Underpinning the past decades of work on the design, initialization, and optimization of neural networks is a seemingly innocuous assumption: that the network is trained on a \textit{stationary} data distribution. In settings where this assumption is violated, e.g.\ deep reinforcement learning, learning algorithms become unstable and brittle with respect to hyperparameters and even random seeds. One factor driving this instability is the loss of plasticity, meaning that updating the network's predictions in response to new information becomes more difficult as training progresses. While many recent works provide analyses and partial solutions to this phenomenon, a fundamental question remains unanswered: to what extent do known mechanisms of plasticity loss overlap, and how can mitigation strategies be combined to best maintain the trainability of a network? This paper addresses these questions, showing that loss of plasticity can be decomposed into multiple independent mechanisms and that, while intervening on any single mechanism is insufficient to avoid the loss of plasticity in all cases, intervening on multiple mechanisms in conjunction results in highly robust learning algorithms. We show that a combination of layer normalization and weight decay is highly effective at maintaining plasticity in a variety of synthetic nonstationary learning tasks, and further demonstrate its effectiveness on naturally arising nonstationarities, including reinforcement learning in the Arcade Learning Environment.
The trend of massive connectivity pushes forward the explosive growth of end devices. The emergence of various applications has prompted a demand for pervasive connectivity and more efficient computing paradigms. On the other hand, the lack of computational capacity of the end devices restricts the implementation of the intelligent applications, and becomes a bottleneck of the multiple access for supporting massive connectivity. Mobile cloud computing (MCC) and mobile edge computing (MEC) techniques enable end devices to offload local computation-intensive tasks to servers by networks. In this paper, we consider the cloud-edge-end collaborative networks to utilize distributed computing resources. Furthermore, we apply task-oriented semantic communications to tackle the fast-varying channel between the end devices and MEC servers and reduce the communication cost. To minimize long-term energy consumption on constraints queue stability and computational delay, a Lyapunov-guided deep reinforcement learning hybrid (DRLH) framework is proposed to solve the mixed integer non-linear programming (MINLP) problem. The long-term energy consumption minimization problem is transformed into the deterministic problem in each time frame. The DRLH framework integrates a model-free deep reinforcement learning algorithm with a model-based mathematical optimization algorithm to mitigate computational complexity and leverage the scenario information, so that improving the convergence performance. Numerical results demonstrate that the proposed DRLH framework achieves near-optimal performance on energy consumption while stabilizing all queues.
We consider the problem of dynamic channel allocation (DCA) in cognitive communication networks with the goal of maximizing a global signal-to-interference-plus-noise ratio (SINR) measure under a specified target quality of service (QoS)-SINR for each network. The shared bandwidth is partitioned into K channels with frequency separation. In contrast to the majority of existing studies that assume perfect orthogonality or a one- to-one user-channel allocation mapping, this paper focuses on real-world systems experiencing inter-carrier interference (ICI) and channel reuse by multiple large-scale networks. This realistic scenario significantly increases the problem dimension, rendering existing algorithms inefficient. We propose a novel multi-agent reinforcement learning (RL) framework for distributed DCA, named Channel Allocation RL To Overlapped Networks (CARLTON). The CARLTON framework is based on the Centralized Training with Decentralized Execution (CTDE) paradigm, utilizing the DeepMellow value-based RL algorithm. To ensure robust performance in the interference-laden environment we address, CARLTON employs a low-dimensional representation of observations, generating a QoS-type measure while maximizing a global SINR measure and ensuring the target QoS-SINR for each network. Our results demonstrate exceptional performance and robust generalization, showcasing superior efficiency compared to alternative state-of-the-art methods, while achieving a marginally diminished performance relative to a fully centralized approach.
Diverse studies in systems neuroscience begin with extended periods of curriculum training known as `shaping' procedures. These involve progressively studying component parts of more complex tasks, and can make the difference between learning a task quickly, slowly or not at all. Despite the importance of shaping to the acquisition of complex tasks, there is as yet no theory that can help guide the design of shaping procedures, or more fundamentally, provide insight into its key role in learning. Modern deep reinforcement learning systems might implicitly learn compositional primitives within their multilayer policy networks. Inspired by these models, we propose and analyse a model of deep policy gradient learning of simple compositional reinforcement learning tasks. Using the tools of statistical physics, we solve for exact learning dynamics and characterise different learning strategies including primitives pre-training, in which task primitives are studied individually before learning compositional tasks. We find a complex interplay between task complexity and the efficacy of shaping strategies. Overall, our theory provides an analytical understanding of the benefits of shaping in a class of compositional tasks and a quantitative account of how training protocols can disclose useful task primitives, ultimately yielding faster and more robust learning.
