A Unified Information-Theoretic Approach to Constrained Multi-Fidelity Multi-Objective Bayesian Optimization

Key points

• We address the challenge of jointly selecting where and at which fidelity to evaluate to identify the highest-fidelity feasible Pareto frontier in this combined setting.1
• From a unified information-theoretic perspective, we measure query utility by the information gain about this frontier, provided by an observation.1
• Since this mutual information is intractable, we derive a variational lower bound using a mixture of under- and over-truncated approximations to the Pareto-consistent region.1
• Multi-fidelity surrogate models propagate the information to arbitrary fidelities, yielding a cost-aware acquisition function without separate heuristics for fidelity selection or constraint handling.1

Sources (1)

• [1]A Unified Information-Theoretic Approach to Constrained Multi-Fidelity Multi-Objective Bayesian Optimization
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 7, 02:47 PM

  Bayesian optimization often involves multiple objectives, constraints, and fidelity levels.

  We address the challenge of jointly selecting where and at which fidelity to evaluate to identify the highest-fidelity feasible Pareto frontier in this combined setting.

Extractive summary: sentences quoted from the sources.