Barbara Engelhardt

NeurIPS Conference 2025 Conference Paper

Preference-Guided Diffusion for Multi-Objective Offline Optimization

• Yashas Annadani
• Syrine Belakaria
• Stefano Ermon
• Stefan Bauer
• Barbara Engelhardt

Offline multi-objective optimization aims to identify Pareto-optimal solutions given a dataset of designs and their objective values. In this work, we propose a preference-guided diffusion model that generates Pareto-optimal designs by leveraging a classifier-based guidance mechanism. Our guidance classifier is a preference model trained to predict the probability that one design dominates another, directing the diffusion model toward optimal regions of the design space. Crucially, this preference model generalizes beyond the training distribution, enabling the discovery of Pareto-optimal solutions outside the observed dataset. We introduce a novel diversity-aware preference guidance, augmenting Pareto dominance preference with diversity criteria. This ensures that generated solutions are optimal and well-distributed across the objective space, a capability absent in prior generative methods for offline multi-objective optimization. We evaluate our approach on various continuous offline multi-objective optimization tasks and find that it consistently outperforms other inverse/generative approaches while remaining competitive with forward/ surrogate-based optimization methods. Our results highlight the effectiveness of classifier-guided diffusion models in generating diverse and high-quality solutions that approximate the Pareto front well.

NeurIPS Conference 2024 Conference Paper

Active Learning for Derivative-Based Global Sensitivity Analysis with Gaussian Processes

• Syrine Belakaria
• Benjamin Letham
• Janardhan R. Doppa
• Barbara Engelhardt
• Stefano Ermon
• Eytan Bakshy

We consider the problem of active learning for global sensitivity analysis of expensive black-box functions. Our aim is to efficiently learn the importance of different input variables, e. g. , in vehicle safety experimentation, we study the impact of the thickness of various components on safety objectives. Since function evaluations are expensive, we use active learning to prioritize experimental resources where they yield the most value. We propose novel active learning acquisition functions that directly target key quantities of derivative-based global sensitivity measures (DGSMs) under Gaussian process surrogate models. We showcase the first application of active learning directly to DGSMs, and develop tractable uncertainty reduction and information gain acquisition functions for these measures. Through comprehensive evaluation on synthetic and real-world problems, our study demonstrates how these active learning acquisition strategies substantially enhance the sample efficiency of DGSM estimation, particularly with limited evaluation budgets. Our work paves the way for more efficient and accurate sensitivity analysis in various scientific and engineering applications.

UAI Conference 2021 Conference Paper

Active multi-fidelity Bayesian online changepoint detection

• Gregory W. Gundersen
• Diana Cai
• Chuteng Zhou
• Barbara Engelhardt
• Ryan P. Adams

Online algorithms for detecting changepoints, or abrupt shifts in the behavior of a time series, are often deployed with limited resources, e. g. , to edge computing settings such as mobile phones or industrial sensors. In these scenarios it may be beneficial to trade the cost of collecting an environmental measurement against the quality or “fidelity” of this measurement and how the measurement affects changepoint estimation. For instance, one might decide between inertial measurements or GPS to determine changepoints for motion. A Bayesian approach to changepoint detection is particularly appealing because we can represent our posterior uncertainty about changepoints and make active, cost-sensitive decisions about data fidelity to reduce this posterior uncertainty. Moreover, the total cost could be dramatically lowered through active fidelity switching, while remaining robust to changes in data distribution. We propose a multi-fidelity approach that makes cost-sensitive decisions about which data fidelity to collect based on maximizing information gain with respect to changepoints. We evaluate this framework on synthetic, video, and audio data and show that this information-based approach results in accurate predictions while reducing total cost.

UAI Conference 2019 Conference Paper

End-to-end Training of Deep Probabilistic CCA on Paired Biomedical Observations

• Gregory W. Gundersen
• Bianca Dumitrascu
• Jordan T. Ash
• Barbara Engelhardt

Medical pathology images are visually evaluated by experts for disease diagnosis, but the connection between image features and the state of the cells in an image is typically unknown. To understand this relationship, we develop a multimodal modeling and inference framework that estimates shared latent structure of joint gene expression levels and medical image features. Our method is built around probabilistic canonical correlation analysis (PCCA), which is fit to image embeddings that are learned using convolutional neural networks and linear embeddings of paired gene expression data. Using a differentiable take on the EM algorithm, we train the model end-to-end so that the PCCA and neural network parameters are estimated simultaneously. We demonstrate the utility of this method in constructing image features that are predictive of gene expression levels on simulated data and the Genotype-Tissue Expression data. We demonstrate that the latent variables are interpretable by disentangling the latent subspace through shared and modality-specific views.

