In this work we propose a planning and acting architecture endowed with a module which learns to select subgoals with Deep Q-Learning. This allows us to decrease the load of a planner when faced with scenarios with real-time restrictions. We have trained this architecture on a video game environment used as a standard test-bed for intelligent systems applications, testing it on different levels of the same game to evaluate its generalization abilities. We have measured the performance of our approach as more training data is made available, as well as compared it with both a state-of-the-art, classical planner and the standard Deep Q-Learning algorithm. The results obtained show our model performs better than the alternative methods considered, when both plan quality (plan length) and time requirements are taken into account. On the one hand, it is more sample-efficient than standard Deep Q-Learning, and it is able to generalize better across levels. On the other hand, it reduces problem-solving time when compared with a state-of-the-art automated planner, at the expense of obtaining plans with only 9% more actions.
Sample efficiency is critical when applying learning-based methods to robotic manipulation due to the high cost of collecting expert demonstrations and the challenges of on-robot policy learning through online Reinforcement Learning (RL). Offline RL addresses this issue by enabling policy learning from an offline dataset collected using any behavioral policy, regardless of its quality. However, recent advancements in offline RL have predominantly focused on learning from large datasets. Given that many robotic manipulation tasks can be formulated as rotation-symmetric problems, we investigate the use of $SO(2)$-equivariant neural networks for offline RL with a limited number of demonstrations. Our experimental results show that equivariant versions of Conservative Q-Learning (CQL) and Implicit Q-Learning (IQL) outperform their non-equivariant counterparts. We provide empirical evidence demonstrating how equivariance improves offline learning algorithms in the low-data regime.
There has been significant interest in the development of personalized and adaptive educational tools that cater to a student's individual learning progress. A crucial aspect in developing such tools is in exploring how mastery can be achieved across a diverse yet related range of content in an efficient manner. While Reinforcement Learning and Multi-armed Bandits have shown promise in educational settings, existing works often assume the independence of learning content, neglecting the prevalent interdependencies between such content. In response, we introduce Education Network Restless Multi-armed Bandits (EdNetRMABs), utilizing a network to represent the relationships between interdependent arms. Subsequently, we propose EduQate, a method employing interdependency-aware Q-learning to make informed decisions on arm selection at each time step. We establish the optimality guarantee of EduQate and demonstrate its efficacy compared to baseline policies, using students modeled from both synthetic and real-world data.
The remarkable adaptability of humans in response to complex environments is often demonstrated by the context-dependent adoption of different behavioral modes. However, the existing game-theoretic studies mostly focus on the single-mode assumption, and the impact of this behavioral multimodality on the evolution of cooperation remains largely unknown. Here, we study how cooperation evolves in a population with two behavioral modes. Specifically, we incorporate Q-learning and Tit-for-Tat (TFT) rules into our toy model, where prisoner's dilemma game is played and we investigate the impact of the mode mixture on the evolution of cooperation. While players in Q-learning mode aim to maximize their accumulated payoffs, players within TFT mode repeat what their neighbors have done to them. In a structured mixing implementation where the updating rule is fixed for each individual, we find that the mode mixture greatly promotes the overall cooperation prevalence. The promotion is even more significant in the probabilistic mixing, where players randomly select one of the two rules at each step. Finally, this promotion is robust when players are allowed to adaptively choose the two modes by real-time comparison. In all three scenarios, players within the Q-learning mode act as catalyzer that turns the TFT players to be more cooperative, and as a result drive the whole population to be highly cooperative. The analysis of Q-tables explains the underlying mechanism of cooperation promotion, which captures the ``psychologic evolution" in the players' mind. Our study indicates that the variety of behavioral modes is non-negligible, and could be crucial to clarify the emergence of cooperation in the real world.
This article provides a methodology and open-source implementation of Reinforcement Learning algorithms for finding optimal routes in a packet-optical network scenario. The algorithm uses measurements provided by the physical layer (pre-FEC bit error rate and propagation delay) and the link layer (link load) to configure a set of latency-based rewards and penalties based on such measurements. Then, the algorithm executes Q-learning based on this set of rewards for finding the optimal routing strategies. It is further shown that the algorithm dynamically adapts to changing network conditions by re-calculating optimal policies upon either link load changes or link degradation as measured by pre-FEC BER.
The primary goal of reinforcement learning is to develop decision-making policies that prioritize optimal performance without considering risk or safety. In contrast, safe reinforcement learning aims to mitigate or avoid unsafe states. This paper presents a risk-sensitive Q-learning algorithm that leverages optimal transport theory to enhance the agent safety. By integrating optimal transport into the Q-learning framework, our approach seeks to optimize the policy's expected return while minimizing the Wasserstein distance between the policy's stationary distribution and a predefined risk distribution, which encapsulates safety preferences from domain experts. We validate the proposed algorithm in a Gridworld environment. The results indicate that our method significantly reduces the frequency of visits to risky states and achieves faster convergence to a stable policy compared to the traditional Q-learning algorithm.
