BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260422T143050Z
LOCATION:B301
DTSTART;TZID=America/New_York:20241118T090000
DTEND;TZID=America/New_York:20241118T123000
UID:submissions.supercomputing.org_SC24_sess746@linklings.com
SUMMARY:Second Workshop on Enabling Predictive Science with Optimization a
 nd Uncertainty Quantification in HPC (EPSOUQ-HPC 2024)
DESCRIPTION:Thrust 1 – Methods driven\nUncertainty quantification theory a
 nd algorithms play a pivotal role in reducing the impact of uncertainties 
 during both optimization and decision-making processes. Bayesian approxima
 tion and ensemble learning techniques are among the most used UQ methods. 
 We are looking to extend this notion to include in the workshop work from 
 researchers representing a variety of UQ methods, focusing perhaps on exam
 ining performance with a view to general applicability in a wide variety o
 f applications,  such as those mentioned in the application section, as we
 ll as e.g. computer vision, image processing, medical image analysis, natu
 ral language processing, bioinformatics, etc.\n\nThrust 2-Model/Data drive
 n\nWith the advent of the Exascale era we expect to see more of these exis
 ting uncertainty propagation methodologies embedded within or wrapped arou
 nd deterministic simulation codes. For example, sampling-based methods rep
 eatedly call a deterministic simulation code for different values of the m
 odel inputs. The independent instantiations of this method makes it a good
  fit for parallel programming paradigms and deployment to compute nodes in
  different HPC systems. However, it is unlikely an Exascale computer will 
 provide enough concurrency for a thousand-fold increase in sample evaluati
 ons for uncertainty propagation applied in this manner. A key topic in mul
 ti-scale simulations is the handling of uncertainty in coupled codes where
  two or more separate codes work together. Coupled codes could typically i
 nclude particle-based and continuum (PDE) based models to bridge scales. T
 his thrust will also include methods and applications that can bring toget
 her AI/ML and physics-based methods to enable UQ and optimization.  Exampl
 e topics in this area may include in-situ surrogate construction, intrusiv
 e reduced order modeling, feature/manifold learning in large data sets.\n\
 nThrust 3 – Use-inspired HPC challenges and solutions\nOf particular inter
 est are real-world experiences with HPC systems and articulations of activ
 e challenges in the use of these systems for optimization and UQ. Potentia
 l topic areas motivated by the desire to perform optimization and UQ inclu
 de strategies for overcoming supercomputing bottlenecks such as the impact
  of queue policies and restrictions on scientific workflows, limitations w
 ith I/O and methods for I/O coupling when using multiple distinct simulati
 on/analysis tools, in-memory and file-based coupling strategies, mixed-lan
 guage and mixed-precision computing, and general memory/compute limitation
 s\n\nScrutinizing Variables for Checkpoint Using Automatic Differentiation
 \n\nCheckpoint/Restart (C/R) saves the running state of the programs perio
 dically, which consumes considerable time and system resources. We observe
  that not every piece of data is involved in the computation in typical HP
 C applications; such unused data should be excluded from checkpointing for
  better ...\n\n\nXin Huang, Weiping Zhang, Shiman Meng, Wubiao Xu, and Xia
 ng Fu (Nanchang Hangkong University); Luanzheng Guo (Pacific Northwest Nat
 ional Laboratory (PNNL)); and Kento Sato (RIKEN)\n---------------------\nO
 ptimizing Uncertainty Estimation on Scientific Visualizations Using Learni
 ng Models\n\nScientific visualizations (SciVis) convert numerical and spat
 ial data into images, enabling deeper insights into complex phenomena. Rec
 ent advancements in machine learning, particularly Deep Learning (DL), hav
 e significantly enhanced SciVis. By combining classical numerical techniqu
 es with DL, we ca...\n\n\nErik Pautsch (Loyola University, Chicago; Argonn
 e National Laboratory (ANL)); David Guerrero-Pantoja and Clara Almeida (Un
 iversity of California, Santa Cruz; Argonne National Laboratory (ANL)); Ma
 ria Pantoja (California Polytechnic State University, San Luis Obispo; Arg
 onne National Laboratory (ANL)); Silvio Rizzi (Argonne National Laboratory
  (ANL)); and George Thiruvathukal (Loyola University, Chicago; Argonne Nat
 ional Laboratory (ANL))\n---------------------\nWelcome and Introduction\n
 \nAntigoni Georgiadou (Oak Ridge National Laboratory (ORNL)) and Tiernan C
 asey (Sandia National Laboratories)\n---------------------\nClosing Remark
 s, OLCF/SNL updates, best paper announcement\n\nAntigoni Georgiadou (Oak R
 idge National Laboratory (ORNL)) and Tiernan Casey (Sandia National Labora
 tories)\n---------------------\nPanel discussion\n\nAntigoni Georgiadou (O
 ak Ridge National Laboratory (ORNL)); Tiernan Casey (Sandia National Labor
 atories); Choongseock (CS) Chang (Princeton Plasma Physics Laboratory); Pe
 ter Coveney (University College London (UCL)); and Shantenu Jha (Rutgers U
 niversity, Princeton Plasma Physics Laboratory)\n---------------------\nEP
 SOUQ-HPC 2024 — Morning Break\n---------------------\nEnsemble Simulations
  on Leadership Computing Systems\n\nScientific productivity can be enhance
 d through workflow management tools, relieving large High Performance Comp
 uting (HPC) system users from the tedious tasks of scheduling and designin
 g the complex computational execution of scientific applications. This pap
 er presents a study on the usage of ense...\n\n\nAntigoni Georgiadou, Henr
 y Monge-Camacho, Tanvir Sohail, and Swarnava Ghosh (Oak Ridge National Lab
 oratory (ORNL)); Arjun Valiya Parambathu (University of Delaware); and Dil
 ipkumar Asthagiri, Dmytro Bykov, Tushar Athawale, and Thomas Beck (Oak Rid
 ge National Laboratory (ORNL))\n---------------------\nA Scalable Training
 -Free Diffusion Model for Uncertainty Quantification\n\nGenerative artific
 ial intelligence extends beyond its success in image/text synthesis, provi
 ng itself a powerful uncertainty quantification (UQ) technique through its
  capability to sample from complex high-dimensional probability distributi
 ons. However, existing methods often require a complicated t...\n\n\nAli H
 aisam Muhammad Rafid (Virginia Tech), Junqi Yin (Oak Ridge National Labora
 tory (ORNL)), Yuwei Geng (University of South Carolina), Siming Liang and 
 Feng Bao (Florida State University), Lili Ju (University of South Carolina
 ), and Guannan Zhang (Oak Ridge National Laboratory (ORNL))\n\nTag: Applic
 ations and Application Frameworks, Algorithms, Performance Evaluation and/
 or Optimization Tools\n\nRegistration Category: Workshop Reg Pass\n\nSessi
 on Chairs: Tiernan Casey (Sandia National Laboratories) and Antigoni Georg
 iadou (Oak Ridge National Laboratory (ORNL))
END:VEVENT
END:VCALENDAR