Cost-aware BO minimizes function evaluations with varying costs.
problem Optimization with varying evaluation costs in expensive functions.
method Cost Apportioned Bayesian Optimization (CArBO) combining initial and cost-cooled phases.
result CArBO finds better hyperparameter configurations with the same cost budget.
The notion of expense in Bayesian optimisation generally refers to the uniformly expensive cost of function evaluations over the whole search space. However, in some scenarios, the cost of evaluation for black-box objective functions is non-uniform since different inputs from search space may incur different costs for …
Cer-Eval saves LLM evaluation costs while maintaining accuracy.
problem Challenges in evaluating large language models due to large dataset requirements.
method Adapts to different evaluation objectives, uses test sample complexity, and develops a partition-based algorithm.
result Cer-Eval can save 20-40% test points with comparable accuracy and 95% confidence guarantee.
Improves classifier evaluation by aligning with Total Classification Cost.
problem Lack of consensus on evaluation metrics and class imbalance issues.
method Introduces Weighted Accuracy (WA) and a reweighting framework for cost-sensitive scenarios.
result WA aligns with Total Classification Cost (TCC) minimization under realistic conditions.
New method optimizes costly evaluations in Bayesian optimization.
problem Costly evaluations in BO methods, especially in hyperparameter tuning.
method Formulated as a CMDP, developed rollout approximation for optimal policy.
result Improved cost efficiency over standard BO methods.
New method optimizes costly functions with unknown costs and budget constraints.
problem Optimizing functions with unknown and heterogeneous evaluation costs under a budget constraint.
method Budgeted multi-step expected improvement acquisition function.
result Our method outperforms existing approaches in various synthetic and real problems.
Optimizes function networks by selectively evaluating parts of the network, reducing costs.
problem Optimizing expensive function networks where full evaluations are costly.
method Knowledge gradient acquisition function for cost-aware partial evaluations.
result Outperforms existing BOFN methods and benchmarks across various problems.
The paper proposes a technique to speed up evolutionary algorithms by using lower-cost approximations of the objective function.
problem Evolutionary algorithms require many evaluations to solve computationally expensive black-box optimization problems.
method The paper introduces a technique to choose an appropriate approximate function cost during the execution of the optimization algorithm.
result The proposed approach can reach the same objective value in less than half the time in certain cases.
Paper optimizes broker performance by estimating execution costs.
problem Minimizing execution costs for large trades.
method Intraday modeling of execution cost components (linear and quadratic).
result Substantial improvements in estimating execution costs.
FlexiBO optimizes deep neural networks by balancing cost and performance.
problem Optimizing deep neural networks for multiple objectives incurs high costs.
method Decouples and weights cost in multi-objective Bayesian optimization.
result FlexiBO discovers designs with lower hypervolume error.
Certified algorithms optimize functions with varying costs, providing error bounds.
problem Optimizing functions with varying evaluation costs and error bounds.
method Formalized as a min-max game, proposed certified MFDOO algorithm with cost complexity bound.
result Proposed certified MFDOO algorithm has near-optimal cost complexity for Lipschitz functions.
New algorithm reduces online hyperparameter optimization costs.
problem High cost of evaluating validation examples in online HPO.
method Modeling online HPO as a time-varying Bayesian optimization problem, proposing a costly feedback setting.
result Cost-efficient GP-UCB algorithm reaches human expert-level performance.
Motivated by safety-critical classification problems, we investigate adversarial attacks against cost-sensitive classifiers. We use current state-of-the-art adversarially-resistant neural network classifiers [1] as the underlying models. Cost-sensitive predictions are then achieved via a final processing step in the fe…
This work improves cost-aware Bayesian optimization by introducing Pareto-efficient acquisition functions.
problem Cost variability in hyperparameter evaluations affects the efficiency of Bayesian optimization.
method Reformulated cost-aware Bayesian optimization as Pareto efficiency, proposing a novel Pareto-efficient expected improvement.
result Pareto-efficient acquisition functions significantly outperform previous solutions, providing finer control over cost-accuracy trade-offs.
Comparison of decision curve analysis and cost curves for model evaluation.
problem Evaluating classification performance across different operating contexts.
method Comparison of Decision Curve Analysis (DCA) and Cost Curves.
result DCA and Cost Curves are closely related, with Brier curves being more generally applicable.
We present PESMO, a Bayesian method for identifying the Pareto set of multi-objective optimization problems, when the functions are expensive to evaluate. The central idea of PESMO is to choose evaluation points so as to maximally reduce the entropy of the posterior distribution over the Pareto set. Critically, the PES…
This study integrates cost-sensitive and causal classification methods.
problem Improving classification model performance in business decision-making.
method A unifying evaluation framework for cost-sensitive and causal classification.
result Conventional classification is a specific case of causal classification.