NeurIPS Conference 2018 Conference Paper

PG-TS: Improved Thompson Sampling for Logistic Contextual Bandits

• Bianca Dumitrascu
• Karen Feng
• Barbara Engelhardt

We address the problem of regret minimization in logistic contextual bandits, where a learner decides among sequential actions or arms given their respective contexts to maximize binary rewards. Using a fast inference procedure with Polya-Gamma distributed augmentation variables, we propose an improved version of Thompson Sampling, a Bayesian formulation of contextual bandits with near-optimal performance. Our approach, Polya-Gamma augmented Thompson Sampling (PG-TS), achieves state-of-the-art performance on simulated and real data. PG-TS explores the action space efficiently and exploits high-reward arms, quickly converging to solutions of low regret. Its explicit estimation of the posterior distribution of the context feature covariance leads to substantial empirical gains over approximate approaches. PG-TS is the first approach to demonstrate the benefits of Polya-Gamma augmentation in bandits and to propose an efficient Gibbs sampler for approximating the analytically unsolvable integral of logistic contextual bandits.

UAI Conference 2017 Conference Paper

A Reinforcement Learning Approach to Weaning of Mechanical Ventilation in Intensive Care Units

• Niranjani Prasad
• Li-Fang Cheng
• Corey Chivers
• Michael Draugelis
• Barbara Engelhardt

The management of invasive mechanical ventilation, and the regulation of sedation and analgesia during ventilation, constitutes a major part of the care of patients admitted to intensive care units. Both prolonged dependence on mechanical ventilation and premature extubation are associated with increased risk of complications and higher hospital costs, but clinical opinion on the best protocol for weaning patients off of a ventilator varies. This work aims to develop a decision support tool that uses available patient information to predict time-to-extubation readiness and to recommend a personalized regime of sedation dosage and ventilator support. To this end, we use off-policy reinforcement learning algorithms to determine the best action at a given patient state from sub-optimal historical ICU data. We compare treatment policies from fitted Qiteration with extremely randomized trees and with feedforward neural networks, and demonstrate that the policies learnt show promise in recommending weaning protocols with improved outcomes, in terms of minimizing rates of reintubation and regulating physiological stability.

ICML Conference 2016 Conference Paper

Hierarchical Compound Poisson Factorization

• Mehmet Emin Basbug
• Barbara Engelhardt

Non-negative matrix factorization models based on a hierarchical Gamma-Poisson structure capture user and item behavior effectively in extremely sparse data sets, making them the ideal choice for collaborative filtering applications. Hierarchical Poisson factorization (HPF) in particular has proved successful for scalable recommendation systems with extreme sparsity. HPF, however, suffers from a tight coupling of sparsity model (absence of a rating) and response model (the value of the rating), which limits the expressiveness of the latter. Here, we introduce hierarchical compound Poisson factorization (HCPF) that has the favorable Gamma-Poisson structure and scalability of HPF to high-dimensional extremely sparse matrices. More importantly, HCPF decouples the sparsity model from the response model, allowing us to choose the most suitable distribution for the response. HCPF can capture binary, non-negative discrete, non-negative continuous, and zero-inflated continuous responses. We compare HCPF with HPF on nine discrete and three continuous data sets and conclude that HCPF captures the relationship between sparsity and response better than HPF.

ICML Conference 2006 Conference Paper

A graphical model for predicting protein molecular function

• Barbara Engelhardt
• Michael I. Jordan
• Steven E. Brenner

IJCAI Conference 2003 Conference Paper

Factored Planning

• EyalAmir
• Barbara Engelhardt

We present a general-purpose method for dynamically factoring a planning domain, whose structure is then exploited by our generic planning method to find sound and complete plans. The planning algorithm's time complexity scales linearly with the size of the domain, and at worst exponentially with the size of the largest subdomain and interaction between subdomains. The factorization procedure divides a planning domain into subdomains that are organized in a tree structure such that interaction between neighboring subdomains in the tree is minimized. The combined planning algorithm is sound and complete, and we demonstrate it on a representative planning domain. The algorithm appears to scale to very large problems regardless of the black box planner used.

IROS Conference 2001 Conference Paper

Balancing deliberation and reaction, planning and execution for space robotic applications

• Russell Knight
• Forest Fisher
• Tara A. Estlin
• Barbara Engelhardt
• Steve A. Chien

Intelligent behavior for robotic agents requires a careful balance of fast reactions and deliberate consideration of long-term ramifications. The need for this balance is particularly acute in space applications, where hostile environments demand fast reactions, and remote locations dictate careful management of consumables that cannot be replenished. However, fast reactions typically require procedural representations with limited scope and handling long-term considerations in a general fashion is often computationally expensive. We describe three major areas for autonomous systems for space exploration: free-flying spacecraft, planetary rovers, and ground communications stations. In each of these broad applications areas, we identify operational considerations requiring rapid response and considerations of long-term ramifications. We describe these issues in the context of ongoing efforts to deploy autonomous systems using planning and task execution systems.

ICAPS Conference 2000 Conference Paper

Using Generic Preferences to Incrementally Improve Plan Quality

• Gregg R. Rabideau
• Barbara Engelhardt
• Steve A. Chien

We describe a methodology for representing and optimizing user preferences on plans. Our approach differs from previous work on plan optimization in that we employ a generalization of commonly occurring plan quality metrics, providing an expressive preference language. We introduce a domain independent algorithm for incrementally improving the quality of feasible plans with respect to preferences described in this language. Finally, we experimentally show that plan quality can be significantly increased with little additional modeling effort for each domain.