Searcharxiv⌕ Search

arXiv subjects

Dimitris Bertsimas

Publications and source records attributed to Dimitris Bertsimas.

At least 19 recordsLinked to original sources

Algorithmic Insurance

When artificial intelligence (AI) systems make errors in high-stakes domains such as medical diagnosis or autonomous driving, a single algorithmic flaw across varying contexts can generate heterogeneous losses that challenge traditional insurance assumptions. Algorithmic insurance, financial coverage for AI-induced damages, is an emerging market, yet insurers struggle to price these risks and AI developers lack frameworks linking system design to liability exposure. We connect a key operational choice in binary classification, how practitioners set decision thresholds, to tail risk exposure. Using conditional value-at-risk (CVaR) to capture extreme losses, we prove that established approaches such as accuracy maximization can significantly increase worst-case losses relative to tail risk optimization, with penalties growing at least quadratically as thresholds deviate from optimal. We propose a liability insurance contract that mandates risk-aware thresholds and characterize when it creates value for AI providers, extending the analysis to degrading model performance and human oversight. In a mammography case study, accuracy-maximizing thresholds carry up to four times the tail risk of CVaR-optimal thresholds. Insurance contracts create 7-8% gains for well-calibrated firms through risk transfer and regulatory capital relief, and up to 105% for poorly calibrated firms through mandatory recalibration, while partially aligned firms can lose under a strict mandate, calling for threshold flexibility in contract design. Unlike traditional insurance that merely transfers risk, algorithmic insurance can act as both a financial instrument and an operational governance mechanism, enabling risk transfer while improving AI safety.

cs.LG↗

Decision-Weighted Limit Theorems: Less Conservative Robust Optimization

Limit theorems provide a natural way to construct tractable models for optimization under uncertainty. Existing approaches typically apply them to unweighted sums of uncertain parameters and use the resulting bounds to construct uncertainty sets. We show that sharper formulations are obtained by applying limit theorems directly to the uncertain linear expressions in the optimization problem, with the decision variables as weights. For independent random variables with finite variance, the weighted central limit theorem yields a second-order cone constraint with a decision-dependent protection term, and for heavy-tailed random variables governed by stable limit laws, the weighted approach yields power-cone constraints. Using norm duality, we show that, under a condition on the individual bounds, the feasible region of the unweighted formulation is contained in that of the weighted one. In 16,000 knapsack instances with a prescribed violation probability of 5%, the weighted formulation improves the optimal objective value by 20.9% on average, with overall average violation probabilities between 3.45% and 4.95%, whereas those of the unweighted formulation are near zero.

math.OC↗

Ensemble Modeling for Time Series Forecasting: an Adaptive Robust Optimization Approach

Accurate time-series forecasting is critical for many problems involving temporal data. Ensemble modeling is a well-established technique that leverages multiple predictive models to improve accuracy and robustness, since a single predictor's performance can vary widely due to shifts in the underlying data distribution. This paper proposes a new methodology for building robust ensembles of time series forecasting models. Our approach uses Adaptive Robust Optimization (ARO) to construct a linear regression ensemble whose model weights adapt over time. We demonstrate the effectiveness of our method through synthetic experiments covering eight disturbance scenarios, from no drift to abrupt regime shifts, and real-world applications, including air pollution management, energy consumption forecasting, and tropical cyclone intensity forecasting. In the real-world applications, our adaptive ensembles outperform the best ensemble member in hindsight by 16-26% in root mean square error (RMSE) and 14-28% in conditional value at risk at the 15% level and improve over the other ensemble methods we compare on most metrics. In the synthetic experiments, they have lower mean RMSE than a rolling-ridge comparator in seven of the eight scenarios, significantly in four, while an abrupt regime shift that first appears after model selection remains a limitation.

