Model Misspecification

Model misspecification, the discrepancy between a statistical model and the true data-generating process, is a pervasive challenge across numerous machine learning and scientific domains. Current research focuses on developing methods to detect, quantify, and mitigate the impact of misspecification, employing techniques like robust optimization, generative adversarial networks, and Bayesian approaches within various model architectures including neural networks, decision trees, and linear models. Addressing model misspecification is crucial for improving the reliability and trustworthiness of inferences drawn from data, impacting fields ranging from reinforcement learning and causal inference to scientific modeling and revenue management.

Papers

January 26, 2023

Model-based Offline Reinforcement Learning with Local Misspecification
Kefan Dong, Yannis Flet-Berliac, Allen Nie, Emma Brunskill
State Action Pair Model Based Offline Reinforcement Learning Model Misspecification Offline Policy Policy Selection

January 12, 2023

Toward Theoretical Guidance for Two Common Questions in Practical Cross-Validation based Hyperparameter Selection
Parikshit Ram, Alexander G. Gray, Horst C. Samulowitz, Gregory Bramble
Yes No Question Cross Validation Model Misspecification Hyperparameter Selection Variance Tolerance Factor

December 1, 2022

Regularization Trade-offs with Fake Features
Martin Hellkvist, Ayça Özçelikkale, Anders Ahlén
Implicit Regularization Ridge Regression Model Misspecification Explicit Regularization Fake Document

November 7, 2022

Max-Min Off-Policy Actor-Critic Method Focusing on Worst-Case Robustness to Model Misspecification
Takumi Tanabe, Rei Sato, Kazuto Fukuchi, Jun Sakuma, Youhei Akimoto
Reinforcement Learning Actor Critic Deep Deterministic Policy Gradient Model Misspecification Policy Algorithm Policy Actor Critic Uncertainty Set Worst Case Robustness Multi Joint

October 30, 2022

A view on model misspecification in uncertainty quantification
Yuko Kato, David M. J. Tax, Marco Loog
Machine Learning Model Human Prediction Uncertainty Quantification High Uncertainty Anticipation Uncertainty Estimation Best View Model Misspecification

October 12, 2022

Robust Neural Posterior Estimation and Statistical Model Criticism
Daniel Ward, Patrick Cannon, Mark Beaumont, Matteo Fasiolo, Sebastian M Schmon
Model Misspecification Neural Posterior Estimation Model Criticism

October 11, 2022

Misspecified Phase Retrieval with Generative Priors
Zhaoqiang Liu, Xinshao Wang, Jiulong Liu
Phase Retrieval Model Misspecification Generative Prior Dimensional Signal Phase Retrieval Problem

September 5, 2022

March 16, 2022

Undersmoothing Causal Estimators with Generative Trees
Damian Machlanski, Spyros Samothrakis, Paul Clarke
Treatment Effect Covariate Shift Observational Data Model Misspecification Treatment Effect Estimation

March 7, 2022

Estimation under Model Misspecification with Fake Features
Martin Hellkvist, Ayça Özçelikkale, Anders Ahlén
Estimation Task Model Misspecification Covariance Conditioning Model Mismatch Fake Document

February 22, 2022

Off-Policy Confidence Interval Estimation with Confounded Markov Decision Process
Chengchun Shi, Jin Zhu, Ye Shen, Shikai Luo, Hongtu Zhu, Rui Song
Markov Decision Process Uncertainty Quantification Model Misspecification Unobserved Variable Policy Estimation

February 12, 2022

Robust Learning from Observation with Model Misspecification
Luca Viano, Yu-Ting Huang, Parameswaran Kamalaruban, Craig Innes, Subramanian Ramamoorthy, Adrian Weller
Imitation Learning Robust Learning Robust Reinforcement Learning Model Misspecification Observation Centric Adversarial Imitation Generative Adversarial Imitation Learning

January 6, 2022

Robust Linear Predictions: Analyses of Uniform Concentration, Fast Rates and Model Misspecification
Saptarshi Chakraborty, Debolina Paul, Swagatam Das
General Analysis Robust Estimation Model Misspecification Fast Rate Robust Prediction General Robustness Linear Predictor Robust Statistic

December 21, 2021

Joint Learning of Linear Time-Invariant Dynamical Systems
Aditya Modi, Mohamad Kazem Shirani Faradonbeh, Ambuj Tewari, George Michailidis
Joint Learning Finite Time Model Misspecification Transition Matrix Linear Time Invariant System

December 16, 2021

Detecting Model Misspecification in Amortized Bayesian Inference with Neural Networks
Marvin Schmitt, Paul-Christian Bürkner, Ullrich Köthe, Stefan T. Radev
Neural Network Simulation Based Inference Model Misspecification Amortized Bayesian Inference Bayesian Flow

November 23, 2021

Uncertainty estimation under model misspecification in neural network regression
Maria R. Cervera, Rafael Dätwyler, Francesco D'Angelo, Hamza Keurti, Benjamin F. Grewe, Christian Henning
Uncertainty Estimation Aleatoric Uncertainty Deep Regression Model Misspecification Flexible Model

November 19, 2021

Composite Goodness-of-fit Tests with Kernels
Oscar Key, Arthur Gretton, François-Xavier Briol, Tamara Fernandez
Probabilistic Model Density Estimation Small Kernel Model Misspecification Goodness of Fit Kernel Stein Discrepancy Kernel Two Sample Test

November 12, 2021

Dynamic treatment effects: high-dimensional inference under model misspecification
Yuqian Zhang, Weijie Ji, Jelena Bradic
High Dimensional Robust Estimation Model Misspecification Double Robustness Dynamic Treatment Effect