Invited editors-in-chief : Benjamin Piwowarski and Fabrizio Silvestri
We are pleased to announce a Special Issue on Information Retrieval (IR) and Natural Language Processing (NLP), with a particular emphasis on Retrieval-Augmented Generation (RAG) and related advancements. This special issue aims to bring together cutting-edge research, innovative methodologies, and practical applications at the intersection of IR and NLP, exploring both foundational and applied aspects of these transformative technologies.
RAG has emerged as a powerful framework that combines the strengths of retrieval-based and generative models, enabling systems to generate more accurate, contextually relevant, and factually grounded responses. Alongside RAG, this special issue also welcomes research on broader IR and NLP themes, such as conversational search, question answering, and knowledge-intensive tasks, to provide a comprehensive view of the field.
However, despite their growing prominence, RAG systems are not without their limitations and continue to be in need of exploration and improvement: they often struggle with scalability bottlenecks when indexing and querying ever-larger corpora; risk amplifying biases present in their retrieval sources; and can still produce misleading or outright incorrect responses when the retrieved evidence is insufficient or poorly aligned. Moreover, ensuring explainability and transparency throughout the two-stage retrieval-generation pipeline remains a challenge, as does the seamless integration of multimodal inputs or structured knowledge (for example, from knowledge graphs) without introducing inconsistencies. Accurately evaluating both the end-to-end RAG system and its individual components—retriever, reranker, generator—requires more robust benchmarks and metrics that faithfully capture factuality, relevance, and efficiency.
Topics of Interest
We invite submissions of original research, review articles, and case studies that address, but are not limited to, the following themes:
1. Retrieval-Augmented Generation (RAG), including Knowledge-Intensive Tasks:
- Theoretical frameworks, models, and architectures for RAG
- Evaluation metrics, benchmarks, and datasets for RAG systems
- Applications of RAG in NLP tasks (e.g. question answering, summarization, and fact-checking, and information verification)
- Handling noisy or incomplete retrieval data in RAG systems
- Scalability, efficiency, and real-time performance of RAG models
- Domain-specific applications (e.g., healthcare, legal, education)
- Multimodal approaches combining text, images, audio, and video
2. Conversational Search and Dialogue Systems:
- Dialogue management, context modeling, and query reformulation
- Personalization and user modeling for conversational search
- Multi-turn interactions and task completion in conversational systems
- Voice assistants, chatbots, and conversational agents for search
- Handling ambiguity, incomplete queries, and user intent in conversational search
- Domain-specific applications (e.g., healthcare, legal, education)
- Multimodal approaches combining text, images, audio, and video
3. Human-AI Collaboration and Explainability:
- User studies and human-in-the-loop approaches
- Explainability, interpretability, and transparency of IR and NLP models
- Trustworthy AI and ethical considerations in IR and NLP
- Ethical considerations, bias, and fairness
4. Emerging Trends and Challenges:
- Synergies between RAG, conversational search, and large language models (LLMs)
- Cross-lingual and multilingual IR and NLP systems
- Zero-shot, few-shot, and self-supervised learning in IR and NLP
- Scalability and efficiency in large-scale IR and NLP systems
- Handling bias, fairness, and inclusivity in IR and NLP models
- Novel applications and case studies in real-world settings
Loading... 
