Laplace Explainer

Laplace methods are emerging as powerful tools for explaining and improving the performance of complex machine learning models, particularly in addressing their "black box" nature. Current research focuses on developing model-agnostic explainers, such as LaPLACE, which leverage probabilistic causal reasoning and Markov blankets to identify key features influencing predictions, surpassing the accuracy and consistency of existing methods like LIME and SHAP. This work extends to large-scale applications through scalable Bayesian inference techniques, enabling the analysis of substantial models like deep neural networks. Ultimately, these advancements enhance model interpretability, trustworthiness, and facilitate improved decision-making across diverse fields.

Papers

October 1, 2023

LaPLACE: Probabilistic Local Model-Agnostic Causal Explanations
Sein Minn
Model Agnostic Local Interpretable Model Agnostic Explanation Probable Explanation Laplace Explainer

October 10, 2022

Sampling-based inference for large linear models, with application to linearised Laplace
Javier Antorán, Shreyas Padhy, Riccardo Barbano, Eric Nalisnick, David Janz, José Miguel Hernández-Lobato
Application Proficiency Scientific Inference Large Model Tomographic Reconstruction Bayesian Regression Scalable Bayesian Laplace Explainer

August 7, 2022

An Empirical Analysis of the Laplace and Neural Tangent Kernels
Ronaldas Paulius Lencevicius
Neural Tangent Kernel Reproducing Kernel Hilbert Space Empirical Analysis Small Kernel Kernel Function Laplacian Kernel Laplace Explainer

June 10, 2022

Neural Laplace: Learning diverse classes of differential equations in the Laplace domain
Samuel Holt, Zhaozhi Qian, Mihaela van der Schaar
Neural ODE Neural Ordinary Differential Equation Differential Equation Class Specific Neural Laplace Laplace Explainer

Laplace Explainer

Papers

LaPLACE: Probabilistic Local Model-Agnostic Causal Explanations

Sampling-based inference for large linear models, with application to linearised Laplace

An Empirical Analysis of the Laplace and Neural Tangent Kernels

Neural Laplace: Learning diverse classes of differential equations in the Laplace domain