Lane-changing decisions, which are crucial for autonomous vehicle path planning, face practical challenges due to rule-based constraints and limited data. Deep reinforcement learning has become a major research focus due to its advantages in data acquisition and interpretability. However, current models often overlook collaboration, which affects not only impacts overall traffic efficiency but also hinders the vehicle's own normal driving in the long run. To address the aforementioned issue, this paper proposes a method named Mix Q-learning for Lane Changing(MQLC) that integrates a hybrid value Q network, taking into account both collective and individual benefits for the greater good. At the collective level, our method coordinates the individual Q and global Q networks by utilizing global information. This enables agents to effectively balance their individual interests with the collective benefit. At the individual level, we integrated a deep learning-based intent recognition module into our observation and enhanced the decision network. These changes provide agents with richer decision information and more accurate feature extraction for improved lane-changing decisions. This strategy enables the multi-agent system to learn and formulate optimal decision-making strategies effectively. Our MQLC model, through extensive experimental results, impressively outperforms other state-of-the-art multi-agent decision-making methods, achieving significantly safer and faster lane-changing decisions.
$Q$-learning is one of the most fundamental reinforcement learning (RL) algorithms. Despite its widespread success in various applications, it is prone to overestimation bias in the $Q$-learning update. To address this issue, double $Q$-learning employs two independent $Q$-estimators which are randomly selected and updated during the learning process. This paper proposes a modified double $Q$-learning, called simultaneous double $Q$-learning (SDQ), with its finite-time analysis. SDQ eliminates the need for random selection between the two $Q$-estimators, and this modification allows us to analyze double $Q$-learning through the lens of a novel switching system framework facilitating efficient finite-time analysis. Empirical studies demonstrate that SDQ converges faster than double $Q$-learning while retaining the ability to mitigate the maximization bias. Finally, we derive a finite-time expected error bound for SDQ.
Semi-gradient Q-learning is applied in many fields, but due to the absence of an explicit loss function, studying its dynamics and implicit bias in the parameter space is challenging. This paper introduces the Fokker--Planck equation and employs partial data obtained through sampling to construct and visualize the effective loss landscape within a two-dimensional parameter space. This visualization reveals how the global minima in the loss landscape can transform into saddle points in the effective loss landscape, as well as the implicit bias of the semi-gradient method. Additionally, we demonstrate that saddle points, originating from the global minima in loss landscape, still exist in the effective loss landscape under high-dimensional parameter spaces and neural network settings. This paper develop a novel approach for probing implicit bias in semi-gradient Q-learning.
Multi-agent reinforcement learning (MARL) has become a significant research topic due to its ability to facilitate learning in complex environments. In multi-agent tasks, the state-action value, commonly referred to as the Q-value, can vary among agents because of their individual rewards, resulting in a Q-vector. Determining an optimal policy is challenging, as it involves more than just maximizing a single Q-value. Various optimal policies, such as a Nash equilibrium, have been studied in this context. Algorithms like Nash Q-learning and Nash Actor-Critic have shown effectiveness in these scenarios. This paper extends this research by proposing a deep Q-networks (DQN) algorithm capable of learning various Q-vectors using Max, Nash, and Maximin strategies. The effectiveness of this approach is demonstrated in an environment where dual robotic arms collaborate to lift a pot.
Despite the recent advancements in offline RL, no unified algorithm could achieve superior performance across a broad range of tasks. Offline \textit{value function learning}, in particular, struggles with sparse-reward, long-horizon tasks due to the difficulty of solving credit assignment and extrapolation errors that accumulates as the horizon of the task grows.~On the other hand, models that can perform well in long-horizon tasks are designed specifically for goal-conditioned tasks, which commonly perform worse than value function learning methods on short-horizon, dense-reward scenarios. To bridge this gap, we propose a hierarchical planner designed for offline RL called PlanDQ. PlanDQ incorporates a diffusion-based planner at the high level, named D-Conductor, which guides the low-level policy through sub-goals. At the low level, we used a Q-learning based approach called the Q-Performer to accomplish these sub-goals. Our experimental results suggest that PlanDQ can achieve superior or competitive performance on D4RL continuous control benchmark tasks as well as AntMaze, Kitchen, and Calvin as long-horizon tasks.