This study tackles the representative yet challenging contact-rich peg-in-hole task of robotic assembly, using a soft wrist that can operate more safely and tolerate lower-frequency control signals than a rigid one. Previous studies often use a fully observable formulation, requiring external setups or estimators for the peg-to-hole pose. In contrast, we use a partially observable formulation and deep reinforcement learning from demonstrations to learn a memory-based agent that acts purely on haptic and proprioceptive signals. Moreover, previous works do not incorporate potential domain symmetry and thus must search for solutions in a bigger space. Instead, we propose to leverage the symmetry for sample efficiency by augmenting the training data and constructing auxiliary losses to force the agent to adhere to the symmetry. Results in simulation with five different symmetric peg shapes show that our proposed agent can be comparable to or even outperform a state-based agent. In particular, the sample efficiency also allows us to learn directly on the real robot within 3 hours.
As a model-free algorithm, deep reinforcement learning (DRL) agent learns and makes decisions by interacting with the environment in an unsupervised way. In recent years, DRL algorithms have been widely applied by scholars for portfolio optimization in consecutive trading periods, since the DRL agent can dynamically adapt to market changes and does not rely on the specification of the joint dynamics across the assets. However, typical DRL agents for portfolio optimization cannot learn a policy that is aware of the dynamic correlation between portfolio asset returns. Since the dynamic correlations among portfolio assets are crucial in optimizing the portfolio, the lack of such knowledge makes it difficult for the DRL agent to maximize the return per unit of risk, especially when the target market permits short selling (i.e., the US stock market). In this research, we propose a hybrid portfolio optimization model combining the DRL agent and the Black-Litterman (BL) model to enable the DRL agent to learn the dynamic correlation between the portfolio asset returns and implement an efficacious long/short strategy based on the correlation. Essentially, the DRL agent is trained to learn the policy to apply the BL model to determine the target portfolio weights. To test our DRL agent, we construct the portfolio based on all the Dow Jones Industrial Average constitute stocks. Empirical results of the experiments conducted on real-world United States stock market data demonstrate that our DRL agent significantly outperforms various comparison portfolio choice strategies and alternative DRL frameworks by at least 42% in terms of accumulated return. In terms of the return per unit of risk, our DRL agent significantly outperforms various comparative portfolio choice strategies and alternative strategies based on other machine learning frameworks.
In this paper, we propose an adaptive panoramic video semantic transmission (APVST) framework enabled by rate splitting multiple access (RSMA). The APVST framework consists of a semantic transmitter and receiver, utilizing a deep joint source-channel coding structure to adaptively extract and encode semantic features from panoramic frames. To achieve higher spectral efficiency and conserve bandwidth, APVST employs an entropy model and a dimension-adaptive module to control the transmission rate. Additionally, we take weighted-to-spherically-uniform peak signal-to-noise ratio (WS-PSNR) and weighted-to-spherically-uniform structural similarity (WS-SSIM) as distortion evaluation metrics for panoramic videos and design a weighted self-attention module for APVST. This module integrates weights and feature maps to enhance the quality of the immersive experience. Considering the overlap in the field of view when users watch panoramic videos, we further utilize RSMA to split the required panoramic video semantic streams into common and private messages for transmission. We propose an RSMA-enabled semantic stream transmission scheme and formulate a joint problem of latency and immersive experience quality by optimizing the allocation ratios of power, common rate, and channel bandwidth, aiming to maximize the quality of service (QoS) scores for users. To address the above problem, we propose a deep reinforcement learning algorithm based on proximal policy optimization (PPO) with high efficiency to handle dynamically changing environments. Simulation results demonstrate that our proposed APVST framework saves up to 20% and 50% of channel bandwidth compared to other semantic and traditional video transmission schemes, respectively. Moreover, our study confirms the efficiency of RSMA in panoramic video transmission, achieving performance gains of 13% and 20% compared to NOMA and OFDMA.
