Parameter Inference

Parameter inference aims to determine the values of unknown parameters within a model based on observed data, a crucial task across diverse scientific fields. Current research emphasizes developing robust and efficient algorithms, including Bayesian methods, neural networks (like Physics-Informed Neural Networks and diffusion models), and evolutionary strategies, often tailored to specific model types (e.g., partial differential equations, agent-based models) and data characteristics (e.g., high-dimensionality, non-stationarity). These advancements improve the accuracy and speed of parameter estimation, enabling more reliable model calibration and predictions in applications ranging from cosmology and climate modeling to battery diagnostics and robotics. The development of efficient surrogates for computationally expensive models is a particularly active area, facilitating high-throughput analyses.

Papers

May 31, 2022

Simulator-Based Inference with Waldo: Confidence Regions by Leveraging Prediction Algorithms and Posterior Estimators for Inverse Problems
Luca Masserano, Tommaso Dorigo, Rafael Izbicki, Mikael Kuusela, Ann B. Lee
Scientific Inference Inverse Problem Posterior Inference Parameter Inference Predictive Algorithm Neural Density Confidence Region

March 22, 2022

Constrained Parameter Inference as a Principle for Learning
Nasir Ahmad, Ellen Schrader, Marcel van Gerven
Stochastic Gradient Descent Back Propagation General Principle Parameter Inference Neural State

March 19, 2022

DiSECt: A Differentiable Simulator for Parameter Inference and Control in Robotic Cutting
Eric Heiden, Miles Macklin, Yashraj Narang, Dieter Fox, Animesh Garg, Fabio Ramos
External Control Parameter Inference Differentiable Simulator Model Slicing Robotic Cutting

March 11, 2022

Parameter Inference of Time Series by Delay Embeddings and Learning Differentiable Operators
Alex Tong Lin, Adrian S. Wong, Robert Martin, Stanley J. Osher, Daniel Eckhardt
Time Series Dynamical System Handwritten Trajectory Parameter Inference Delay Embeddings

Parameter Inference

Papers

Simulator-Based Inference with Waldo: Confidence Regions by Leveraging Prediction Algorithms and Posterior Estimators for Inverse Problems

Constrained Parameter Inference as a Principle for Learning

DiSECt: A Differentiable Simulator for Parameter Inference and Control in Robotic Cutting

Parameter Inference of Time Series by Delay Embeddings and Learning Differentiable Operators