Computational Sciences

Explore opportunities in Computational Sciences

Computational Sciences (CS) Foundational research competency focus on advancing the fundamentals of predictive simulation sciences, data intensive sciences, heterogeneous computing, and emerging computing architectures to transform the future of complex Army multi-domain operations.  Gains made through these underpinning multidisciplinary research efforts provide fundamental insight and understanding of complex Army systems and operating environments.  By harnessing the synergism between emerging heterogeneous computing systems and innovative computational algorithms, the Army is provided computational overmatch in tactical environments where speed and accuracy of decision making is disruptive to OPFOR operations.  Resulting technologies will also provide accelerated innovation of Army materiel.  The envisioned capability specialty areas for Computational Sciences include:

  • AI & ML in SWAP-T (size, weight, power, time) resource-constrained environments
  • Domain-specific and heterogeneous tactical High Performance Computing (HPC)
  • Real-time, scalable big data analytics for Army applications
  • Computational predictive design for complex materials and systems

And the envisioned technical specialty areas include:

  • HPC–enabled combined simulation, emulation, and field testing
  • Multi-scale modeling and uncertainty quantification for predictive computational design
  • Advanced and unconventional computing architectures and algorithms research
  • Data reduction, transformation, and correlation

Computational Sciences have a number of AI Cross-Cutting Research (AICCR) elements that address key AI research challenges for MDO.  It is recognized that the military domains are frequently distinct from commercial applications because of: rapidly changing situations; limited access to real data to train AI and limited resources with SWAP-T constraints, and noisy incomplete, uncertain, and erroneous data inputs during operations; and peer adversaries that employ deceptive techniques to defeat algorithms.  Thus, the primary goal of AICCR is to research and develop artificially intelligent agents (heterogeneous & distributed) that rapidly learn, adapt, reason and act in contested, austere and congested environments by addressing the following AI & ML research challenges in Computational Sciences:

  • Learning in complex environments (AI & ML with small samples, dirty data, high clutter)
  • Resource-constrained processing at the point-of-need (AI & ML computing with extremely low SWAP-T)
  • Generalizable & predictable AI
  • Explainability & programmability for AI & ML
  • AI & ML with integrated quantitative models