The unmanned aerial vehicle (UAV) network is popular these years due to its various applications. In the UAV network, routing is significantly affected by the distributed network topology, leading to the issue that UAVs are vulnerable to deliberate damage. Hence, this paper focuses on the routing plan and recovery for UAV networks with attacks. In detail, a deliberate attack model based on the importance of nodes is designed to represent enemy attacks. Then, a node importance ranking mechanism is presented, considering the degree of nodes and link importance. However, it is intractable to handle the routing problem by traditional methods for UAV networks, since link connections change with the UAV availability. Hence, an intelligent algorithm based on reinforcement learning is proposed to recover the routing path when UAVs are attacked. Simulations are conducted and numerical results verify the proposed mechanism performs better than other referred methods.

The emergence of price comparison websites (PCWs) has presented insurers with unique challenges in formulating effective pricing strategies. Operating on PCWs requires insurers to strike a delicate balance between competitive premiums and profitability, amidst obstacles such as low historical conversion rates, limited visibility of competitors' actions, and a dynamic market environment. In addition to this, the capital intensive nature of the business means pricing below the risk levels of customers can result in solvency issues for the insurer. To address these challenges, this paper introduces reinforcement learning (RL) framework that learns the optimal pricing policy by integrating model-based and model-free methods. The model-based component is used to train agents in an offline setting, avoiding cold-start issues, while model-free algorithms are then employed in a contextual bandit (CB) manner to dynamically update the pricing policy to maximise the expected revenue. This facilitates quick adaptation to evolving market dynamics and enhances algorithm efficiency and decision interpretability. The paper also highlights the importance of evaluating pricing policies using an offline dataset in a consistent fashion and demonstrates the superiority of the proposed methodology over existing off-the-shelf RL/CB approaches. We validate our methodology using synthetic data, generated to reflect private commercially available data within real-world insurers, and compare against 6 other benchmark approaches. Our hybrid agent outperforms these benchmarks in terms of sample efficiency and cumulative reward with the exception of an agent that has access to perfect market information which would not be available in a real-world set-up.

Dynamic Algorithm Configuration (DAC) tackles the question of how to automatically learn policies to control parameters of algorithms in a data-driven fashion. This question has received considerable attention from the evolutionary community in recent years. Having a good benchmark collection to gain structural understanding on the effectiveness and limitations of different solution methods for DAC is therefore strongly desirable. Following recent work on proposing DAC benchmarks with well-understood theoretical properties and ground truth information, in this work, we suggest as a new DAC benchmark the controlling of the key parameter $\lambda$ in the $(1+(\lambda,\lambda))$~Genetic Algorithm for solving OneMax problems. We conduct a study on how to solve the DAC problem via the use of (static) automated algorithm configuration on the benchmark, and propose techniques to significantly improve the performance of the approach. Our approach is able to consistently outperform the default parameter control policy of the benchmark derived from previous theoretical work on sufficiently large problem sizes. We also present new findings on the landscape of the parameter-control search policies and propose methods to compute stronger baselines for the benchmark via numerical approximations of the true optimal policies.

Deep learning-based recommender models (DLRMs) have become an essential component of many modern recommender systems. Several companies are now building large compute clusters reserved only for DLRM training, driving new interest in cost- and time- saving optimizations. The systems challenges faced in this setting are unique; while typical deep learning training jobs are dominated by model execution, the most important factor in DLRM training performance is often online data ingestion. In this paper, we explore the unique characteristics of this data ingestion problem and provide insights into DLRM training pipeline bottlenecks and challenges. We study real-world DLRM data processing pipelines taken from our compute cluster at Netflix to observe the performance impacts of online ingestion and to identify shortfalls in existing pipeline optimizers. We find that current tooling either yields sub-optimal performance, frequent crashes, or else requires impractical cluster re-organization to adopt. Our studies lead us to design and build a new solution for data pipeline optimization, InTune. InTune employs a reinforcement learning (RL) agent to learn how to distribute the CPU resources of a trainer machine across a DLRM data pipeline to more effectively parallelize data loading and improve throughput. Our experiments show that InTune can build an optimized data pipeline configuration within only a few minutes, and can easily be integrated into existing training workflows. By exploiting the responsiveness and adaptability of RL, InTune achieves higher online data ingestion rates than existing optimizers, thus reducing idle times in model execution and increasing efficiency. We apply InTune to our real-world cluster, and find that it increases data ingestion throughput by as much as 2.29X versus state-of-the-art data pipeline optimizers while also improving both CPU & GPU utilization.

This paper presents an extension of the Mirror Descent method to overcome challenges in cooperative Multi-Agent Reinforcement Learning (MARL) settings, where agents have varying abilities and individual policies. The proposed Heterogeneous-Agent Mirror Descent Policy Optimization (HAMDPO) algorithm utilizes the multi-agent advantage decomposition lemma to enable efficient policy updates for each agent while ensuring overall performance improvements. By iteratively updating agent policies through an approximate solution of the trust-region problem, HAMDPO guarantees stability and improves performance. Moreover, the HAMDPO algorithm is capable of handling both continuous and discrete action spaces for heterogeneous agents in various MARL problems. We evaluate HAMDPO on Multi-Agent MuJoCo and StarCraftII tasks, demonstrating its superiority over state-of-the-art algorithms such as HATRPO and HAPPO. These results suggest that HAMDPO is a promising approach for solving cooperative MARL problems and could potentially be extended to address other challenging problems in the field of MARL.

