Learning Theory

Learning theory seeks to understand how and why machine learning algorithms generalize from training data to unseen examples. Current research focuses on refining generalization bounds for various settings, including dependent data and distribution regression, often employing techniques like Rademacher complexity and algorithmic stability to analyze model performance. This work is crucial for improving the reliability and efficiency of machine learning algorithms across diverse applications, from predicting dynamical systems to understanding the implicit biases of deep learning architectures like graph convolutional networks and neural networks designed for distribution inputs. Ultimately, these advancements aim to provide a more rigorous theoretical foundation for the design and application of machine learning models.

Papers

October 10, 2024

Exploring Natural Language-Based Strategies for Efficient Number Learning in Children through Reinforcement Learning
Tirthankar Mittra
Reinforcement Learning LeArning Abstract Deep Reinforcement Learning High Efficiency General Strategy Nine Year Old Child Numerical Reasoning Learning Theory Internal Numeracy

May 1, 2024

Error Exponent in Agnostic PAC Learning
Adi Hendel, Meir Feder
Statistical Learning Theory Probably Approximately Correct Learning Theory Agnostic PAC Error Scaling

February 9, 2024

The Complexity of Sequential Prediction in Dynamical Systems
Vinod Raman, Unique Subedi, Ambuj Tewari
Dynamical System Complexity Matter Sequential Prediction Optimal Error Learning Theory

February 8, 2024

Sharp Rates in Dependent Learning Theory: Avoiding Sample Size Deflation for the Square Loss
Ingvar Ziemann, Stephen Tu, George J. Pappas, Nikolai Matni
Statistical Learning Fast Rate Global Mixing Hypothesis Class Sample Size Sub Gaussian Square Loss Learning Theory

November 26, 2023

Applying statistical learning theory to deep learning
Cédric Gerbelot, Avetik Karagulyan, Stefani Karp, Kavya Ravichandran, Menachem Stern, Nathan Srebro
Deep Learning Gradient Descent Inductive Bias Mirror Descent Statistical Learning Theory Learning Theory

July 7, 2023

Learning Theory of Distribution Regression with Neural Networks
Zhongjie Shi, Zhan Yu, Ding-Xuan Zhou
Neural Network Regression Model Classical Neural Network Fully Connected Neural Network Learning Theory Distribution Regression

May 31, 2023

Hypothesis Transfer Learning with Surrogate Classification Losses: Generalization Bounds through Algorithmic Stability
Anass Aghbalou, Guillaume Staerman
Domain Adaptation Strong Generalization Machine Learning Algorithm Surrogate Loss Algorithmic Stability Learning Theory Hypothesis Transfer

September 7, 2022

A learning theory for quantum photonic processors and beyond
Matteo Rosati
Quantum Physic Quantum Circuit Quantum Advantage Learning Theory Continuous Variable Quantum Photonic Processor

April 17, 2022

Learning Theory of Mind via Dynamic Traits Attribution
Dung Nguyen, Phuoc Nguyen, Hung Le, Kien Do, Svetha Venkatesh, Truyen Tran
Latent Variable Human Mind Theory of Mind Learning Theory Trait Estimation Social Agent False Belief

February 9, 2022

The no-free-lunch theorems of supervised learning
Tom F. Sterkenburg, Peter D. Grünwald
Supervised Learning Inductive Bias Learning Algorithm Justification Theory Learning Theory

December 7, 2021

Learning Theory Can (Sometimes) Explain Generalisation in Graph Neural Networks
Pascal Mattia Esser, Leena Chennuru Vankadara, Debarghya Ghoshdastidar
Graph Neural Network Strong Generalization Graph Convolution Generalization Property Rademacher Complexity Transductive Inference Learning Theory