cs.LG↗

Online LLM Personalization from Pairwise Preferences Across Users

Large language models can answer the same question in different styles, but users' preferred styles are hidden and vary across users and questions. We study inference-time personalization through online selection from a fixed set of prompts, each representing a persona. Users arrive sequentially, each asks a short sequence of questions, and the system displays two responses per question and observes one pairwise choice. We formulate this interaction as a variant of the contextual dueling bandit problem in which the learner starts without preference data, adapts within each session, and exploits information shared across users. The objective is to minimize cumulative strong regret, summing the utility gaps of both displayed responses relative to the user's best response among the available options. We make two contributions. First, we introduce the problem and release a reproducible preference benchmark and dataset built around a tutoring case study, containing educational questions, LLM-generated answers in six teaching styles, and pairwise judgments from LLM-simulated students with hidden preference profiles. We fit Bradley-Terry models to these judgments to obtain reference utilities for regret evaluation. Second, we propose a baseline that learns a mixture prior over users from completed sessions and adapts to each new user within the session. On students and topics never used in development, it removes 44-48% of random pairs' excess regret and achieves over half the performance of a per-user linear oracle fitted in hindsight. We release code, prompts, and the benchmark.

cs.LG↗

Towards Optimal Valve Prescription for Transcatheter Aortic Valve Replacement (TAVR) Surgery: A Machine Learning Approach

Transcatheter Aortic Valve Replacement (TAVR) has emerged as a prominent, minimally invasive treatment for patients with severe aortic stenosis, a life-threatening cardiovascular condition. Multiple transcatheter heart valves (THV) have been approved for use in TAVR, but current guidelines regarding valve type prescription remain a topic of ongoing debate within the medical community. We propose a data-driven clinical support tool to identify the optimal valve type with the objective of minimizing the risk of permanent pacemaker implantation (PPI), a predominant postoperative complication. We synthesize a novel dataset, combining U.S. and Greek patient populations, that integrates data from three distinct sources (patient demographics, computed tomography scans, echocardiograms) while harmonizing the different encoding processes specific to each country's record system. We propose leaf-level analysis to leverage the heterogeneity of the patient populations and avoid benchmarking against uncertain counterfactual risk estimates. The final prescriptive model shows a reduction in PPI rates of 26% and 16% compared to the current standard of care in our internal U.S. population and external, Greek validation set, respectively. To the best of our knowledge, this work represents the first unified, personalized prescription strategy for THV selection in TAVR.

cs.LG↗

Bespoke: Generating MOOC-Quality Industry-Personalized Lecture Videos at Scale

Lifelong learners from different industries often watch the same recorded lecture, even when they will apply the material in different workplaces. We present Bespoke, a system that takes the transcript of an existing lecture and generates a new video customized to a target industry and duration, with new slides, narration, and charts. From 31 graduate lectures in analytics, machine learning, and optimization, we generated 209 videos for these three industries, plus a generic-audience version of each lecture. Twenty-five experts in the corresponding domains rated 92 videos on a five-point rubric. They judged 87% at or above the rubric midpoint corresponding to ``a standard MOOC lecture's quality'' (mean overall score 3.42 out of 5; 48% scored 4 or above). Overall quality was similar across industries, durations, and a 21-lecture held-out set unused while developing the system, at about \$0.22 API cost per minute of video.

cs.HC↗

WFDroneBench: A Benchmark for Sensor Placement and Drone Routing for Wildfire Detection

Increasingly frequent and severe wildfires threaten ecosystems, public health, and infrastructure. Early detection is vital but limited by existing monitoring systems. Drones offer mobile, real-time coverage, but optimizing sensor placement and drone routing in dynamic fire zones remains challenging. To address this, we introduce WFDroneBench, an open-source Python benchmarking library for early wildfire detection that integrates machine-learned risk maps with optimization-based deployment strategies for sensors, charging stations, and drones. It evaluates risk maps, optimization strategies, and monitoring equipment using standardized metrics and realistic wildfire simulations. The framework supports benchmarking across predictive and decision-making components: machine learning researchers can assess risk models and compare routing strategies. WFDroneBench includes 7746 scenarios across 49 locations, built from historical ignitions, real-world wildfire risk maps, and simulated fire spread, along with two ground detector and three drone routing strategies. Our experiments show that the risk-aware strategy Max-Coverage significantly outperforms other baselines when risk maps are sufficiently accurate, achieving the fastest detection on the most difficult fires. We further find that risk-aware static infrastructure helps even under an imperfect risk map and drone-based detection outperforms ground sensors. Finally, our results reveal two key open challenges: (i) detecting small fires rapidly and reliably, and (ii) improving risk-map prediction, where the gap between ground-truth ignition patterns and available risk maps highlights a significant opportunity for ML innovation. We openly release all code, data, and documentation.

