Special Sessions
Link to the Call for Special Session Proposals (closed)
IE’27 will feature two Special Sessions addressing emerging research directions closely related to Intelligent Environments.
Accepted Special Sessions
Intelligent environments increasingly rely on multiple sensing modalities and artificial intelligence to understand people, activities and context. This special session explores how multimodal intelligence can be deployed at or near the edge, enabling real-time, privacy-aware and inclusive interaction while reducing dependence on remote cloud infrastructure.
The session brings together research on multimodal perception and sensor fusion, edge-based learning and inference, resource-efficient AI, human activity and behaviour understanding, assistive technologies, distributed intelligence, privacy and trustworthy AI. It also welcomes work on multimodal foundation models, datasets, benchmarks and real-world deployments of multimodal edge AI systems.
Industry 5.0 is extending intelligent industrial environments beyond automation and efficiency towards systems that are human-centric, sustainable and resilient. This special session explores how AI, IoT, edge and cloud computing, digital twins, cyber-physical systems and intelligent sensing can support the transition from Industry 4.0 to Industry 5.0.
The session brings together research on human-AI and human-robot collaboration, trustworthy and explainable AI, sustainable and energy-efficient operation, resilience to disruptions, smart manufacturing, intelligent buildings and infrastructure, and connected industrial environments. Contributions addressing security, privacy, interoperability and real-world Industry 5.0 applications are also encouraged.
Intelligent medical environments bring together connected systems, diverse data, and AI models, with software agents offering new ways to coordinate workflows. Developing the models that underpin these environments often requires collaboration across institutions while sensitive data remain under local control. Federated learning offers an approach to collaborative model development without centralizing training data.
This 90-minute tutorial introduces federated learning and its relevance to intelligent agent-based medical environments, using NVIDIA FLARE (NVFlare), an open-source framework. Through lectures and presenter-led live demonstrations, participants will explore how existing machine-learning workflows can be adapted to a federated setting. The tutorial will discuss how federated learning can support models used in intelligent medical applications and how AI agents can assist with the development and coordination of federated workflows. It will address practical considerations including data heterogeneity, privacy, security, evaluation, and human oversight, and highlight emerging directions in multimodal learning and foundation models. Researchers, developers, and practitioners will gain a practical introduction and resources for exploring federated AI in their own healthcare applications.
Submission
Papers submitted to Special Sessions follow the same submission and peer-review process as regular IE2027 conference papers. Submissions are made through the main IE2027 submission system, where authors should select the corresponding Special Session track.
The same conference deadlines, paper formats, review criteria, registration requirements, and publication conditions apply. Authors may submit either a Full Paper (up to 8 pages) or a Short Paper (up to 4 pages) in IEEE conference format.