PyTorchKR Papers

최신 논문

총 13,743편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2501.07146

TIMRL: A Novel Meta-Reinforcement Learning Framework for Non-Stationary and Multi-Task Environments

In recent years, meta-reinforcement learning (meta-RL) algorithm has been proposed to improve sample efficiency in the field of decision-making and control, enabling agents to learn new knowledge from a small number of samples. However, most research uses the Gaussian distribution to extract task representation, which is poorly adapted to tasks that change in non-stationary environment. To address this problem, we propose a novel meta-reinforcement learning method by leveraging Gaussian mixture model and the transformer network to construct task inference model. The Gaussian mixture model is utilized to extend the task representation and conduct explicit encoding of tasks. Specifically, the classification of tasks is encoded through transformer network to determine the Gaussian component corresponding to the task. By leveraging task labels, the transformer network is trained using supervised learning. We validate our method on MuJoCo benchmarks with non-stationary and multi-task environments. Experimental results demonstrate that the proposed method dramatically improves sample efficiency and accurately recognizes the classification of the tasks, while performing excellently in the environment.

arXiv:2501.06404공식 코드 · Papers with Code

A Hybrid Framework for Reinsurance Optimization: Integrating Generative Models and Reinforcement Learning

Reinsurance optimization is critical for insurers to manage risk exposure, ensure financial stability, and maintain solvency. Traditional approaches often struggle with dynamic claim distributions, high-dimensional constraints, and evolving market conditions. This paper introduces a novel hybrid framework that integrates {Generative Models}, specifically Variational Autoencoders (VAEs), with {Reinforcement Learning (RL)} using Proximal Policy Optimization (PPO). The framework enables dynamic and scalable optimization of reinsurance strategies by combining the generative modeling of complex claim distributions with the adaptive decision-making capabilities of reinforcement learning. The VAE component generates synthetic claims, including rare and catastrophic events, addressing data scarcity and variability, while the PPO algorithm dynamically adjusts reinsurance parameters to maximize surplus and minimize ruin probability. The framework's performance is validated through extensive experiments, including out-of-sample testing, stress-testing scenarios (e.g., pandemic impacts, catastrophic events), and scalability analysis across portfolio sizes. Results demonstrate its superior adaptability, scalability, and robustness compared to traditional optimization techniques, achieving higher final surpluses and computational efficiency. Key contributions include the development of a hybrid approach for high-dimensional optimization, dynamic reinsurance parameterization, and validation against stochastic claim distributions. The proposed framework offers a transformative solution for modern reinsurance challenges, with potential applications in multi-line insurance operations, catastrophe modeling, and risk-sharing strategy design.

arXiv:2305.14133공식 코드 · Papers with Code

Conditional Mutual Information for Disentangled Representations in Reinforcement Learning

Reinforcement Learning (RL) environments can produce training data with spurious correlations between features due to the amount of training data or its limited feature coverage. This can lead to RL agents encoding these misleading correlations in their latent representation, preventing the agent from generalising if the correlation changes within the environment or when deployed in the real world. Disentangled representations can improve robustness, but existing disentanglement techniques that minimise mutual information between features require independent features, thus they cannot disentangle correlated features. We propose an auxiliary task for RL algorithms that learns a disentangled representation of high-dimensional observations with correlated features by minimising the conditional mutual information between features in the representation. We demonstrate experimentally, using continuous control tasks, that our approach improves generalisation under correlation shifts, as well as improving the training performance of RL algorithms in the presence of correlated features.

arXiv:2501.05329

Knowledge Transfer in Model-Based Reinforcement Learning Agents for Efficient Multi-Task Learning

We propose an efficient knowledge transfer approach for model-based reinforcement learning, addressing the challenge of deploying large world models in resource-constrained environments. Our method distills a high-capacity multi-task agent (317M parameters) into a compact 1M parameter model, achieving state-of-the-art performance on the MT30 benchmark with a normalized score of 28.45, a substantial improvement over the original 1M parameter model's score of 18.93. This demonstrates the ability of our distillation technique to consolidate complex multi-task knowledge effectively. Additionally, we apply FP16 post-training quantization, reducing the model size by 50% while maintaining performance. Our work bridges the gap between the power of large models and practical deployment constraints, offering a scalable solution for efficient and accessible multi-task reinforcement learning in robotics and other resource-limited domains.