math.OC↗

Rapid drone-based wildfire detection at a fraction of current prevention spending

Early wildfire detection is critical to prevent small ignitions from escalating into large-scale disasters, yet current monitoring systems lack a quantitative framework for allocating detection infrastructure at scale. We jointly optimize the placement of monitoring infrastructure and the routing of autonomous drones under realistic operational constraints to quantify the investment required for rapid, large-scale wildfire detection. Evaluated out-of-sample on 3,693 California ignitions from 2021-2024, an optimized drone network operating at a $100 million five-year budget detects 97.3% of fires, including 74% within the first hour. Amortized over five years, that budget is about $20 million per year, roughly 5% of California's annual wildfire-prevention expenditure. Under current technology costs, drone-based monitoring is substantially more cost-effective than static ground sensors. Detection is governed primarily by spatial coverage, while routing strategy mainly determines detection speed.

math.OC↗

Strike Price Optimization for ISO New England's Day-Ahead Ancillary Services

ISO New England's (ISO-NE) Day-Ahead Ancillary Services Initiative settles reserve products as financial call options on real-time energy prices. The system-wide strike price creates an efficiency-reliability tradeoff: increasing it lowers competitive reserve offers, but weakens resources' incentives to incur preparation costs and remain available for real-time performance. The existing strike price rule does not explicitly account for heterogeneous resource incentives and reserve requirements. We develop an optimization framework that selects the highest strike price while ensuring that enough resources retain an incentive to prepare and collectively satisfy the reserve requirements. Because a resource's preparation decision may affect the resulting real-time price distribution, its incentive depends on an unobservable counterfactual. To address this, we derive a tight lower-bound certificate using only the available conditional price distribution and a bound on the resource's price impact. We show that each resource enters the optimization through a single incentive threshold and that any finite optimal strike price occurs at one of these thresholds. This yields a tractable exact solution method based on threshold calculations and a small number of linear feasibility checks. Using reconstructed ISO-NE conditional price distributions and representative gas-fired resources, we find that higher heat-rate combustion turbines are more likely than combined-cycle resources to constrain the strike price choice.

math.OC↗

A Causal DAG Prior for Synthetic Time-Series Classification Datasets

A Prior-data fitted Network learns the posterior predictive induced by its training prior; bringing this paradigm to multivariate time-series classification therefore calls for a synthetic generator that produces complete labelled datasets with temporal structure. We introduce a causal prior that synthesizes each dataset from a randomly sampled DAG over typed nodes across two modalities (tabular attributes and time series), natively producing multivariate, multi-class TSC datasets with cross-modal causal structure across channels, timesteps and labels, a regime not addressed by existing synthetic priors. To validate the prior, we finetune TabPFN v2.5 with minimal adaptations and evaluate on 75 UCR/UEA datasets within TabPFN's operating regime. Finetuning on our generator significantly outperforms both the unmodified upstream model and a tabular-only ablation of the same prior (Wilcoxon signed-rank $p=3.0\times 10^{-8}$ on ROC-AUC), isolating the contribution of the cross-modal temporal structure.

cs.LG↗

RocketPFN: Accurate Time Series Classification via In-Context Learning

We introduce RocketPFN, a training-free pipeline for time series classification that combines random convolutional feature extraction (Rocket) with in-context classification via a pretrained tabular foundation model (TabPFN v2.5). On 92 UCR datasets (30-resample protocol), RocketPFN matches HC2, the strongest published method on the archive, in mean accuracy (both 0.900, Wilcoxon p=0.50), with no training on the target data and a median inference time of 30 seconds per fold. It also significantly outperforms every individual classifier in the HC2 ensemble. On UEA (20 datasets) the difference is likewise not statistically significant. A separate comparison concerns TSC foundation models: when paired with the same downstream classifier, MOMENT, Mantis, and MantisV2 are all significantly outperformed by RocketPFN using fewer extracted features and no learned parameters (p<0.001 in each case). This holds even when the encoders were pretrained on corpora that include the UCR training samples. We propose this two-stage pipeline as a reference point for evaluating zero-shot TSC foundation models.

