This paper presents a {\delta}-PI algorithm which is based on damped Newton method for the H{\infty} tracking control problem of unknown continuous-time nonlinear system. A discounted performance function and an augmented system are used to get the tracking Hamilton-Jacobi-Isaac (HJI) equation. Tracking HJI equation is a nonlinear partial differential equation, traditional reinforcement learning methods for solving the tracking HJI equation are mostly based on the Newton method, which usually only satisfies local convergence and needs a good initial guess. Based upon the damped Newton iteration operator equation, a generalized tracking Bellman equation is derived firstly. The {\delta}-PI algorithm can seek the optimal solution of the tracking HJI equation by iteratively solving the generalized tracking Bellman equation. On-policy learning and off-policy learning {\delta}-PI reinforcement learning methods are provided, respectively. Off-policy version {\delta}-PI algorithm is a model-free algorithm which can be performed without making use of a priori knowledge of the system dynamics. NN-based implementation scheme for the off-policy {\delta}-PI algorithms is shown. The suitability of the model-free {\delta}-PI algorithm is illustrated with a nonlinear system simulation.

The inability to naturally enforce safety in Reinforcement Learning (RL), with limited failures, is a core challenge impeding its use in real-world applications. One notion of safety of vast practical relevance is the ability to avoid (unsafe) regions of the state space. Though such a safety goal can be captured by an action-value-like function, a.k.a. safety critics, the associated operator lacks the desired contraction and uniqueness properties that the classical Bellman operator enjoys. In this work, we overcome the non-contractiveness of safety critic operators by leveraging that safety is a binary property. To that end, we study the properties of the binary safety critic associated with a deterministic dynamical system that seeks to avoid reaching an unsafe region. We formulate the corresponding binary Bellman equation (B2E) for safety and study its properties. While the resulting operator is still non-contractive, we fully characterize its fixed points representing--except for a spurious solution--maximal persistently safe regions of the state space that can always avoid failure. We provide an algorithm that, by design, leverages axiomatic knowledge of safe data to avoid spurious fixed points.

Phosphorus removal is vital in wastewater treatment to reduce reliance on limited resources. Deep reinforcement learning (DRL) is a machine learning technique that can optimize complex and nonlinear systems, including the processes in wastewater treatment plants, by learning control policies through trial and error. However, applying DRL to chemical and biological processes is challenging due to the need for accurate simulators. This study trained six models to identify the phosphorus removal process and used them to create a simulator for the DRL environment. Although the models achieved high accuracy (>97%), uncertainty and incorrect prediction behavior limited their performance as simulators over longer horizons. Compounding errors in the models' predictions were identified as one of the causes of this problem. This approach for improving process control involves creating simulation environments for DRL algorithms, using data from supervisory control and data acquisition (SCADA) systems with a sufficient historical horizon without complex system modeling or parameter estimation.

In the pursuit of reducing the number of trainable parameters in deep transformer networks, we employ Reinforcement Learning to dynamically select layers during training and tie them together. Every few iterations, the RL agent is asked whether to train each layer $i$ independently or to copy the weights of a previous layer $j

Learning the behavior of large agent populations is an important task for numerous research areas. Although the field of multi-agent reinforcement learning (MARL) has made significant progress towards solving these systems, solutions for many agents often remain computationally infeasible and lack theoretical guarantees. Mean Field Games (MFGs) address both of these issues and can be extended to Graphon MFGs (GMFGs) to include network structures between agents. Despite their merits, the real world applicability of GMFGs is limited by the fact that graphons only capture dense graphs. Since most empirically observed networks show some degree of sparsity, such as power law graphs, the GMFG framework is insufficient for capturing these network topologies. Thus, we introduce the novel concept of Graphex MFGs (GXMFGs) which builds on the graph theoretical concept of graphexes. Graphexes are the limiting objects to sparse graph sequences that also have other desirable features such as the small world property. Learning equilibria in these games is challenging due to the rich and sparse structure of the underlying graphs. To tackle these challenges, we design a new learning algorithm tailored to the GXMFG setup. This hybrid graphex learning approach leverages that the system mainly consists of a highly connected core and a sparse periphery. After defining the system and providing a theoretical analysis, we state our learning approach and demonstrate its learning capabilities on both synthetic graphs and real-world networks. This comparison shows that our GXMFG learning algorithm successfully extends MFGs to a highly relevant class of hard, realistic learning problems that are not accurately addressed by current MARL and MFG methods.