Bayesian optimization with cost-awareness using Gittins index.
problem Optimizing unknown functions with limited data evaluations and costs.
method Developed a connection between cost-aware Bayesian optimization and the Pandora's Box problem, using the Gittins index as an acquisition function.
result The Gittins index-based acquisition function performs well in cost-aware Bayesian optimization, especially in high dimensions.
Paper introduces impact curves for evaluating binarized regression models with varying costs.
problem Evaluating binarized regression models with varying costs and instance-specific utility.
method Proposes impact curves to optimize binary decisions across different utilities.
result Impact curves identify conditions where one model is favored over another and quantify model improvement.
Active testing reduces label costs for efficient model evaluation.
problem Real-world applications require expensive test labels, disconnecting from existing model evaluation methods.
method Derives acquisition strategies to select test points efficiently, addressing label bias and variance.
result Active testing improves model evaluation efficiency without sacrificing accuracy.
Adaptive Bayesian Optimization for resource-constrained experiments with switching costs.
problem Sequential experimental design with varying costs for changing design variables.
method Adapted batch algorithms to sequential problem, proposing cost-aware and cost-ignorant methods.
result Cost-aware algorithm outperforms tuned process-constrained algorithms in all settings considered.
A new method for faster multi-objective optimization by evaluating objectives separately.
problem Finding the Pareto front of trade-offs between multiple objectives efficiently.
method Knowledge Gradient with decoupled evaluations, accounting for different costs.
result The method significantly outperforms existing approaches in terms of evaluation cost.
Machine learning has automated much of financial fraud detection, notifying firms of, or even blocking, questionable transactions instantly. However, data imbalance starves traditionally trained models of the content necessary to detect fraud. This study examines three separate factors of credit card fraud detection vi…
New method assesses energy storage value beyond cost reduction.
problem Improving energy storage value beyond cost reduction.
method Market potential method to evaluate and compare energy storage technologies.
result High-cost hydrogen storage can be more valuable than low-cost hydrogen storage.
Efficiently allocate budgets for LLM-assisted virtual screening to reduce costs.
problem Reducing the cost of evaluating alternatives in large-scale screening tasks.
method Propose a top-m greedy evaluation mechanism and the EFG-m algorithm for efficient budget allocation. result Prove that EFG-m is both sample-optimal and consistent in large-scale virtual screening. Generative AI reduces IR evaluation costs but introduces errors; this work provides reliable CIs.
problem Generating relevance annotations using AI introduces errors that affect IR evaluation metrics.
method Proposes two methods: prediction-powered inference and conformal risk control to place reliable CIs around IR metrics.
result Proposed methods accurately capture both variance and bias in evaluation based on AI-generated annotations.
CJE calibrates cheap LLM judges against an oracle, achieving high accuracy at a fraction of the cost.
problem Inexpensive LLM judges can produce biased rankings, leading to unreliable outcomes.
method CJE uses a small oracle to calibrate cheap scores, then evaluates at scale with valid uncertainty.
result CJE achieves 99% pairwise ranking accuracy at 14x lower cost compared to a 16x oracle/judge cost ratio.
We propose a novel Bayesian Optimization approach for black-box functions with an environmental variable whose value determines the tradeoff between evaluation cost and the fidelity of the evaluations. Further, we use a novel approach to sampling support points, allowing faster construction of the acquisition function.…
Recurrent neural networks are a powerful tool, but they are very sensitive to their hyper-parameter configuration. Moreover, training properly a recurrent neural network is a tough task, therefore selecting an appropriate configuration is critical. Varied strategies have been proposed to tackle this issue. However, mos…
Algorithm reduces audit costs by identifying best service configurations from biased textual evidence.
problem Designing service systems from textual evidence requires accurate selection despite biased automated scoring.
method Developed PP-LUCB algorithm combining LLM scores and selective audits to minimize costs.
result Correctly identified the best model in 40/40 trials with 90% cost reduction.
Framework evaluates privacy cost of non-private pre-processing in DP pipelines.
problem Privacy cost of non-private data-dependent pre-processing in DP machine learning pipelines.
method Establishes upper bounds on overall privacy guarantees using Smooth DP and bounded sensitivity.
result Explicit overall privacy guarantees for various pre-processing algorithms.
A simple formula approximates AUM fees' cumulative costs.
problem Estimating the total cost of AUM fees over time.
method Intuitive explanation and analytical derivation of a formula.
result Investments lose almost Nε% of their value over N years with an annual fee of ε%.