arXiv:2501.04870

Deep Transfer $Q$-Learning for Offline Non-Stationary Reinforcement Learning

In dynamic decision-making scenarios across business and healthcare, leveraging sample trajectories from diverse populations can significantly enhance reinforcement learning (RL) performance for specific target populations, especially when sample sizes are limited. While existing transfer learning methods primarily focus on linear regression settings, they lack direct applicability to reinforcement learning algorithms. This paper pioneers the study of transfer learning for dynamic decision scenarios modeled by non-stationary finite-horizon Markov decision processes, utilizing neural networks as powerful function approximators and backward inductive learning. We demonstrate that naive sample pooling strategies, effective in regression settings, fail in Markov decision processes.To address this challenge, we introduce a novel ``re-weighted targeting procedure'' to construct ``transferable RL samples'' and propose ``transfer deep $Q^*$-learning'', enabling neural network approximation with theoretical guarantees. We assume that the reward functions are transferable and deal with both situations in which the transition densities are transferable or nontransferable. Our analytical techniques for transfer learning in neural network approximation and transition density transfers have broader implications, extending to supervised transfer learning with neural networks and domain shift scenarios. Empirical experiments on both synthetic and real datasets corroborate the advantages of our method, showcasing its potential for improving decision-making through strategically constructing transferable RL samples in non-stationary reinforcement learning contexts.

arXiv:2501.05408

TimeRL: Efficient Deep Reinforcement Learning with Polyhedral Dependence Graphs

Modern deep learning (DL) workloads increasingly use complex deep reinforcement learning (DRL) algorithms that generate training data within the learning loop. This results in programs with several nested loops and dynamic data dependencies between tensors. While DL systems with eager execution support such dynamism, they lack the optimizations and smart scheduling of graph-based execution. Graph-based execution, however, cannot express dynamic tensor shapes, instead requiring the use of multiple static subgraphs. Either execution model for DRL thus leads to redundant computation, reduced parallelism, and less efficient memory management. We describe TimeRL, a system for executing dynamic DRL programs that combines the dynamism of eager execution with the whole-program optimizations and scheduling of graph-based execution. TimeRL achieves this by introducing the declarative programming model of recurrent tensors, which allows users to define dynamic dependencies as intuitive recurrence equations. TimeRL translates recurrent tensors into a polyhedral dependence graph (PDG) with dynamic dependencies as symbolic expressions. Through simple PDG transformations, TimeRL applies whole-program optimizations, such as automatic vectorization, incrementalization, and operator fusion. The PDG also allows for the computation of an efficient program-wide execution schedule, which decides on buffer deallocations, buffer donations, and GPU/CPU memory swapping. We show that TimeRL executes current DRL algorithms up to 47$\times$ faster than existing DRL systems, while using 16$\times$ less GPU peak memory.

arXiv:2501.04982

CuRLA: Curriculum Learning Based Deep Reinforcement Learning for Autonomous Driving

In autonomous driving, traditional Computer Vision (CV) agents often struggle in unfamiliar situations due to biases in the training data. Deep Reinforcement Learning (DRL) agents address this by learning from experience and maximizing rewards, which helps them adapt to dynamic environments. However, ensuring their generalization remains challenging, especially with static training environments. Additionally, DRL models lack transparency, making it difficult to guarantee safety in all scenarios, particularly those not seen during training. To tackle these issues, we propose a method that combines DRL with Curriculum Learning for autonomous driving. Our approach uses a Proximal Policy Optimization (PPO) agent and a Variational Autoencoder (VAE) to learn safe driving in the CARLA simulator. The agent is trained using two-fold curriculum learning, progressively increasing environment difficulty and incorporating a collision penalty in the reward function to promote safety. This method improves the agent's adaptability and reliability in complex environments, and understand the nuances of balancing multiple reward components from different feedback signals in a single scalar reward function. Keywords: Computer Vision, Deep Reinforcement Learning, Variational Autoencoder, Proximal Policy Optimization, Curriculum Learning, Autonomous Driving.

