MARV Workshop

IEEE ISC2 2022

MARV workshop

Multimodal and AI-Responsible data processing and deliVery in smart cities

Cities have become actual “data engines” exploiting a huge variety of IoT urban sensors and devices recording multiple everyday activities in the city environment and producing large-scale heterogeneous datasets. This leads to new challenges for extracting valuable knowledge and technological and commercial value from data, leading to more accurate predictions and better analytics, which hold an unprecedented opportunity to shift the traditional methodologies, techniques, and tools of information extraction into new dimensions by cracking the problem of extreme-scale data analytics. To address these challenges, new scientific and technological contributions in the relevant domains are needed, to jointly boost efficiency, reliability, and performance in real-world applications with noisy data. The Workshop will be a forum for presenting the latest scientific contributions in multi-modal data analysis over heterogeneous, edge-cloud continuum architectures and high-performance (de-)centralized computing, along with the required technological solutions and frameworks that can enable their real-life implementations.

MARV aspires the convergence of a set of technologies in the areas of AI, analytics, multimodal perception, software engineering, networks, HPC as part of an Edge-Cloud Computing Continuum paradigm, as well as Edge-Cloud orchestration, to support data-driven real-time application workflows and decision making in modern cities, showcasing the potential to address technological and societal challenges very effectively.

Contributions can be analytical, empirical, technological, methodological, systemic or a combination of those. Papers reporting strong data-oriented systems engineering contributions backed by solid and appropriate evaluations are strongly encouraged. The impact of the contributions should be demonstrated in the context of the data-related aspects in the smart cities realm. Research contributions are solicited in all application areas pertinent to multimodal extreme scale data analytics for smart cities environments, including but not limited to:

Machine learning for Internet of Things
Audio-visual and multimodal data analytics and AI
Multimodal data fusion for streaming and batch analytics
Edge-cloud data management and distribution
Automated and advanced data labelling and annotation
Security in the edge-cloud continuum
Privacy preservation and data protection
Responsible AI and ethical governance
Anonymization techniques preserving contextual information
MLOps for multimodal analytics and AI in heterogeneous environments
Deployment, orchestration and offloading in the edge-cloud continuum
Distributed and federated learning
Just-in-time deep learning models (e.g. early exiting, dynamic computation graphs)
Collaborative edge computing with machine/deep learning
Resource-efficient ML/DL at the edge
Communications and networking for smart cities applications
Advanced visualizations and decision making
Multimodal datasets and open data
Novel and emerging applications in multimodal smart city analytics and AI
Human-in-the-loop for improving decision making in Smart City environments