DDDAS2024 is the premier conference on Science and Technology capabilities stemming from the Dynamic Data Driven Applications Systems (DDDAS/InfoSymbiotics) paradigm, where instrumentation data are integrated dynamically (at runtime) into models, which in turn can control dynamically the use and acquisition of instrumentation data. DDDAS-based methods provide enhanced “systems-analytics” capabilities and for systems’ dynamic conditions. These capabilities go beyond the typical AI “data-analytics” approaches, based on Machine Learning (ML) and Deep Learning (DL), and other Neural Networks (NN), all of which do not support dynamic systems conditions.
The DDDAS2024 conference will address state-of-the-art efforts that support “systems analytics”, and enable capabilities for “reliable AI” (that is: consistent, trusted, and with quality performance and applicable for dynamic conditions), by combining DDDAS and typical AI approaches, to create frameworks for enabling new and advanced Science and Technology capabilities, which address challenges and create opportunities in important areas, including Sustainable Development, Human Security, Healthcare, and Renewable Energy.
The DDDAS/InfoSymbiotics scientific paradigm has been driving advances in foundational and in applications methods, through novel system-cognizant (as well as subsystems-level) representations/models. These include multimodal, multiscale, multilevel, coupled systems-based models and instrumentation, uncertainty quantification, estimation, observation, sampling, planning, control and intelligent decision making, and where ML and NN are used as tools in the DDDAS-based systems-cognizant models. The DDDAS paradigm, enabling “systems-analytics” and “Dynamic Digital Twin” capabilities, has demonstrated its impact in several important application domains and systems under dynamic conditions; such include aerospace, bio-, cyber-, geo-, space-sciences, computer vision, medical sciences, as well as critical infrastructures’ security, resiliency, and optimized performance. The scope of application areas, and more – ranging “from the nano-scale to the extra-terra-scale”.
Current, typical AI methods are predominantly based on the use of ML, DL, and other NN methods, and mostly use static-data. These approaches require (human initiated) training and retraining when the learning outcomes are not satisfactory. Moreover, for dynamic conditions, repeated, rapid and continuous retraining is required, which may not be computationally feasible. The DDDAS paradigm presents an opportunity to generalize present AI methods by integrating the data acquisition process, together with system-cognizant modeling, which allows incorporating domain knowledge for interpretability, explainability, and predictive power. Integrating the DDDAS approach with AI will enable new ways to enhance capabilities in many research areas, foster development of novel methods, and coordinate data among training, testing, and deployment environments.
In addition to the traditional DDDAS approaches and typical AI methods (ML, DL, and other NN), DDDAS2024 will entertain an expanded conference scope which showcases methods and applications based on coupling DDDAS approaches with AI methods. The DDDAS emphasis on dynamically coupled data and models, as well as their analysis, prediction of behavior and operational control, together with AI methods, entail advances in fundamental areas and permeate many topics of contemporary interest.
Areas of interest include foundational and applications methods:
Coupled DDDAS and AI – theory and methods, and applications DDDAS and AI for analysis, prediction of behavior, and operational control of systems Dynamic control of complex, heterogeneous, multimodal, distributed systems Systems Analytics and beyond Big-Data analytics Dynamic Digital Twins (also referred to as DDDAS-based Digital Twins) Systems-cognizant representations/modeling (e.g., physics-based modeling, agent-, graph-based, etc) coupled with ML/NN methods Informative approaches for Estimation, Control and (Machine) Learning, Planning and Decision support Multimodal learning Optimization methods (beyond /additionally-to deep learning) Multimodality decision making / decision support Information fusion and inference New and advanced Computer Vision methods Planning and control Efficient and Scalable methods for stochastic systems, modeling, simulation, and sensing Learning, optimization and awareness methods Explainable and Interpretable Methods Development of architectures of DDDAS-based reliable AI Coupled dynamic sensors with dynamic decision making and prediction for critical applications Adversarial attack and defense Modeling for threat-awareness effects Novel Scientific Applications enabled through methods above, including areas such as: Materials – Fundaments & Design – Structural Health Monitoring – Advanced Manufacturing Smart Civil Infrastructures – Transportation -Power-grids –Water Distribution – Smart Cities – Smart Agriculture – Energy Systems -Communications Networks (5G/6G, Land, Air, Space) Ecological Systems – Atmospheric Weather – Adverse events (Hurricanes, Tornadoes, Earthquakes) – Environmental Disasters (Wildfires – Oil Spills) – Emergency Response CyberSecurity – Network traffic, Navigation integrity Space Domain Awareness – Space Weather, Space Object Tracking Enterprise Resource Planning – Supply-Chain Logistics – Model-based Real-time Decision Support Health systems – advanced medical diagnostics and intervention –epidemics/pandemics
The conference program includes plenary peer reviewed papers, keynotes, panels, and Best Student Paper recognition awards. The conference proceedings will be published by Springer. The DDDAS 2024 will serve as a forum to present and discuss advances and opportunities in a wide set of application areas and their underlying foundational methods and multidisciplinary collaborative research approaches. Participants from academia, industry, government, and international counterparts will report original work where DDDAS and AI research is advancing scientific frontiers, engendering new science and engineering capabilities, and adaptively optimizing operational processes, on the broad set of topics and interests as delineated above.