Offline reinforcement learning (RL) is a compelling paradigm to extend RL's practical utility by leveraging pre-collected, static datasets, thereby avoiding the limitations associated with collecting online interactions. The major difficulty in offline RL is mitigating the impact of approximation errors when encountering out-of-distribution (OOD) actions; doing so ineffectively will lead to policies that prefer OOD actions, which can lead to unexpected and potentially catastrophic results. Despite the variety of works proposed to address this issue, they tend to excessively suppress the value function in and around OOD regions, resulting in overly pessimistic value estimates. In this paper, we propose a novel framework called Strategically Conservative Q-Learning (SCQ) that distinguishes between OOD data that is easy and hard to estimate, ultimately resulting in less conservative value estimates. Our approach exploits the inherent strengths of neural networks to interpolate, while carefully navigating their limitations in extrapolation, to obtain pessimistic yet still property calibrated value estimates. Theoretical analysis also shows that the value function learned by SCQ is still conservative, but potentially much less so than that of Conservative Q-learning (CQL). Finally, extensive evaluation on the D4RL benchmark tasks shows our proposed method outperforms state-of-the-art methods. Our code is available through \url{https://github.com/purewater0901/SCQ}.
Radio Resource Management is a challenging topic in future 6G networks where novel applications create strong competition among the users for the available resources. In this work we consider the frequency scheduling problem in a multi-user MIMO system. Frequency resources need to be assigned to a set of users while allowing for concurrent transmissions in the same sub-band. Traditional methods are insufficient to cope with all the involved constraints and uncertainties, whereas reinforcement learning can directly learn near-optimal solutions for such complex environments. However, the scheduling problem has an enormous action space accounting for all the combinations of users and sub-bands, so out-of-the-box algorithms cannot be used directly. In this work, we propose a scheduler based on action-branching over sub-bands, which is a deep Q-learning architecture with parallel decision capabilities. The sub-bands learn correlated but local decision policies and altogether they optimize a global reward. To improve the scaling of the architecture with the number of sub-bands, we propose variations (Unibranch, Graph Neural Network-based) that reduce the number of parameters to learn. The parallel decision making of the proposed architecture allows to meet short inference time requirements in real systems. Furthermore, the deep Q-learning approach permits online fine-tuning after deployment to bridge the sim-to-real gap. The proposed architectures are evaluated against relevant baselines from the literature showing competitive performance and possibilities of online adaptation to evolving environments.
In offline reinforcement learning, in-sample learning methods have been widely used to prevent performance degradation caused by evaluating out-of-distribution actions from the dataset. Extreme Q-learning (XQL) employs a loss function based on the assumption that Bellman error follows a Gumbel distribution, enabling it to model the soft optimal value function in an in-sample manner. It has demonstrated strong performance in both offline and online reinforcement learning settings. However, issues remain, such as the instability caused by the exponential term in the loss function and the risk of the error distribution deviating from the Gumbel distribution. Therefore, we propose Maclaurin Expanded Extreme Q-learning to enhance stability. In this method, applying Maclaurin expansion to the loss function in XQL enhances stability against large errors. This approach involves adjusting the modeled value function between the value function under the behavior policy and the soft optimal value function, thus achieving a trade-off between stability and optimality depending on the order of expansion. It also enables adjustment of the error distribution assumption from a normal distribution to a Gumbel distribution. Our method significantly stabilizes learning in online RL tasks from DM Control, where XQL was previously unstable. Additionally, it improves performance in several offline RL tasks from D4RL.
Many classic Reinforcement Learning (RL) algorithms rely on a Bellman operator, which involves an expectation over the next states, leading to the concept of bootstrapping. To introduce a form of pessimism, we propose to replace this expectation with an expectile. In practice, this can be very simply done by replacing the $L_2$ loss with a more general expectile loss for the critic. Introducing pessimism in RL is desirable for various reasons, such as tackling the overestimation problem (for which classic solutions are double Q-learning or the twin-critic approach of TD3) or robust RL (where transitions are adversarial). We study empirically these two cases. For the overestimation problem, we show that the proposed approach, ExpectRL, provides better results than a classic twin-critic. On robust RL benchmarks, involving changes of the environment, we show that our approach is more robust than classic RL algorithms. We also introduce a variation of ExpectRL combined with domain randomization which is competitive with state-of-the-art robust RL agents. Eventually, we also extend \ExpectRL with a mechanism for choosing automatically the expectile value, that is the degree of pessimism
Nowadays, a significant share of the Business-to-Consumer sector is based on online platforms like Amazon and Alibaba and uses Artificial Intelligence for pricing strategies. This has sparked debate on whether pricing algorithms may tacitly collude to set supra-competitive prices without being explicitly designed to do so. Our study addresses these concerns by examining the risk of collusion when Reinforcement Learning algorithms are used to decide on pricing strategies in competitive markets. Prior research in this field focused on Tabular Q-learning (TQL) and led to opposing views on whether learning-based algorithms can lead to supra-competitive prices. Our work contributes to this ongoing discussion by providing a more nuanced numerical study that goes beyond TQL by additionally capturing off- and on-policy Deep Reinforcement Learning (DRL) algorithms. We study multiple Bertrand oligopoly variants and show that algorithmic collusion depends on the algorithm used. In our experiments, TQL exhibits higher collusion and price dispersion phenomena compared to DRL algorithms. We show that the severity of collusion depends not only on the algorithm used but also on the characteristics of the market environment. We further find that Proximal Policy Optimization appears to be less sensitive to collusive outcomes compared to other state-of-the-art DRL algorithms.