cs.LG↗

Robust Optimization for Green Ammonia Production

The central challenge in optimizing green ammonia systems is satisfying the minimum-load requirements of the Haber-Bosch (HB) process under renewable uncertainty. We develop a robust optimization framework consisting of a strategic capacity planning model and an operational flow model under solar and wind uncertainty. The strategic model is a mixed-integer optimization (MIO) problem with flexible HB operating modes, namely hot-idling and shutdowns. To address the resulting computational challenges, we propose a robust scenario-reduction framework that combines k-means clustering with robust optimization to generate adversarial renewable trajectories. For the operational model, we develop adaptive robust rolling-horizon formulations under forecast uncertainty. Computational results show that the proposed framework produces feasible capacity plans under out-of-sample simulation, whereas existing approaches based on constraint aggregation fail to satisfy HB minimum-load requirements. Adaptive policies achieve higher ammonia production than static robust policies for a given robustness level, but provide weaker protection against realizations outside the uncertainty set.

math.OC↗

Prescribe-then-Select: Adaptive Policy Selection for Contextual Stochastic Optimization

We address the problem of policy selection in contextual stochastic optimization (CSO), where covariates are available as contextual information and decisions must satisfy hard feasibility constraints. In many CSO settings, multiple candidate policies--arising from different modeling paradigms--exhibit heterogeneous performance across the covariate space, with no single policy uniformly dominating. We propose Prescribe-then-Select (PS), a modular framework that first constructs a library of feasible candidate policies and then learns a meta-policy to select the best policy for the observed covariates. We implement the meta-policy using ensembles of Optimal Policy Trees trained via cross-validation on the training set, making policy choice entirely data-driven. Across two benchmark CSO problems--single-stage newsvendor and two-stage shipment planning--PS consistently outperforms the best single policy in heterogeneous regimes of the covariate space and converges to the dominant policy when such heterogeneity is absent. All the code to reproduce the results can be found at https://anonymous.4open.science/r/Prescribe-then-Select-TMLR.

cs.LG↗

Optimal Control of Multiclass Fluid Queueing Networks: A Machine Learning Approach

We propose a machine learning approach to the optimal control of multiclass fluid queueing networks (MFQNETs) that provides explicit and insightful control policies. We prove that a piecewise constant optimal policy exists for MFQNET control problems, with segments separated by hyperplanes passing through the origin. We use Optimal Classification Trees with hyperplane splits (OCT-H) to learn an optimal control policy for MFQNETs. We use numerical solutions of MFQNET control problems as a training set and apply OCT-H to learn explicit control policies. Furthermore, we show that both the theoretical results and the proposed algorithm extend to robust MFQNETs with uncertain service and arrival rates. We report experimental results with up to 33 servers and 99 classes that demonstrate that the learned policies achieve 100% accuracy on the test set. While the offline training of OCT-H can take days in large networks, the online application takes milliseconds.

cs.LG↗

Optimal Control of Fluid Restless Multi-armed Bandits: A Machine Learning Approach

We present a novel machine learning framework for the optimal control of fluid restless multi-armed bandit problems (FRMABPs) with state equations that are either affine or quadratic in the state variables. By establishing fundamental properties of FRMABPs, we develop an efficient numerical algorithm that generates a comprehensive training set by solving multiple instances with diverse initial states. We further enhance this training set by applying a nonlinear transformation to the feature vectors, leveraging structural properties of FRMABPs. A time-dependent state feedback policy is then learned using Optimal Classification Trees with hyperplane splits (OCT-H). We test our approach on machine maintenance, epidemic control, and fisheries control problems, demonstrating that our method yields high-quality state feedback policies. Furthermore, once a policy is learned, it achieves a speed-up of up to 26 million times compared to the direct numerical algorithm.