Low Earth Orbit (LEO) satellite-to-handheld connections herald a new era in satellite communications. Space-Division Multiple Access (SDMA) precoding is a method that mitigates interference among satellite beams, boosting spectral efficiency. While optimal SDMA precoding solutions have been proposed for ideal channel knowledge in various scenarios, addressing robust precoding with imperfect channel information has primarily been limited to simplified models. However, these models might not capture the complexity of LEO satellite applications. We use the Soft Actor-Critic (SAC) deep Reinforcement Learning (RL) method to learn robust precoding strategies without the need for explicit insights into the system conditions and imperfections. Our results show flexibility to adapt to arbitrary system configurations while performing strongly in terms of achievable rate and robustness to disruptive influences compared to analytical benchmark precoders.
Deep Reinforcement Learning (DRL) techniques have received significant attention in control and decision-making algorithms. Most applications involve complex decision-making systems, justified by the algorithms' computational power and cost. While model-based versions are emerging, model-free DRL approaches are intriguing for their independence from models, yet they remain relatively less explored in terms of performance, particularly in applied control. This study conducts a thorough performance analysis comparing the data-driven DRL paradigm with a classical state feedback controller, both designed based on the same cost (reward) function of the linear quadratic regulator (LQR) problem. Twelve additional performance criteria are introduced to assess the controllers' performance, independent of the LQR problem for which they are designed. Two Deep Deterministic Policy Gradient (DDPG)-based controllers are developed, leveraging DDPG's widespread reputation. These controllers are aimed at addressing a challenging setpoint tracking problem in a Non-Minimum Phase (NMP) system. The performance and robustness of the controllers are assessed in the presence of operational challenges, including disturbance, noise, initial conditions, and model uncertainties. The findings suggest that the DDPG controller demonstrates promising behavior under rigorous test conditions. Nevertheless, further improvements are necessary for the DDPG controller to outperform classical methods in all criteria. While DRL algorithms may excel in complex environments owing to the flexibility in the reward function definition, this paper offers practical insights and a comparison framework specifically designed to evaluate these algorithms within the context of control engineering.
Recently, the increasing use of deep reinforcement learning for flow control problems has led to a new area of research, focused on the coupling and the adaptation of the existing algorithms to the control of numerical fluid dynamics environments. Although still in its infancy, the field has seen multiple successes in a short time span, and its fast development pace can certainly be partly imparted to the open-source effort that drives the expansion of the community. Yet, this emerging domain still misses a common ground to (i) ensure the reproducibility of the results, and (ii) offer a proper ad-hoc benchmarking basis. To this end, we propose Beacon, an open-source benchmark library composed of seven lightweight 1D and 2D flow control problems with various characteristics, action and observation space characteristics, and CPU requirements. In this contribution, the seven considered problems are described, and reference control solutions are provided. The sources for the following work are available at https://github.com/jviquerat/beacon.
This paper introduces a Multi-Agent Deep Reinforcement Learning (MA-DRL) approach for routing in Low Earth Orbit Satellite Constellations (LSatCs). Each satellite is an independent decision-making agent with a partial knowledge of the environment, and supported by feedback received from the nearby agents. Building on our previous work that introduced a Q-routing solution, the contribution of this paper is to extend it to a deep learning framework able to quickly adapt to the network and traffic changes, and based on two phases: (1) An offline exploration learning phase that relies on a global Deep Neural Network (DNN) to learn the optimal paths at each possible position and congestion level; (2) An online exploitation phase with local, on-board, pre-trained DNNs. Results show that MA-DRL efficiently learns optimal routes offline that are then loaded for an efficient distributed routing online.
Deep reinforcement learning excels in numerous large-scale practical applications. However, existing performance analyses ignores the unique characteristics of continuous-time control problems, is unable to directly estimate the generalization error of the Bellman optimal loss and require a boundedness assumption. Our work focuses on continuous-time control problems and proposes a method that is applicable to all such problems where the transition function satisfies semi-group and Lipschitz properties. Under this method, we can directly analyze the \emph{a priori} generalization error of the Bellman optimal loss. The core of this method lies in two transformations of the loss function. To complete the transformation, we propose a decomposition method for the maximum operator. Additionally, this analysis method does not require a boundedness assumption. Finally, we obtain an \emph{a priori} generalization error without the curse of dimensionality.
