Heterogeneity Environment

Heterogeneity in environments, encompassing variations in data distribution, hardware capabilities, and environmental dynamics, poses significant challenges across diverse machine learning applications. Current research focuses on developing robust algorithms and models, such as federated learning approaches with adaptive client selection and novel neural network architectures, to mitigate the negative impacts of this heterogeneity on model training and performance. These efforts aim to improve the accuracy, efficiency, and generalizability of machine learning models in complex, real-world scenarios, with applications ranging from reinforcement learning to distributed computing and environmental monitoring. The ultimate goal is to create more reliable and adaptable systems capable of handling the inherent variability found in many practical settings.

Papers

February 2, 2024

FedShift: Tackling Dual Heterogeneity Problem of Federated Learning via Weight Shift Aggregation
Jungwon Seo, Chunming Rong, Minhoe Kim
Weight Aggregation Heterogeneity Challenge System Heterogeneity Unknown Heterogeneity Heterogeneity Environment

June 14, 2023

Off-policy Evaluation in Doubly Inhomogeneous Environments
Zeyu Bian, Chengchun Shi, Zhengling Qi, Lan Wang
Reinforcement Learning Policy Evaluation Homogeneity Bias Heterogeneity Environment Testing Stationarity

May 18, 2023

Client Selection for Federated Policy Optimization with Environment Heterogeneity
Zhijie Xie, S. H. Song
Reinforcement Learning Policy Gradient Policy Iteration Client Selection Centralized Learning Heterogeneity Environment

February 4, 2023

Federated Temporal Difference Learning with Linear Function Approximation under Environmental Heterogeneity
Han Wang, Aritra Mitra, Hamed Hassani, George J. Pappas, James Anderson
Markov Decision Process Linear Function Approximation Federated Reinforcement Learning Finite Time Markovian Sampling MDP Model Heterogeneity Environment

October 4, 2022

Stateful active facilitator: Coordination and Environmental Heterogeneity in Cooperative Multi-Agent Reinforcement Learning
Dianbo Liu, Vedant Shah, Oussama Boussif, Cristian Meo, Anirudh Goyal, Tianmin Shu, Michael Mozer, Nicolas Heess, Yoshua Bengio
Cooperative Multi Agent Reinforcement Learning Prior Coordination Heterogeneity Environment Complex Coordination

July 22, 2022

JAM: The JavaScript Agent Machine for Distributed Computing and Simulation with reactive and mobile Multi-agent Systems -- A Technical Report
Stefan Bosse
Multi Agent Simulation Study Web Agent Reactive Behavior Heterogeneity Environment

April 6, 2022

Federated Reinforcement Learning with Environment Heterogeneity
Hao Jin, Yang Peng, Wenhao Yang, Shusen Wang, Zhihua Zhang
Federated Reinforcement Learning Policy Design Agent Environment Interaction Heterogeneity Environment

December 27, 2021

Estimating Parameters of the Tree Root in Heterogeneous Soil Environments via Mask-Guided Multi-Polarimetric Integration Neural Network
Hai-Han Sun, Yee Hui Lee, Qiqi Dai, Chongyi Li, Genevieve Ow, Mohamed Lokman Mohd Yusof, Abdulkadir C. Yucel
Parameter Estimation Polarimetric Synthetic Aperture Radar SHUSS Root Exploration Term Heterogeneity Environment

September 24, 2021

Optimization-based Causal Estimation from Heterogenous Environments
Mingzhang Yin, Yixin Wang, David M. Blei
Causal Inference Causal Model Causal Relationship Causal Effect Estimation Causal Explanation Heterogeneity Environment

Heterogeneity Environment

Papers

FedShift: Tackling Dual Heterogeneity Problem of Federated Learning via Weight Shift Aggregation

Off-policy Evaluation in Doubly Inhomogeneous Environments

Client Selection for Federated Policy Optimization with Environment Heterogeneity

Federated Temporal Difference Learning with Linear Function Approximation under Environmental Heterogeneity

Stateful active facilitator: Coordination and Environmental Heterogeneity in Cooperative Multi-Agent Reinforcement Learning

JAM: The JavaScript Agent Machine for Distributed Computing and Simulation with reactive and mobile Multi-agent Systems -- A Technical Report

Federated Reinforcement Learning with Environment Heterogeneity

Estimating Parameters of the Tree Root in Heterogeneous Soil Environments via Mask-Guided Multi-Polarimetric Integration Neural Network

Optimization-based Causal Estimation from Heterogenous Environments