Data analytics is transforming the world of science, health, commerce, defense, and social activities. Complex scientific and human systems are being designed, managed, and optimized using first-principles simulations, data science, machine learning, and graph methods. Many real-world analytic workloads are a mix of algorithms and data types best supported by different programming and parallel execution models.
The composability of models and the capability of computer systems to efficiently and transparently support the diverse model is key to achieving performance and productivity requirements of emerging real-world uses.
This workshop seeks paper on mixed data analytic workflows, algorithms, composability, optimizations, programming environments, hardware designs, and benchmarking studies. Besides regular papers, extended abstracts papers describing innovative ideas related to the workshop theme are also encouraged. Topics of interest, of both theoretical and practical significance, include but are not limited to:
- Applications and workflows integrating scientific simulation, data analytics, and learning.
- Libraries, Runtime systems and programming models in support for Big Data Analytics workflows.
- Data Structures and Algorithms supporting hybrid data models (e.g., Graphs and Tables and Attributed Graphs) and Machine Learning.
- Machine Learning and Combinatorial Optimization algorithms.
- Approaches for managing massive unstructured datasets (including graph databases and solutions combining learning approaches with graph analytics).
- Explainable AI and Fairness in algorithms.
- Novel computer architecture design in support of Big Data Analytics workflows: including micro and system level design, accelerators, custom processors and reconfigurable computing.