Many rescue missions demand effective perception and real-time decision making, which highly rely on effective data collection and processing. In this study, we propose a three-layer architecture of emergency caching networks focusing on data collection and reliable transmission, by leveraging efficient perception and edge caching technologies. Based on this architecture, we propose a disaster map collection framework that integrates coded caching technologies. Our framework strategically caches coded fragments of maps across unmanned aerial vehicles (UAVs), fostering collaborative uploading for augmented transmission reliability. Additionally, we establish a comprehensive probability model to assess the effective recovery area of disaster maps. Towards the goal of utility maximization, we propose a deep reinforcement learning (DRL) based algorithm that jointly makes decisions about cooperative UAVs selection, bandwidth allocation and coded caching parameter adjustment, accommodating the real-time map updates in a dynamic disaster situation. Our proposed scheme is more effective than the non-coding caching scheme, as validated by simulation.
In recent years, deep reinforcement learning has emerged as a technique to solve closed-loop flow control problems. Employing simulation-based environments in reinforcement learning enables a priori end-to-end optimization of the control system, provides a virtual testbed for safety-critical control applications, and allows to gain a deep understanding of the control mechanisms. While reinforcement learning has been applied successfully in a number of rather simple flow control benchmarks, a major bottleneck toward real-world applications is the high computational cost and turnaround time of flow simulations. In this contribution, we demonstrate the benefits of model-based reinforcement learning for flow control applications. Specifically, we optimize the policy by alternating between trajectories sampled from flow simulations and trajectories sampled from an ensemble of environment models. The model-based learning reduces the overall training time by up to $85\%$ for the fluidic pinball test case. Even larger savings are expected for more demanding flow simulations.
In the rapidly evolving domain of electrical power systems, the Volt-VAR optimization (VVO) is increasingly critical, especially with the burgeoning integration of renewable energy sources. Traditional approaches to learning-based VVO in expansive and dynamically changing power systems are often hindered by computational complexities. To address this challenge, our research presents a novel framework that harnesses the potential of Deep Reinforcement Learning (DRL), specifically utilizing the Importance Weighted Actor-Learner Architecture (IMPALA) algorithm, executed on the RAY platform. This framework, built upon RLlib-an industry-standard in Reinforcement Learning-ingeniously capitalizes on the distributed computing capabilities and advanced hyperparameter tuning offered by RAY. This design significantly expedites the exploration and exploitation phases in the VVO solution space. Our empirical results demonstrate that our approach not only surpasses existing DRL methods in achieving superior reward outcomes but also manifests a remarkable tenfold reduction in computational requirements. The integration of our DRL agent with the RAY platform facilitates the creation of RLlib-IMPALA, a novel framework that efficiently uses RAY's resources to improve system adaptability and control. RLlib-IMPALA leverages RAY's toolkit to enhance analytical capabilities and significantly speeds up training to become more than 10 times faster than other state-of-the-art DRL methods.
This paper tackles the growing issue of excessive data transmission in networks. With increasing traffic, backhaul links and core networks are under significant traffic, leading to the investigation of caching solutions at edge routers. Many existing studies utilize Markov Decision Processes (MDP) to tackle caching problems, often assuming decision points at fixed intervals; however, real-world environments are characterized by random request arrivals. Additionally, critical file attributes such as lifetime, size, and priority significantly impact the effectiveness of caching policies, yet existing research fails to integrate all these attributes in policy design. In this work, we model the caching problem using a Semi-Markov Decision Process (SMDP) to better capture the continuous-time nature of real-world applications, enabling caching decisions to be triggered by random file requests. We then introduce a Proximal Policy Optimization (PPO)--based caching strategy that fully considers file attributes like lifetime, size, and priority. Simulations show that our method outperforms a recent Deep Reinforcement Learning-based technique. To further advance our research, we improved the convergence rate of PPO by prioritizing transitions within the replay buffer through an attention mechanism. This mechanism evaluates the similarity between the current state and all stored transitions, assigning higher priorities to transitions that exhibit greater similarity.
A key challenge in realizing fault-tolerant quantum computers is circuit optimization. Focusing on the most expensive gates in fault-tolerant quantum computation (namely, the T gates), we address the problem of T-count optimization, i.e., minimizing the number of T gates that are needed to implement a given circuit. To achieve this, we develop AlphaTensor-Quantum, a method based on deep reinforcement learning that exploits the relationship between optimizing T-count and tensor decomposition. Unlike existing methods for T-count optimization, AlphaTensor-Quantum can incorporate domain-specific knowledge about quantum computation and leverage gadgets, which significantly reduces the T-count of the optimized circuits. AlphaTensor-Quantum outperforms the existing methods for T-count optimization on a set of arithmetic benchmarks (even when compared without making use of gadgets). Remarkably, it discovers an efficient algorithm akin to Karatsuba's method for multiplication in finite fields. AlphaTensor-Quantum also finds the best human-designed solutions for relevant arithmetic computations used in Shor's algorithm and for quantum chemistry simulation, thus demonstrating it can save hundreds of hours of research by optimizing relevant quantum circuits in a fully automated way.
This work focuses on using advanced techniques for structural health monitoring (SHM) for bridges with Traffic. We propose an approach using deep reinforcement learning (DRL)-based control for Unmanned Aerial Vehicle (UAV). Our approach conducts a concrete bridge deck survey while traffic is ongoing and detects cracks. The UAV performs the crack detection, and the location of cracks is initially unknown. We use two edge detection techniques. First, we use canny edge detection for crack detection. We also use a Convolutional Neural Network (CNN) for crack detection and compare it with canny edge detection. Transfer learning is applied using CNN with pre-trained weights obtained from a crack image dataset. This enables the model to adapt and improve its performance in identifying and localizing cracks. Proximal Policy Optimization (PPO) is applied for UAV control and bridge surveys. The experimentation across various scenarios is performed to evaluate the performance of the proposed methodology. Key metrics such as task completion time and reward convergence are observed to gauge the effectiveness of the approach. We observe that the Canny edge detector offers up to 40\% lower task completion time, while the CNN excels in up to 12\% better damage detection and 1.8 times better rewards.
Robot navigation has transitioned from prioritizing obstacle avoidance to adopting socially aware navigation strategies that accommodate human presence. As a result, the recognition of socially aware navigation within dynamic human-centric environments has gained prominence in the field of robotics. Although reinforcement learning technique has fostered the advancement of socially aware navigation, defining appropriate reward functions, especially in congested environments, has posed a significant challenge. These rewards, crucial in guiding robot actions, demand intricate human-crafted design due to their complex nature and inability to be automatically set. The multitude of manually designed rewards poses issues with hyperparameter redundancy, imbalance, and inadequate representation of unique object characteristics. To address these challenges, we introduce a transformable gaussian reward function (TGRF). The TGRF significantly reduces the burden of hyperparameter tuning, displays adaptability across various reward functions, and demonstrates accelerated learning rates, particularly excelling in crowded environments utilizing deep reinforcement learning (DRL). We introduce and validate TGRF through sections highlighting its conceptual background, characteristics, experiments, and real-world application, paving the way for a more effective and adaptable approach in robotics.The complete source code is available on https://github.com/JinnnK/TGRF
A traveling salesman problem (TSP) is a well-known NP-complete problem. Traditional TSP presumes that the locations of customers and the traveling time among customers are fixed and constant. In real-life cases, however, the traffic conditions and customer requests may change over time. To find the most economic route, the decisions can be made constantly upon the time-point when the salesman completes his service of each customer. This brings in a dynamic version of the traveling salesman problem (DTSP), which takes into account the information of real-time traffic and customer requests. DTSP can be extended to a dynamic pickup and delivery problem (DPDP). In this article, we ameliorate the attention model to make it possible to perceive environmental changes. A deep reinforcement learning algorithm is proposed to solve DTSP and DPDP instances with a size of up to 40 customers in 100 locations. Experiments show that our method can capture the dynamic changes and produce a highly satisfactory solution within a very short time. Compared with other baseline approaches, more than 5% improvements can be observed in many cases.