Low Generalization

Low generalization, the inability of a model to perform well on unseen data, is a significant challenge across various machine learning domains, hindering the reliable application of powerful models like deep neural networks and reinforcement learning agents. Current research focuses on improving generalization through techniques such as instance selection for training data, refined optimization methods like sharpness-aware minimization, and ensemble averaging to reduce prediction variance. Addressing this limitation is crucial for building robust and reliable AI systems applicable to diverse real-world problems, from time series forecasting to dynamic algorithm configuration.

Papers

July 18, 2024

Instance Selection for Dynamic Algorithm Configuration with Reinforcement Learning: Improving Generalization
Carolin Benjamins, Gjorgjina Cenikj, Ana Nikolikj, Aditya Mohan, Tome Eftimov, Marius Lindauer
Reinforcement Learning Deep Reinforcement Learning Strong Generalization Generalization Performance Instance Selection Algorithm Configuration Low Generalization

June 18, 2024

Generalization bounds for mixing processes via delayed online-to-PAC conversions
Baptiste Abeles, Eugenio Clerico, Gergely Neu
Strong Generalization Complex Process Statistical Learning Online Learning Algorithm Online to PAC Conversion Low Generalization

February 15, 2024

SAMformer: Unlocking the Potential of Transformers in Time Series Forecasting with Sharpness-Aware Minimization and Channel-Wise Attention
Romain Ilbert, Ambroise Odonnat, Vasilii Feofanov, Aladin Virmaux, Giuseppe Paolo, Themis Palpanas, Ievgen Redko
Transformer Megatron Decepticons Time Series Forecasting Channel Attention Sharpness Aware Minimization Long Term Forecasting Sharpness Aware Optimization Linear Forecasting Low Generalization

February 25, 2023

Explaining Generalization Power of a DNN Using Interactive Concepts
Huilin Zhou, Hao Zhang, Huiqi Deng, Dongrui Liu, Wen Shen, Shih-Han Chan, Quanshi Zhang
Deep Neural Network Strong Generalization Generalization Capacity Low Generalization

December 21, 2022

Uncontrolled Lexical Exposure Leads to Overestimation of Compositional Generalization in Pretrained Models
Najoung Kim, Tal Linzen, Paul Smolensky
Strong Generalization Compositional Generalization Lexical Knowledge Low Generalization

March 7, 2022

Robust Modeling of Unknown Dynamical Systems via Ensemble Averaged Learning
Victor Churchill, Steve Manns, Zhen Chen, Dongbin Xiu
Deep Neural Network Dynamical System Generalization Error DNN Model Robust Model Ensemble Averaging DNN Prediction Low Generalization

Low Generalization

Papers

Instance Selection for Dynamic Algorithm Configuration with Reinforcement Learning: Improving Generalization

Generalization bounds for mixing processes via delayed online-to-PAC conversions

SAMformer: Unlocking the Potential of Transformers in Time Series Forecasting with Sharpness-Aware Minimization and Channel-Wise Attention

Explaining Generalization Power of a DNN Using Interactive Concepts

Uncontrolled Lexical Exposure Leads to Overestimation of Compositional Generalization in Pretrained Models

Robust Modeling of Unknown Dynamical Systems via Ensemble Averaged Learning