arXiv:2501.04538

HypeRL: Parameter-Informed Reinforcement Learning for Parametric PDEs

In this work, we devise a new, general-purpose reinforcement learning strategy for the optimal control of parametric partial differential equations (PDEs). Such problems frequently arise in applied sciences and engineering and entail a significant complexity when control and/or state variables are distributed in high-dimensional space or depend on varying parameters. Traditional numerical methods, relying on either iterative minimization algorithms or dynamic programming, while reliable, often become computationally infeasible. Indeed, in either way, the optimal control problem must be solved for each instance of the parameters, and this is out of reach when dealing with high-dimensional time-dependent and parametric PDEs. In this paper, we propose HypeRL, a deep reinforcement learning (DRL) framework to overcome the limitations shown by traditional methods. HypeRL aims at approximating the optimal control policy directly. Specifically, we employ an actor-critic DRL approach to learn an optimal feedback control strategy that can generalize across the range of variation of the parameters. To effectively learn such optimal control laws, encoding the parameter information into the DRL policy and value function neural networks (NNs) is essential. To do so, HypeRL uses two additional NNs, often called hypernetworks, to learn the weights and biases of the value function and the policy NNs. We validate the proposed approach on two PDE-constrained optimal control benchmarks, namely a 1D Kuramoto-Sivashinsky equation and a 2D Navier-Stokes equations, by showing that the knowledge of the PDE parameters and how this information is encoded, i.e., via a hypernetwork, is an essential ingredient for learning parameter-dependent control policies that can generalize effectively to unseen scenarios and for improving the sample efficiency of such policies.

arXiv:2501.04481

Safe Reinforcement Learning with Minimal Supervision

Reinforcement learning (RL) in the real world necessitates the development of procedures that enable agents to explore without causing harm to themselves or others. The most successful solutions to the problem of safe RL leverage offline data to learn a safe-set, enabling safe online exploration. However, this approach to safe-learning is often constrained by the demonstrations that are available for learning. In this paper we investigate the influence of the quantity and quality of data used to train the initial safe learning problem offline on the ability to learn safe-RL policies online. Specifically, we focus on tasks with spatially extended goal states where we have few or no demonstrations available. Classically this problem is addressed either by using hand-designed controllers to generate data or by collecting user-generated demonstrations. However, these methods are often expensive and do not scale to more complex tasks and environments. To address this limitation we propose an unsupervised RL-based offline data collection procedure, to learn complex and scalable policies without the need for hand-designed controllers or user demonstrations. Our research demonstrates the significance of providing sufficient demonstrations for agents to learn optimal safe-RL policies online, and as a result, we propose optimistic forgetting, a novel online safe-RL approach that is practical for scenarios with limited data. Further, our unsupervised data collection approach highlights the need to balance diversity and optimality for safe online exploration.

arXiv:2501.04472

Hybrid Artificial Intelligence Strategies for Drone Navigation

Objective: This paper describes the development of hybrid artificial intelligence strategies for drone navigation. Methods: The navigation module combines a deep learning model with a rule-based engine depending on the agent state. The deep learning model has been trained using reinforcement learning. The rule-based engine uses expert knowledge to deal with specific situations. The navigation module incorporates several strategies to explain the drone decision based on its observation space, and different mechanisms for including human decisions in the navigation process. Finally, this paper proposes an evaluation methodology based on defining several scenarios and analyzing the performance of the different strategies according to metrics adapted to each scenario. Results: Two main navigation problems have been studied. For the first scenario (reaching known targets), it has been possible to obtain a 90% task completion rate, reducing significantly the number of collisions thanks to the rule-based engine. For the second scenario, it has been possible to reduce 20% of the time required to locate all the targets using the reinforcement learning model. Conclusions: Reinforcement learning is a very good strategy to learn policies for drone navigation, but in critical situations, it is necessary to complement it with a rule-based module to increase task success rate.

arXiv:2501.04228

Constraints as Rewards: Reinforcement Learning for Robots without Reward Functions

Reinforcement learning has become an essential algorithm for generating complex robotic behaviors. However, to learn such behaviors, it is necessary to design a reward function that describes the task, which often consists of multiple objectives that needs to be balanced. This tuning process is known as reward engineering and typically involves extensive trial-and-error. In this paper, to avoid this trial-and-error process, we propose the concept of Constraints as Rewards (CaR). CaR formulates the task objective using multiple constraint functions instead of a reward function and solves a reinforcement learning problem with constraints using the Lagrangian-method. By adopting this approach, different objectives are automatically balanced, because Lagrange multipliers serves as the weights among the objectives. In addition, we will demonstrate that constraints, expressed as inequalities, provide an intuitive interpretation of the optimization target designed for the task. We apply the proposed method to the standing-up motion generation task of a six-wheeled-telescopic-legged robot and demonstrate that the proposed method successfully acquires the target behavior, even though it is challenging to learn with manually designed reward functions.

