Workshop on Intelligent Information Retrieval and Agentic Systems for Specialised Domains (IIRAS'2026)

Scope

The rapid evolution of Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) techniques has opened numerous opportunities for developing intelligent systems capable of autonomous decision-making and adaptive information processing. However, applying these technologies to specialised domains—such as legal, medical, technical, and regulatory contexts—presents unique challenges that require domain-specific approaches, rigorous accuracy standards, and robust compliance frameworks. This workshop focuses on the intersection of advanced AI techniques and domain-specific applications within the broader context of ambient intelligence and artificial intelligence. As intelligent systems become increasingly embedded in our daily environments and routines, the need for accurate, contextually-aware, and trustworthy information retrieval becomes essential. There is a need to addresses the critical gap between general-purpose LLMs and the specialised requirements of professional domains, where errors can have significant legal, financial, or safety implications. By providing a platform for knowledge exchange between academia and industry, this workshop facilitates the transfer of cutting-edge research into practical application. We aim to bring together researchers working on theoretical advances in LLMs and RAG with practitioners deploying these technologies in real-world specialised domains. We invite a broad spectrum of contributions, ranging from original research and architectural proposals to industrial use cases and technological demonstrators. We specifically welcome work addressing the rigorous application of AI and emerging technologies in industrial environments, including research summaries and ongoing PhD studies.

---

---

---

Topics

• RAG and Advanced Information Retrieval
• GraphRAG and knowledge graph integration
• Retrieval optimisation for specialised corpora
• Domain-Specific Language Models
• Fine-tuning and specialisation of foundation models
• Few-shot and zero-shot learning in specialised contexts
• Evaluation benchmarks for domain-specific LLMs
• Agentic AI Architectures
• Memory systems for long-context reasoning
• Tool use and external API integration
• Multi-agent coordination and communication
• Frameworks for agentic systems and benchmarking
• Document processing and analysis automation
• Workflow automation using LLM agents
• Natural language interfaces for complex systems
• Process optimisation through intelligent agents
• Applications in Specialised Domains
• Human in the loop AI
• Trust and transparency in automated systems
• Collaborative intelligence between humans and AI agents
• User preference modelling and personalisation
• Metrics for RAG system performance
• Domain-specific evaluation frameworks
• Robustness testing and adversarial scenarios