Zero Trust is a new security vision for 6G networks that emphasises the philosophy of never trust and always verify. However, there is a fundamental trade-off between the wireless transmission efficiency and the trust level, which is reflected by the verification interval and its adaptation strategy. More importantly, the mathematical framework to characterise the trust level of the adaptive verification strategy is still missing. Inspired by this vision, we propose a concept called age of trust (AoT) to capture the characteristics of the trust level degrading over time, with the definition of the time elapsed since the last verification of the target user's trust plus the initial age, which depends on the trust level evaluated at that verification. The higher the trust level, the lower the initial age. To evaluate the trust level in the long term, the average AoT is used. We then investigate how to find a compromise between average AoT and wireless transmission efficiency with limited resources. In particular, we address the bi-objective optimization (BOO) problem between average AoT and throughput over a single link with arbitrary service process, where the identity of the receiver is constantly verified, and we devise a periodic verification scheme and a Q-learning-based scheme for constant process and random process, respectively. We also tackle the BOO problem in a multiple random access scenario, where a trust-enhanced frame-slotted ALOHA is designed. Finally, the numerical results show that our proposals can achieve a fair compromise between trust level and wireless transmission efficiency, and thus have a wide application prospect in various zero-trust architectures.
The rapid adoption of Large Language Models (LLMs) has exposed critical security and ethical vulnerabilities, particularly their susceptibility to adversarial manipulations. This paper introduces QROA, a novel black-box jailbreak method designed to identify adversarial suffixes that can bypass LLM alignment safeguards when appended to a malicious instruction. Unlike existing suffix-based jailbreak approaches, QROA does not require access to the model's logit or any other internal information. It also eliminates reliance on human-crafted templates, operating solely through the standard query-response interface of LLMs. By framing the attack as an optimization bandit problem, QROA employs a surrogate model and token level optimization to efficiently explore suffix variations. Furthermore, we propose QROA-UNV, an extension that identifies universal adversarial suffixes for individual models, enabling one-query jailbreaks across a wide range of instructions. Testing on multiple models demonstrates Attack Success Rate (ASR) greater than 80\%. These findings highlight critical vulnerabilities, emphasize the need for advanced defenses, and contribute to the development of more robust safety evaluations for secure AI deployment. The code is made public on the following link: https://github.com/qroa/QROA
The Whittle index policy is a heuristic that has shown remarkably good performance (with guaranteed asymptotic optimality) when applied to the class of problems known as Restless Multi-Armed Bandit Problems (RMABPs). In this paper we present QWI and QWINN, two reinforcement learning algorithms, respectively tabular and deep, to learn the Whittle index for the total discounted criterion. The key feature is the use of two time-scales, a faster one to update the state-action Q -values, and a relatively slower one to update the Whittle indices. In our main theoretical result we show that QWI, which is a tabular implementation, converges to the real Whittle indices. We then present QWINN, an adaptation of QWI algorithm using neural networks to compute the Q -values on the faster time-scale, which is able to extrapolate information from one state to another and scales naturally to large state-space environments. For QWINN, we show that all local minima of the Bellman error are locally stable equilibria, which is the first result of its kind for DQN-based schemes. Numerical computations show that QWI and QWINN converge faster than the standard Q -learning algorithm, neural-network based approximate Q-learning and other state of the art algorithms.
Q-learning is widely recognized as an effective approach for synthesizing controllers to achieve specific goals. However, handling challenges posed by continuous state-action spaces remains an ongoing research focus. This paper presents a systematic analysis that highlights a major drawback in space discretization methods. To address this challenge, the paper proposes a symbolic model that represents behavioral relations, such as alternating simulation from abstraction to the controlled system. This relation allows for seamless application of the synthesized controller based on abstraction to the original system. Introducing a novel Q-learning technique for symbolic models, the algorithm yields two Q-tables encoding optimal policies. Theoretical analysis demonstrates that these Q-tables serve as both upper and lower bounds on the Q-values of the original system with continuous spaces. Additionally, the paper explores the correlation between the parameters of the space abstraction and the loss in Q-values. The resulting algorithm facilitates achieving optimality within an arbitrary accuracy, providing control over the trade-off between accuracy and computational complexity. The obtained results provide valuable insights for selecting appropriate learning parameters and refining the controller. The engineering relevance of the proposed Q-learning based symbolic model is illustrated through two case studies.