Function Approximator

Function approximators, primarily neural networks (including multilayer perceptrons and convolutional neural networks), aim to accurately represent complex functions from data. Current research focuses on improving approximation accuracy, particularly addressing limitations like spectral bias and achieving machine precision, while also exploring the theoretical underpinnings of generalization and robustness in the face of model misspecification. This work is crucial for advancing fields like deep learning, reinforcement learning, and scientific computing, where accurate function approximation is essential for solving complex problems and improving model performance.

Papers

September 18, 2024

Bridging the Gap Between Approximation and Learning via Optimal Approximation by ReLU MLPs of Maximal Regularity
Ruiyang Hong, Anastasis Kratsios
LeArning Abstract Average Approximation Universal Approximators Lipschitz Network Geometric Regularity Function Approximator

July 18, 2024

LLMs as Function Approximators: Terminology, Taxonomy, and Questions for Evaluation
David Schlangen
Natural Language Processing Medical LLM Global Evaluation Pre Trained Comprehensive Taxonomy Artificial General Intelligence Generalist Model Dark Jargon Natural Language Specification Function Approximator

November 8, 2023

Optimal Deep Neural Network Approximation for Korobov Functions with respect to Sobolev Norms
Yahong Yang, Yulong Lu
Deep Neural Network Average Approximation Simple Function Approximation Rate Approximation Property Deep Neural Network Approximation Sobolev Norm Function Approximator

July 18, 2023

Multi-stage Neural Networks: Function Approximator of Machine Precision
Yongji Wang, Ching-Yao Lai
Physic Informed Neural Network Multi Stage Multidimensional Local Precision Rate Single Neural Network Function Approximator

June 19, 2023

On the Model-Misspecification in Reinforcement Learning
Yunfan Li, Lin Yang
Reinforcement Learning Sample Efficient Model Misspecification Approximate Policy Iteration Function Approximator

January 4, 2022

Efficient Quantum Feature Extraction for CNN-based Learning
Tong Dou, Guofeng Zhang, Wei Cui
Quantum Circuit Convolution Layer Classical Convolutional Neural Network Quantum Classical Neural Network Quantum Feature Function Approximator