cs.LG↗

A Machine Learning Approach to Two-Stage Adaptive Robust Optimization

We propose an approach based on machine learning to solve two-stage linear adaptive robust optimization (ARO) problems with binary here-and-now variables and polyhedral uncertainty sets. We encode the optimal here-and-now decisions, the worst-case scenarios associated with the optimal here-and-now decisions, and the optimal wait-and-see decisions into what we denote as the strategy. We solve multiple similar ARO instances in advance using the column and constraint generation algorithm and extract the optimal strategies to generate a training set. We train machine learning models that predict high-quality strategies for the here-and-now decisions, the worst-case scenarios associated with the optimal here-and-now decisions, and the wait-and-see decisions. The models can be applied to problems with varying dimensions. We also introduce novel methods to expedite training data generation and reduce the number of different target classes the machine learning algorithm needs to be trained on. We apply the proposed approach to the facility location, the multi-item inventory control and the unit commitment problems. Our approach solves ARO problems drastically faster than the state-of-the-art algorithms with high accuracy.

cs.LG↗

A Multimodal and Explainable Machine Learning Approach to Diagnosing Multi-Class Ejection Fraction from Electrocardiograms

Left ventricular ejection fraction (LVEF) assessment depends on echocardiography, limiting access in primary care and resource-constrained settings. We developed a multimodal machine-learning framework that combines engineered 12-lead ECG timeseries features with structured EHR variables to classify LVEF into four clinically used strata: normal (>50%), mildly reduced (40-50%), moderately reduced (30-40%), and severely reduced (<30%). To support model explainability, we identified the most influential ECG and EHR features via SHAP attributions. Using retrospective data from Hartford HealthCare, we trained XGBoost models on 36,784 ECG-echocardiogram pairs from 30,952 outpatients and evaluated temporal generalizability on 19,966 ECGs from a subsequent period. The multimodal model achieved one-vs-rest AUROCs of 0.95 (severe), 0.92 (moderate), 0.82 (mild), and 0.91 (normal), outperforming ECG-only and EHR-only baselines, and maintained performance under temporal validation. This work supports ECG-based, multimodal LVEF stratification as a practical screening and triage aid to prioritize confirmatory imaging where resources are limited.

cs.LG↗

Robustifying and Selecting Cohort-Appropriate Prognostic Models under Distributional Shifts

External validation is widely regarded as the gold standard for prognostic model evaluation. In this study, we challenge the assumption that successful external calibration guarantees model generalizability and propose two complementary strategies to improve transportability of prognostic models across cohorts. Using six real-world surgical cohorts from tertiary academic centers, we tested whether successful external calibration depends largely on similarity in covariates and outcomes between training and validation cohorts, quantified using Kullback-Leibler (KL) divergence, with calibration assessed by the Integrated Calibration Index (ICI). From the model-developer's perspective, we trained the "best-on-average" prognostic model by tuning toward a meta-analysis-derived covariate and outcome distribution as an approximation of the broader target population. From the end-user perspective, we proposed a simple measure for cohort outcome similarity to identify, among published models, the one most suitable for a given target cohort in terms of both calibration and clinical utility. External calibration worsened as distributional mismatch increased. Higher KL divergence was associated with higher ICI in both surgery-alone (Spearman $ρ=0.614$, $p=0.004$) and surgery + adjuvant chemotherapy cohorts (Spearman $ρ=0.738$, $p<0.001$). Meta-analysis-informed weighting improved calibration in most settings without materially affecting discrimination, with the clearest benefit when evaluated on the aggregated external population ($p=0.037$). Models developed in more similar cohorts achieved lower ICI in surgery-alone (Spearman $ρ=0.803$, $p<0.001$) and surgery + adjuvant chemotherapy cohorts (Spearman $ρ=0.737$, $p<0.001$), and provided greater clinical utility on DCA.

stat.ME↗