The workshop will comprise three tracks:
Robust Multimodal AI and Data: This track focuses on all aspects of multimodal AI/ML models, and relevant data needs, including
- Techniques for multimodal data fusion
- Foundation models and their applications
- Learning and representation of transient dynamics in complex systems
- Knowledge informed, data driven methods
- Data driven control and examples of applications to real-world systems
- AI supporting data acquisition, communication, data management, and data validation
- Novel techniques for data reduction
- Novel experimental in-process and post-process characterization of systems
- Techniques for optimizing sensor selection and placement for digital twins
- Detecting and dealing with data bias
- Standards for data for digital twin technologies
Deployment on the computing continuum: This track focuses on the ecosystem and methods necessary for practical use of digital twins, such as:
- Real-time data integration, online and offline continual learning on edge-based systems
- Edge deployment for real-time and power- efficient deployment of AI models, digital twins, and integrating HPC and edge systems
- Federated learning, including data privacy
- Security and resilience
- Human-machine interface design
- Interoperability of digital twins
- Standards needs, and the use of digital twin technologies in standards
Assured AI: This track focuses on the technical challenges associated with developing and using robust digital models, such as
- Integrating physics or other knowledge into machine learning
- Methods for multi-scale prediction (especially multi-scale time-series prediction)
- Scaling issues associated with large models
- AI algorithms for control of complex systems
- Uncertainty quantification
- Assurance, including causal inference, explainability, and interpretability
- Verification, validation, and calibration
In addition, topics associated with detecting and dealing with bias, and diversity, equity, and inclusion associated with the development and use of AI solutions for robust engineering and science are of interest.