CAPO optimizes LLM prompts more efficiently and cost-effectively.
problem Costly and inefficient automatic prompt optimization for LLMs.
method Integrates AutoML techniques for evolutionary optimization of instructions and few-shot examples.
result Significantly improves prompt optimization efficiency and accuracy.
iMOCA optimizes multiple objectives with continuous approximations for resource efficiency.
problem Optimizing multiple objectives with continuous function approximations that balance accuracy and evaluation cost.
method Information-Theoretic Multi-Objective Bayesian Optimization with Continuous Approximations (iMOCA) selects input and function approximations to maximize information gain per unit cost.
result iMOCA significantly improves over existing single-fidelity methods in approximating the optimal Pareto set.
The paper develops a method to optimize individualized treatment rules for cost-effectiveness.
problem Developing cost-effective individualized treatment rules for healthcare policy.
method Using conditional random forest and net-monetary-benefit (NMB) to estimate optimal CE-ITR.
result The approach optimizes healthcare resource allocation by maximizing health gains and minimizing costs.
Bayesian optimization reduces hyperparameter tuning cost for stochastic models.
problem Hyperparameter tuning under uncertainty in noisy function evaluations.
method Bayesian optimization framework for scale parameter in stochastic models, using statistical surrogate and closed-form optimizer.
result Significant reduction in computational cost (40 times fewer data points, 40-fold reduction in cost).
Label embedding (LE) is an important family of multi-label classification algorithms that digest the label information jointly for better performance. Different real-world applications evaluate performance by different cost functions of interest. Current LE algorithms often aim to optimize one specific cost function, b…
We study the problem of black-box optimization of a noisy function in the presence of low-cost approximations or fidelities, which is motivated by problems like hyper-parameter tuning. In hyper-parameter tuning evaluating the black-box function at a point involves training a learning algorithm on a large data-set at a …
Accelerates Bayesian optimization of function networks with partial evaluations.
problem Optimizing expensive-to-evaluate function networks with varying node costs.
method Proposes an accelerated algorithm that uses global Monte Carlo simulations to select node-specific candidate inputs.
result Achieves up to a 16x speedup over the original p-KGFN algorithm while maintaining competitive query efficiency.
Developed a cost and revenue model for HEMS to estimate breakeven transport volumes under different reimbursement and labor cost assumptions.
problem Estimating breakeven transport volumes for HEMS under varying reimbursement and labor cost assumptions.
method Developed a two-part model: cost framework and actuarial revenue model using healthcare encounter data and payer reimbursement rates.
result Estimated breakeven transport volumes under different reimbursement and labor cost assumptions.
Adaptive replication improves stochastic function optimization.
problem Challenges in accurately estimating functions with high variance.
method Trust-region-based Bayesian optimization with adaptive replication.
result Adaptive replication substantially improves solution accuracy and efficiency.
New method reduces mixture model evaluation cost for large models.
problem Computational infeasibility of evaluating all mixture components.
method Combining EM and Metropolis-Hastings for stochastic sampling.
result Significantly reduced computational cost for large models.
In this study we model the warranty claims process and evaluate the warranty servicing costs under non-renewing and renewing free repair warranties. We assume that the repair time for rectifying the claims is non-zero and the repair cost is a function of the length of the repair time. To accommodate the ageing of the p…
AFA evaluates AI feature acquisition strategies in domains with high costs.
problem Evaluate AI feature acquisition strategies in domains with high costs.
method Apply missing data methods and offline reinforcement learning under NDE and NUC assumptions.
result Propose a novel semi-offline reinforcement learning framework with three new estimators.
MUMBO optimizes multiple tasks efficiently, even with low-cost related functions.
problem Efficiently optimizing multiple related functions with low-cost evaluations.
method Derives a novel multi-task version of entropy search.
result Robust performance with low computational overhead across various optimization challenges.
In binary classification framework, we are interested in making cost sensitive label predictions in the presence of uniform/symmetric label noise. We first observe that 0-1 Bayes classifiers are not (uniform) noise robust in cost sensitive setting. To circumvent this impossibility result, we present two schemes; un…
Ability for accurate hospital case cost modelling and prediction is critical for efficient health care financial management and budgetary planning. A variety of regression machine learning algorithms are known to be effective for health care cost predictions. The purpose of this experiment was to build an Azure Machine…
Statistical test evaluates if personalizing interventions is cost-effective.
problem Balancing the benefits of personalizing interventions with their potential costs.
method Developed a statistical hypothesis test to assess the performance of personalized interventions.
result The test shows that personalized interventions can outperform standard approaches under certain conditions.