arXiv:2501.04068

Explainable Reinforcement Learning for Formula One Race Strategy

In Formula One, teams compete to develop their cars and achieve the highest possible finishing position in each race. During a race, however, teams are unable to alter the car, so they must improve their cars' finishing positions via race strategy, i.e. optimising their selection of which tyre compounds to put on the car and when to do so. In this work, we introduce a reinforcement learning model, RSRL (Race Strategy Reinforcement Learning), to control race strategies in simulations, offering a faster alternative to the industry standard of hard-coded and Monte Carlo-based race strategies. Controlling cars with a pace equating to an expected finishing position of P5.5 (where P1 represents first place and P20 is last place), RSRL achieves an average finishing position of P5.33 on our test race, the 2023 Bahrain Grand Prix, outperforming the best baseline of P5.63. We then demonstrate, in a generalisability study, how performance for one track or multiple tracks can be prioritised via training. Further, we supplement model predictions with feature importance, decision tree-based surrogate models, and decision tree counterfactuals towards improving user trust in the model. Finally, we provide illustrations which exemplify our approach in real-world situations, drawing parallels between simulations and reality.

arXiv:2501.03687

Run-and-tumble chemotaxis using reinforcement learning

Bacterial cells use run-and-tumble motion to climb up attractant concentration gradient in their environment. By extending the uphill runs and shortening the downhill runs the cells migrate towards the higher attractant zones. Motivated by this, we formulate a reinforcement learning (RL) algorithm where an agent moves in one dimension in the presence of an attractant gradient. The agent can perform two actions: either persistent motion in the same direction or reversal of direction. We assign costs for these actions based on the recent history of the agent's trajectory. We ask the question: which RL strategy works best in different types of attractant environment. We quantify efficiency of the RL strategy by the ability of the agent (a) to localize in the favorable zones after large times, and (b) to learn about its complete environment. Depending on the attractant profile and the initial condition, we find an optimum balance is needed between exploration and exploitation to ensure the most efficient performance.

arXiv:2501.03902

Explainable Reinforcement Learning via Temporal Policy Decomposition

We investigate the explainability of Reinforcement Learning (RL) policies from a temporal perspective, focusing on the sequence of future outcomes associated with individual actions. In RL, value functions compress information about rewards collected across multiple trajectories and over an infinite horizon, allowing a compact form of knowledge representation. However, this compression obscures the temporal details inherent in sequential decision-making, presenting a key challenge for interpretability. We present Temporal Policy Decomposition (TPD), a novel explainability approach that explains individual RL actions in terms of their Expected Future Outcome (EFO). These explanations decompose generalized value functions into a sequence of EFOs, one for each time step up to a prediction horizon of interest, revealing insights into when specific outcomes are expected to occur. We leverage fixed-horizon temporal difference learning to devise an off-policy method for learning EFOs for both optimal and suboptimal actions, enabling contrastive explanations consisting of EFOs for different state-action pairs. Our experiments demonstrate that TPD generates accurate explanations that (i) clarify the policy's future strategy and anticipated trajectory for a given action and (ii) improve understanding of the reward composition, facilitating fine-tuning of the reward function to align with human expectations.

arXiv:2501.03824

Online Reinforcement Learning-Based Dynamic Adaptive Evaluation Function for Real-Time Strategy Tasks

Effective evaluation of real-time strategy tasks requires adaptive mechanisms to cope with dynamic and unpredictable environments. This study proposes a method to improve evaluation functions for real-time responsiveness to battle-field situation changes, utilizing an online reinforcement learning-based dynam-ic weight adjustment mechanism within the real-time strategy game. Building on traditional static evaluation functions, the method employs gradient descent in online reinforcement learning to update weights dynamically, incorporating weight decay techniques to ensure stability. Additionally, the AdamW optimizer is integrated to adjust the learning rate and decay rate of online reinforcement learning in real time, further reducing the dependency on manual parameter tun-ing. Round-robin competition experiments demonstrate that this method signifi-cantly enhances the application effectiveness of the Lanchester combat model evaluation function, Simple evaluation function, and Simple Sqrt evaluation function in planning algorithms including IDABCD, IDRTMinimax, and Port-folio AI. The method achieves a notable improvement in scores, with the en-hancement becoming more pronounced as the map size increases. Furthermore, the increase in evaluation function computation time induced by this method is kept below 6% for all evaluation functions and planning algorithms. The pro-posed dynamic adaptive evaluation function demonstrates a promising approach for real-time strategy task evaluation.

