The field of reinforcement learning and control is rapidly advancing, with a focus on developing more efficient, robust, and adaptable algorithms. Recent research has explored the use of quantum computing, adversarial robustness, and multi-agent systems to improve the performance of reinforcement learning agents. Additionally, there has been a growing interest in developing more advanced control architectures, such as those using model predictive control and nonlinear model predictive control, to improve the stability and robustness of systems. Notable papers in this area include 'Quantum Boltzmann Machines for Sample-Efficient Reinforcement Learning', which introduces a novel quantum-classical model for reinforcement learning, and 'Adversarially Robust Multitask Adaptive Control', which proposes a clustered multitask approach to mitigate corrupted model updates. Other significant contributions include the development of novel control architectures, such as 'Stable and Robust SLIP Model Control via Energy Conservation-Based Feedback Cancellation for Quadrupedal Applications' and 'A Tilting-Rotor Enhanced Quadcopter Fault-Tolerant Control Based on Non-Linear Model Predictive Control'. These advancements have the potential to significantly impact a wide range of applications, from robotics and autonomous systems to finance and healthcare.
Advancements in Reinforcement Learning and Control
Sources
Quantum Boltzmann Machines for Sample-Efficient Reinforcement Learning
Adversarially Robust Multitask Adaptive Control
Stable and Robust SLIP Model Control via Energy Conservation-Based Feedback Cancellation for Quadrupedal Applications
A Tilting-Rotor Enhanced Quadcopter Fault-Tolerant Control Based on Non-Linear Model Predictive Control
SAD-Flower: Flow Matching for Safe, Admissible, and Dynamically Consistent Planning
Partial Action Replacement: Tackling Distribution Shift in Offline MARL
Distributed Adaptive Estimation over Sensor Networks with Partially Unknown Source Dynamics
Comparative Study of Q-Learning for State-Feedback LQG Control with an Unknown Model
Algorithm-Relative Trajectory Valuation in Policy Gradient Control
Dual-MPC Footstep Planning for Robust Quadruped Locomotion
Test-driven Reinforcement Learning
Multi-layer barrier function-based adaptive super-twisting controller
PrefPoE: Advantage-Guided Preference Fusion for Learning Where to Explore
X-IONet: Cross-Platform Inertial Odometry Network with Dual-Stage Attention
Beyond Distributions: Geometric Action Control for Continuous Reinforcement Learning
Stability of Certainty-Equivalent Adaptive LQR for Linear Systems with Unknown Time-Varying Parameters
Learning Omnidirectional Locomotion for a Salamander-Like Quadruped Robot
LPPG-RL: Lexicographically Projected Policy Gradient Reinforcement Learning with Subproblem Exploration
Recursive Binary Identification under Data Tampering and Non-Persistent Excitation with Application to Emission Control
Interpretable by Design: Query-Specific Neural Modules for Explainable Reinforcement Learning
Information-Driven Fault Detection and Identification for Multi-Agent Spacecraft Systems: Collaborative On-Orbit Inspection Mission
Diffusion Policies with Value-Conditional Optimization for Offline Reinforcement Learning
APEX: Action Priors Enable Efficient Exploration for Robust Motion Tracking on Legged Robots
Data Fusion-Enhanced Decision Transformer for Stable Cross-Domain Generalization
Robust Estimation and Control for Heterogeneous Multi-agent Systems Based on Decentralized k-hop Prescribed Performance Observers
MARBLE: Multi-Armed Restless Bandits in Latent Markovian Environment
Quantum Meet-in-the-Middle Attacks on Key-Length Extension Constructions
Quasi-Newton Compatible Actor-Critic for Deterministic Policies