Abstract:libEnsemble is a Python-based toolkit for running dynamic ensembles, developed as part of the DOE Exascale Computing Project. The toolkit utilizes a unique generator--simulator--allocator paradigm, where generators produce input for simulators, simulators evaluate those inputs, and allocators decide whether and when a simulator or generator should be called. The generator steers the ensemble based on simulation results. Generators may, for example, apply methods for numerical optimization, machine learning, or statistical calibration. libEnsemble communicates between a manager and workers. We overview the unique characteristics of libEnsemble as well as current and potential interoperability with other packages in the workflow ecosystem. We highlight libEnsemble's dynamic resource features: libEnsemble can detect system resources, such as available nodes, cores, and GPUs, and assign these in a portable way. These features allow users to specify the number of processors and GPUs required for each simulation; and resources will be automatically assigned on a wide range of systems, including Frontier, Aurora, and Perlmutter. Such ensembles can include multiple simulation types, some using GPUs and others using only CPUs, sharing nodes for maximum efficiency. We also describe the benefits of libEnsemble's generator--simulator coupling, which easily exposes to the user the ability to cancel, and portably kill, running simulations based on models that are updated with intermediate simulation output. We demonstrate libEnsemble's capabilities, scalability, and scientific impact via a Gaussian process surrogate training problem for the longitudinal density profile at the exit of a plasma accelerator stage. The study uses gpCAM for the surrogate model and employs either Wake-T or WarpX simulations, highlighting efficient use of resources that can easily extend to exascale.

hubEnsembles: Ensembling Methods in R

Assembling ensembling: An adventure in approaches across disciplines

Developing parsimonious ensembles using predictor diversity within a reinforcement learning framework

Developing parsimonious ensembles using ensemble diversity within a reinforcement learning framework

Diversified Ensembling: An Experiment in Crowdsourced Machine Learning

Dynamic Post-Hoc Neural Ensemblers

Deep interpretable ensembles

Pooling homogeneous ensembles to build heterogeneous ones

Aggregating distribution forecasts from deep ensembles

Multi-model ensembles for ecosystem prediction

Enabling Ensemble Learning for Heterogeneous Large Language Models with Deep Parallel Collaboration.

Ensemble Subset Regression (ENSURE): Efficient High-dimensional Prediction

Context-dependent representation of within- and between-model uncertainty: aggregating probabilistic predictions in infectious disease epidemiology

Ensemble Clustering Based on Meta-Learning and Hyperparameter Optimization

FastEnsemble: A new scalable ensemble clustering method

Hybrid Cluster Ensemble Framework Based on the Random Combination of Data Transformation Operators

Ensemble Learning for Heterogeneous Large Language Models with Deep Parallel Collaboration

Ensemble Methods: Foundations and Algorithms

Portable, heterogeneous ensemble workflows at scale using libEnsemble

H-ensemble: An Information Theoretic Approach to Reliable Few-Shot Multi-Source-Free Transfer

Ensemble Methodology:Innovations in Credit Default Prediction Using LightGBM, XGBoost, and LocalEnsemble