Trajectory Truncation

Trajectory truncation involves shortening or selectively modifying trajectories, sequences of states or actions, to improve efficiency or performance in various applications. Current research focuses on leveraging truncation to enhance reinforcement learning algorithms, particularly by improving the stability of value estimation, optimizing data collection strategies in Monte Carlo simulations, and mitigating uncertainty in model-based offline reinforcement learning. This technique is proving valuable in diverse fields, from improving the efficiency of drone swarm coordination through control-aware trajectory prediction to enhancing data compression and reconstruction in GPS tracking. The overall impact lies in improving the efficiency and robustness of algorithms across various domains by strategically managing trajectory information.

Papers

October 17, 2024

Truncating Trajectories in Monte Carlo Policy Evaluation: an Adaptive Approach
Riccardo Poiani, Nicole Nobili, Alberto Maria Metelli, Marcello Restelli
Reinforcement Learning Constructive Approach Adaptive Importance Monte Carlo Trajectory Truncation

May 31, 2024

Target Networks and Over-parameterization Stabilize Off-policy Bootstrapping with Function Approximation
Fengdi Che, Chenjun Xiao, Jincheng Mei, Bo Dai, Ramki Gummadi, Oscar A Ramirez, Christopher K Harris, A. Rupam Mahmood, Dale Schuurmans
Function Approximation Episodic Markov Decision Process Bootstrapping End to End Goal Node Trajectory Truncation

January 23, 2024

Control-Aware Trajectory Predictions for Communication-Efficient Drone Swarm Coordination in Cluttered Environments
Longhao Yan, Jingyuan Zhou, Kaidi Yang
Unmanned Aerial Vehicle Various Fast Moving Drone Cluttered Environment Trajectory Truncation

May 7, 2023

Truncating Trajectories in Monte Carlo Reinforcement Learning
Riccardo Poiani, Alberto Maria Metelli, Marcello Restelli
Reinforcement Learning Policy OpTimization Trajectory Truncation

April 10, 2023

Uncertainty-driven Trajectory Truncation for Data Augmentation in Offline Reinforcement Learning
Junjie Zhang, Jiafei Lyu, Xiaoteng Ma, Jiangpeng Yan, Jun Yang, Le Wan, Xiu Li
Data Augmentation Offline Reinforcement Learning Model Based Offline Reinforcement Learning Synthetic Trajectory Trajectory Truncation

January 18, 2023

Compression of GPS Trajectories using Autoencoders
Michael Kölle, Steffen Illium, Carsten Hahn, Lorenz Schauer, Johannes Hutter, Claudia Linnhoff-Popien
Supervised Autoencoder Linear Compression Dynamic Time Warping GPS Trajectory LSTM Autoencoder Trajectory Truncation