Learning Algorithm

Learning algorithms are computational methods designed to automatically improve performance on a task through experience, aiming to create models that generalize well to unseen data. Current research emphasizes developing algorithms with improved efficiency and robustness, particularly in distributed and continual learning settings, exploring architectures like graph neural networks, transformers, and variations of gradient descent. These advancements are crucial for addressing challenges in diverse fields, including autonomous systems, healthcare, and scientific discovery, by enabling more efficient and reliable data analysis and decision-making.

Papers

February 8, 2022

Convergence of a New Learning Algorithm
Feng Lin
Neural Network Early Stage Convergence Back Propagation Learning Algorithm

January 28, 2022

A Simple Guard for Learned Optimizers
Isabeau Prémont-Schwarz, Jaroslav Vítků, Jan Feyereisl
Gradient Descent Learning Algorithm Better Optimizers Learned Optimizers General Purpose Air Guardian System L2O Model

December 4, 2021

Towards the One Learning Algorithm Hypothesis: A System-theoretic Approach
Christos Mavridis, John Baras
Human Cognition Learning Algorithm Hierarchical Attention Learning Model Progressive Learning Learning Architecture System Theoretic

November 26, 2021

Enforcing and Discovering Structure in Machine Learning
Francesco Locatello
Machine Learning Causal Structure Inner Structure Prior Knowledge Learning Algorithm Structure Discovery Exact Enforcement

November 17, 2021

CONFAIR: Configurable and Interpretable Algorithmic Fairness
Ankit Kulshrestha, Ilya Safro
Procedural Fairness Algorithmic Fairness Learning Algorithm Fair Algorithm Fairness Criterion Complex Machine Learning

November 9, 2021

August 11, 2021

Approximation and generalization properties of the random projection classification method
Mireille Boutin, Evzenie Coupkova
Average Approximation Learning Algorithm Generalization Gap Generalization Property Linear Classifier Random Projection