Interactive Machine Learning (IML) seeks to integrate human expertise into machine learning processes. However, most existing algorithms cannot be applied to Realworld Scenarios because their state spaces and/or action spaces are limited to discrete values. Furthermore, the interaction of all existing methods is restricted to deciding between multiple proposals. We therefore propose a novel framework based on Bayesian Optimization (BO). Interactive Bayesian Optimization (IBO) enables collaboration between machine learning algorithms and humans. This framework captures user preferences and provides an interface for users to shape the strategy by hand. Additionally, we've incorporated a new acquisition function, Preference Expected Improvement (PEI), to refine the system's efficiency using a probabilistic model of the user preferences. Our approach is geared towards ensuring that machines can benefit from human expertise, aiming for a more aligned and effective learning process. In the course of this work, we applied our method to simulations and in a real world task using a Franka Panda robot to show human-robot collaboration.

Understanding the emergence of cooperation in systems of computational agents is crucial for the development of effective cooperative AI. Interaction among individuals in real-world settings are often sparse and occur within a broad spectrum of incentives, which often are only partially known. In this work, we explore how cooperation can arise among reinforcement learning agents in scenarios characterised by infrequent encounters, and where agents face uncertainty about the alignment of their incentives with those of others. To do so, we train the agents under a wide spectrum of environments ranging from fully competitive, to fully cooperative, to mixed-motives. Under this type of uncertainty we study the effects of mechanisms, such as reputation and intrinsic rewards, that have been proposed in the literature to foster cooperation in mixed-motives environments. Our findings show that uncertainty substantially lowers the agents' ability to engage in cooperative behaviour, when that would be the best course of action. In this scenario, the use of effective reputation mechanisms and intrinsic rewards boosts the agents' capability to act nearly-optimally in cooperative environments, while greatly enhancing cooperation in mixed-motive environments as well.

In this work, we compare emergent communication (EC) built upon multi-agent deep reinforcement learning (MADRL) and language-oriented semantic communication (LSC) empowered by a pre-trained large language model (LLM) using human language. In a multi-agent remote navigation task, with multimodal input data comprising location and channel maps, it is shown that EC incurs high training cost and struggles when using multimodal data, whereas LSC yields high inference computing cost due to the LLM's large size. To address their respective bottlenecks, we propose a novel framework of language-guided EC (LEC) by guiding the EC training using LSC via knowledge distillation (KD). Simulations corroborate that LEC achieves faster travel time while avoiding areas with poor channel conditions, as well as speeding up the MADRL training convergence by up to 61.8% compared to EC.

Two desiderata of reinforcement learning (RL) algorithms are the ability to learn from relatively little experience and the ability to learn policies that generalize to a range of problem specifications. In factored state spaces, one approach towards achieving both goals is to learn state abstractions, which only keep the necessary variables for learning the tasks at hand. This paper introduces Causal Bisimulation Modeling (CBM), a method that learns the causal relationships in the dynamics and reward functions for each task to derive a minimal, task-specific abstraction. CBM leverages and improves implicit modeling to train a high-fidelity causal dynamics model that can be reused for all tasks in the same environment. Empirical validation on manipulation environments and Deepmind Control Suite reveals that CBM's learned implicit dynamics models identify the underlying causal relationships and state abstractions more accurately than explicit ones. Furthermore, the derived state abstractions allow a task learner to achieve near-oracle levels of sample efficiency and outperform baselines on all tasks.

Soft growing robots, are a type of robots that are designed to move and adapt to their environment in a similar way to how plants grow and move with potential applications where they could be used to navigate through tight spaces, dangerous terrain, and hard-to-reach areas. This research explores the application of deep reinforcement Q-learning algorithm for facilitating the navigation of the soft growing robots in cluttered environments. The proposed algorithm utilizes the flexibility of the soft robot to adapt and incorporate the interaction between the robot and the environment into the decision-making process. Results from simulations show that the proposed algorithm improves the soft robot's ability to navigate effectively and efficiently in confined spaces. This study presents a promising approach to addressing the challenges faced by growing robots in particular and soft robots general in planning obstacle-aware paths in real-world scenarios.