Learning to control an agent from data collected offline in a rich pixel-based visual observation space is vital for real-world applications of reinforcement learning (RL). A major challenge in this setting is the presence of input information that is hard to model and irrelevant to controlling the agent. This problem has been approached by the theoretical RL community through the lens of exogenous information, i.e, any control-irrelevant information contained in observations. For example, a robot navigating in busy streets needs to ignore irrelevant information, such as other people walking in the background, textures of objects, or birds in the sky. In this paper, we focus on the setting with visually detailed exogenous information, and introduce new offline RL benchmarks offering the ability to study this problem. We find that contemporary representation learning techniques can fail on datasets where the noise is a complex and time dependent process, which is prevalent in practical applications. To address these, we propose to use multi-step inverse models, which have seen a great deal of interest in the RL theory community, to learn Agent-Controller Representations for Offline-RL (ACRO). Despite being simple and requiring no reward, we show theoretically and empirically that the representation created by this objective greatly outperforms baselines.

Repeated principal-agent theory is a well-established paradigm that studies sequential interactions between two self-interested decision-makers. In particular, it offers a framework to analyze the problem of a primary party (i.e., principal) in a system who seeks to optimize the ultimate performance of the system by repeatedly delegating some operational control to another strategic party (i.e., agent) with a private decision-making process. This privacy imposes an information asymmetry between the principal and the agent that can appear as either an adverse selection setting, in which the information about the agent's true preferences or rewards are hidden from the principal, or a moral hazard setting, in which the actions chosen by the agent are hidden from the principal (Bolton and Dewatripont 2004). In either case, the principal's problem can be defined along two main dimensions: i) learning some private information about the agent by training a consistent estimator, ii) designing an incentive mechanism to lead the agent's algorithm in favor of the principal. In this paper, we study these two research problems for an unexplored adverse selection setting by marrying the classical principal-agent theory to statistics and reinforcement learning. In a repeated principal-agent game, the main theoretical challenge is sourced from the dynamic and sequential interactions taking place between the two strategic decision-makers. In each play of the game, first the principal offers a menu of incentives to the agent, and then the agent makes a choice from a finite set of actions, which in turn determines the rewards collected by both players. In other words, there is a two-sided sequential externality in this setting, whereby the agent's imperfect knowledge imposes additional costs on the principal and the principal's incentives impose a more challenging decision-making environment for the agent with imperfect knowledge. This paper considers that both the principal and the agent observe stochastic rewards with unknown (to both) expectations, and that both parties aim to maximize their own cumulative expected rewards Dogan et.

This work presents In-Context Policy Iteration, an algorithm for performing Reinforcement Learning (RL), in-context, using foundation models. While the application of foundation models to RL has received considerable attention, most approaches rely on either (1) the curation of expert demonstrations (either through manual design or task-specific pretraining) or (2) adaptation to the task of interest using gradient methods (either fine-tuning or training of adapter layers). Both of these techniques have drawbacks. Collecting demonstrations is labor-intensive, and algorithms that rely on them do not outperform the experts from which the demonstrations were derived. All gradient techniques are inherently slow, sacrificing the "few-shot" quality that made in-context learning attractive to begin with. In this work, we present an algorithm, ICPI, that learns to perform RL tasks without expert demonstrations or gradients. Instead we present a policy-iteration method in which the prompt content is the entire locus of learning. ICPI iteratively updates the contents of the prompt from which it derives its policy through trial-and-error interaction with an RL environment. In order to eliminate the role of in-weights learning (on which approaches like Decision Transformer rely heavily), we demonstrate our algorithm using Codex, a language model with no prior knowledge of the domains on which we evaluate it.

At present, the mainstream artificial intelligence generally adopts the technical path of "attention mechanism + deep learning" + "reinforcement learning". It has made great progress in the field of AIGC (Artificial Intelligence Generated Content), setting off the technical wave of big models[ 2][13 ]. But in areas that need to interact with the actual environment, such as elderly care, home nanny, agricultural production, and vehicle driving, trial and error are expensive and a reinforcement learning process that requires much trial and error is difficult to achieve. Therefore, in order to achieve Artificial General Intelligence(AGI) that can be applied to any field, we need to use both existing technologies and solve the defects of existing technologies, so as to further develop the technological wave of artificial intelligence. In this paper, we analyze the limitations of the technical route of large models, and by addressing these limitations, we propose solutions, thus solving the inherent defects of large models. In this paper, we will reveal how to achieve true AGI step by step.

Autonomous navigation in offroad environments has been extensively studied in the robotics field. However, navigation in covert situations where an autonomous vehicle needs to remain hidden from outside observers remains an underexplored area. In this paper, we propose a novel Deep Reinforcement Learning (DRL) based algorithm, called CoverNav, for identifying covert and navigable trajectories with minimal cost in offroad terrains and jungle environments in the presence of observers. CoverNav focuses on unmanned ground vehicles seeking shelters and taking covers while safely navigating to a predefined destination. Our proposed DRL method computes a local cost map that helps distinguish which path will grant the maximal covertness while maintaining a low cost trajectory using an elevation map generated from 3D point cloud data, the robot's pose, and directed goal information. CoverNav helps robot agents to learn the low elevation terrain using a reward function while penalizing it proportionately when it experiences high elevation. If an observer is spotted, CoverNav enables the robot to select natural obstacles (e.g., rocks, houses, disabled vehicles, trees, etc.) and use them as shelters to hide behind. We evaluate CoverNav using the Unity simulation environment and show that it guarantees dynamically feasible velocities in the terrain when fed with an elevation map generated by another DRL based navigation algorithm. Additionally, we evaluate CoverNav's effectiveness in achieving a maximum goal distance of 12 meters and its success rate in different elevation scenarios with and without cover objects. We observe competitive performance comparable to state of the art (SOTA) methods without compromising accuracy.