number of accepted pages. They are otherwise published and indexed in the exact same way.

While we encourage full paper submissions, the “Abstract Only” option caters to researchers who can only publish in specific journals or work for companies in circumstances such that they cannot publish at all, but still want to present their work and discuss it with their peers at ICCS. In the “Abstract Only” option, a short abstract is included in the conference program, but not in LNCS.

After the conference, the best papers will be invited for a special issue of the Journal of Computational Science (Impact Factor: 3.1)

Papers submission

Manuscripts must be written in English and follow the LNCS template guidelines. By submitting a paper, at least one author agrees to register, attend the conference, and present the work.

Accepted papers will be published in Springer Lecture Notes in Computer Science (LNCS) series and indexed by Scopus, EI Engineering Index, Thomson Reuters Conference Proceedings Citation Index (part of ISI Web of Science), and other indexing services. Each paper will include linked references, XML versions, and citable DOI numbers.

The latest versions of the templates can be downloaded here and are also accessible on the Overleaf platform. Submissions must be made via EasyChair platform.

More info & Instruction for Authors:

https://www.iccs-meeting.org/iccs2025/call-for-papers/

Workshop organizers:

Program Committee: