RELAX 2025: Workshop on Relaxed Semantics in Data Analytics Pipelines

The pervasiveness of complex AI and data analytics technologies has created new challenges for distributed data systems. These systems must manage diverse and high-velocity data sources, ranging from multivariate, audiovisual, relational, and unstructured data. Additionally, real-time streaming data requires novel computational approaches to handle its evolving nature efficiently.

The RELAX 2025 workshop focuses on developing scalable and efficient data-intensive software systems that can navigate the trade-offs between strict correctness guarantees and system performance. As modern data analytics systems must process incomplete, uncertain, or inconsistent data while maintaining efficiency and usability, this workshop seeks to explore:

Federated AI models to enhance robustness without centralized data collection.

Computational approximations for efficiency and sustainability in AI processing.

Addressing data inaccuracies, bias, and ethical considerations in analytics systems.

Methods to balance accuracy, consistency, and performance in large-scale data pipelines.

Techniques for ensuring data quality in dynamic, streaming, and distributed environments.

We aim to foster discussions on designing flexible data management systems that balance decentralization, approximation, and ethics while addressing the challenges of high-velocity and diverse data.

For more information, please visit the workshop webpage.