Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
The Bias-Variance Tradeoff in Data-Driven Optimization: A Local Misspecification Perspective
Authors: Haixiang Lan, Luofeng Liao, Adam N. Elmachtoub, Christian Kroer, Henry Lam, Haofeng Zhang
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section, we validate our findings by conducting numerical experiments on the newsvendor problem, a classic example in operations research with non-linear cost objectives. We show and compare the performances of the three data-driven methods in the finite-sample regimes under different local misspecification settings, including different directions and degrees of misspecification. The experimental results in the finite-sample regime are consistent with our asymptotic comparisons. |
| Researcher Affiliation | Collaboration | Haixiang Lan1 , Luofeng Liao1 , Adam N. Elmachtoub1, Christian Kroer1, Henry Lam1, Haofeng Zhang1,2 1Department of Industrial Engineering and Operations Research, Columbia University 2Morgan Stanley EMAIL |
| Pseudocode | No | The paper describes methods and processes in narrative text and mathematical formulas but does not include any clearly labeled pseudocode blocks or algorithms. |
| Open Source Code | Yes | Question: Does the paper provide open access to the data and code, with sufficient instructions to faithfully reproduce the main experimental results, as described in supplemental material? Answer: [Yes] Justification: The paper provides open access to the data and code. |
| Open Datasets | No | We describe the local misspecified setting by using the framework of Example 5 and building a model and generating a random demand dataset as follows. We denote the training dataset as n z(j) i on i=1, where n is the training sample size. |
| Dataset Splits | No | The paper discusses a 'training dataset' and uses it for numerical experiments but does not explicitly describe any train/test/validation splits, proportions, or specific splitting methodologies. |
| Hardware Specification | Yes | All computations were carried out on a personal desktop computer without GPU acceleration. |
| Software Dependencies | No | The paper does not specify any software names with version numbers for libraries, frameworks, or solvers used in the experiments. |
| Experiment Setup | Yes | The newsvendor problem has the objective function c(w, z) = a (w z)+ +d (z w)+, where for each j [dz]: (1) z(j) is the customers random demand of product j; (2) w(j) is the decision variable, the ordering quantity for product j; (3) a(j) is the holding cost for product j; (4) d(j) is the backlogging cost for product j. We assume the random demand for each product are independent and the holding cost and backlogging cost is uniform among all products by setting a(j) = 5 and d(j) = 1 for all j [dz]. We set (1) α = 0.1 to denote the severely misspecified setting, (2) α = 0.5 to denote the balanced setting and α = 2 to denote the mildly misspecified setting. We discuss two types of directions: (1) u(z) = Qdz j=1 z(j) 2; (2) u(z) = Qdz j=1 z(j) 3j 2 /2. |