arXiv:2501.03676

SALE-Based Offline Reinforcement Learning with Ensemble Q-Networks

In this work, we build upon the offline reinforcement learning algorithm TD7, which incorporates State-Action Learned Embeddings (SALE) and a prioritized experience replay buffer (LAP). We propose a model-free actor-critic algorithm that integrates ensemble Q-networks and a gradient diversity penalty from EDAC. The ensemble Q-networks introduce penalties to guide the actor network toward in-distribution actions, effectively addressing the challenge of out-of-distribution actions. Meanwhile, the gradient diversity penalty encourages diverse Q-value gradients, further suppressing overestimation for out-of-distribution actions. Additionally, our method retains an adjustable behavior cloning (BC) term that directs the actor network toward dataset actions during early training stages, while gradually reducing its influence as the precision of the Q-ensemble improves. These enhancements work synergistically to improve the stability and precision of the training. Experimental results on the D4RL MuJoCo benchmarks demonstrate that our algorithm achieves higher convergence speed, stability, and performance compared to existing methods.

arXiv:2501.03562

Rethinking Adversarial Attacks in Reinforcement Learning from Policy Distribution Perspective

Deep Reinforcement Learning (DRL) suffers from uncertainties and inaccuracies in the observation signal in realworld applications. Adversarial attack is an effective method for evaluating the robustness of DRL agents. However, existing attack methods targeting individual sampled actions have limited impacts on the overall policy distribution, particularly in continuous action spaces. To address these limitations, we propose the Distribution-Aware Projected Gradient Descent attack (DAPGD). DAPGD uses distribution similarity as the gradient perturbation input to attack the policy network, which leverages the entire policy distribution rather than relying on individual samples. We utilize the Bhattacharyya distance in DAPGD to measure policy similarity, enabling sensitive detection of subtle but critical differences between probability distributions. Our experiment results demonstrate that DAPGD achieves SOTA results compared to the baselines in three robot navigation tasks, achieving an average 22.03% higher reward drop compared to the best baseline.

arXiv:2501.02205

Digital Twin Calibration with Model-Based Reinforcement Learning

This paper presents a novel methodological framework, called the Actor-Simulator, that incorporates the calibration of digital twins into model-based reinforcement learning for more effective control of stochastic systems with complex nonlinear dynamics. Traditional model-based control often relies on restrictive structural assumptions (such as linear state transitions) and fails to account for parameter uncertainty in the model. These issues become particularly critical in industries such as biopharmaceutical manufacturing, where process dynamics are complex and not fully known, and only a limited amount of data is available. Our approach jointly calibrates the digital twin and searches for an optimal control policy, thus accounting for and reducing model error. We balance exploration and exploitation by using policy performance as a guide for data collection. This dual-component approach provably converges to the optimal policy, and outperforms existing methods in extensive numerical experiments based on the biopharmaceutical manufacturing domain.

arXiv:2501.02709

Horizon Generalization in Reinforcement Learning

We study goal-conditioned RL through the lens of generalization, but not in the traditional sense of random augmentations and domain randomization. Rather, we aim to learn goal-directed policies that generalize with respect to the horizon: after training to reach nearby goals (which are easy to learn), these policies should succeed in reaching distant goals (which are quite challenging to learn). In the same way that invariance is closely linked with generalization is other areas of machine learning (e.g., normalization layers make a network invariant to scale, and therefore generalize to inputs of varying scales), we show that this notion of horizon generalization is closely linked with invariance to planning: a policy navigating towards a goal will select the same actions as if it were navigating to a waypoint en route to that goal. Thus, such a policy trained to reach nearby goals should succeed at reaching arbitrarily-distant goals. Our theoretical analysis proves that both horizon generalization and planning invariance are possible, under some assumptions. We present new experimental results and recall findings from prior work in support of our theoretical results. Taken together, our results open the door to studying how techniques for invariance and generalization developed in other areas of machine learning might be adapted to achieve this alluring property.

arXiv:2501.02481공식 코드 · Papers with Code

Representation Convergence: Mutual Distillation is Secretly a Form of Regularization

In this paper, we argue that mutual distillation between reinforcement learning policies serves as an implicit regularization, preventing them from overfitting to irrelevant features. We highlight two key contributions: (a) Theoretically, for the first time, we prove that enhancing the policy robustness to irrelevant features leads to improved generalization performance. (b) Empirically, we demonstrate that mutual distillation between policies contributes to such robustness, enabling the spontaneous emergence of invariant representations over pixel inputs. Overall, our findings challenge the conventional view of distillation as merely a means of knowledge transfer, offering a novel perspective on the generalization in deep reinforcement learning.