2,955 Hits in 3.4 sec

Leveraging Communication Topologies Between Learning Agents in Deep Reinforcement Learning [article]

Dhaval Adjodah, Dan Calacci, Abhimanyu Dubey, Anirudh Goyal, Peter Krafft, Esteban Moro, Alex Pentland
2020 arXiv   pre-print
A common technique to improve learning performance in deep reinforcement learning (DRL) and many other machine learning algorithms is to run multiple learning agents in parallel.  ...  Overall, our work suggests that distributed machine learning algorithms could be made more effective if the communication topology between learning agents was optimized.  ...  ACKNOWLEDGEMENTS The authors wish to thank Yan Leng for her help in early analysis of the properties of networks, Alia Braley for proofreading, and Tim Salimans for his help with replicating the OpenAI  ... 
arXiv:1902.06740v2 fatcat:57dxolh2nnba5fwjnpwjtzkpy4

DeepConfig: Automating Data Center Network Topologies Management with Machine Learning [article]

Christopher Streiffer, Huan Chen, Theophilus Benson, Asim Kadav
2017 arXiv   pre-print
We develop a framework, DeepConfig, that simplifies the processing of configuring and training deep learning agents that use the intermediate representation to learns different tasks.  ...  We present a design for developing general intermediate representations of network topologies using deep learning that is amenable to solving classes of data center problems.  ...  communication between the DeepRL agents and the network, and the DeepRL agents, called DeepConf-agents, which encapsulate data center functionality.  ... 
arXiv:1712.03890v1 fatcat:rjkbk2yzzbat7megrnudwc4yiy

Selective Network Discovery via Deep Reinforcement Learning on Embedded Spaces [article]

Peter Morales, Rajmonda Sulo Caceres, Tina Eliassi-Rad
2020 arXiv   pre-print
We propose a framework, called Network Actor Critic (NAC), which learns a policy and notion of future reward in an offline setting via a deep reinforcement learning algorithm.  ...  Finally, we outline learning regimes where planning is critical in addressing sparse and changing reward signals.  ...  Finally, [18, 21] leverage deep Reinforcement Learning techniques to learn a class of graph greedy optimization heuristics on fully observed networks.  ... 
arXiv:1909.07294v2 fatcat:ogmaukqkgrb7foyptitksvjn4e

Toward Intelligent Vehicular Networks: A Machine Learning Framework

Le Liang, Hao Ye, Geoffrey Ye Li
2019 IEEE Internet of Things Journal  
In particular, we discuss in greater detail the application of reinforcement learning in managing network resources as an alternative to the prevalent optimization approach.  ...  In this article, we first identify the distinctive characteristics of high mobility vehicular networks and motivate the use of machine learning to address the resulting challenges.  ...  The structure of reinforcement learning for V2V communications is shown in Fig. 3 , where an agent, corresponding to a V2V link, interacts with the environment.  ... 
doi:10.1109/jiot.2018.2872122 fatcat:n25uma5isfduvk3hh5mvnai4fy

Series Editorial The Fourth Issue of the Series on Machine Learning in Communications and Networks

Geoffrey Y. Li, Walid Saad, Ayfer Ozgur, Peter Kairouz, Zhijin Qin, Jakob Hoydis, Zhu Han, Deniz Gunduz, Jaafar Elmirghani
2022 IEEE Journal on Selected Areas in Communications  
HE third call for papers of the Series on Machine Learning in Communications and Networks has continued to receive a great number of high-quality papers covering various aspects of intelligent communications  ...  SIGNAL PROCESSING This issue consists of six papers that address various problems in signal processing using machine learning.  ...  A combination of state-of-the-art methods in the machine learning literature, namely, Soft-Actor Critic (SAC) reinforcement learning agents and the attention-based transformer architecture, is leveraged  ... 
doi:10.1109/jsac.2021.3126188 fatcat:6aohhlq55fco5gnndq6cusjbbi

Toward Packet Routing with Fully-distributed Multi-agent Deep Reinforcement Learning [article]

Xinyu You, Xuanjie Li, Yuedong Xu, Hui Feng, Jin Zhao, Huaicheng Yan
2019 arXiv   pre-print
In this paper, we propose a novel packet routing framework based on multi-agent deep reinforcement learning (DRL) in which each router possess an independent LSTM recurrent neural network for training  ...  Experimental results manifest that our multi-agent DRL policy can strike the delicate balance between congestion-aware and shortest routes, and significantly reduce the packet delivery time in general  ...  Index Terms-Packet routing, multi-agent learning, deep reinforcement learning, local communications I.  ... 
arXiv:1905.03494v2 fatcat:5vvgnwuh6jc57bos67h4gwgely

Towards Resilient Access Equality for 6G Serverless p-LEO Satellite Networks [article]

Lin Shih-Chun, Lin Chia-Hung, Chu Liang C., Lien Shao-Yu
2022 arXiv   pre-print
The proposed design dynamically orchestrates communications and computation functionalities and resources among heterogeneous physical units to efficiently fulfill multi-agent deep reinforcement learning  ...  , and learning performance.  ...  Specifically, a multi-agent deep reinforcement learning (MADRL) algorithm has three design components: observation processors, action predictors, and deep reinforce-ment learners (DRLs).  ... 
arXiv:2205.08430v1 fatcat:75l42xqr5zcbnat4ott22yeweq

How to Organize your Deep Reinforcement Learning Agents: The Importance of Communication Topology [article]

Dhaval Adjodah, Dan Calacci, Abhimanyu Dubey, Peter Krafft, Esteban Moro, Alex `Sandy' Pentland
2019 arXiv   pre-print
In this empirical paper, we investigate how learning agents can be arranged in more efficient communication topologies for improved learning.  ...  This is an important problem because a common technique to improve speed and robustness of learning in deep reinforcement learning and many other machine learning algorithms is to run multiple learning  ...  Optimizing the communication topology between agents is a hard problem as Related Work Running parallel (and sometimes asynchronous) agents is very common in modern deep reinforcement learning (DRL).  ... 
arXiv:1811.12556v2 fatcat:6wlehljve5aslb2gvv5w4rlonq

Modeling Complex Networks Based on Deep Reinforcement Learning

Wenbo Song, Wei Sheng, Dong Li, Chong Wu, Jun Ma
2022 Frontiers in Physics  
To overcome this limitation, we attempt to construct a network model based on deep reinforcement learning, named as NMDRL.  ...  In the new model, each node in complex networks is regarded as an intelligent agent, which reacts with the agents around it for refreshing its relationships at every moment.  ...  Based on the above facts, leveraging deep reinforcement learning to build a network model should be a meaningful attempt.  ... 
doi:10.3389/fphy.2021.822581 fatcat:teijmzaabnadbj7gxamtqffc54

A survey on Machine Learning Techniques for Routing Optimization in SDN

Rashid Amin, Elisa Rojas, Aqsa Aqdus, Sadia Ramzan, David Casillas-Perez, Jose M. Arco
2021 IEEE Access  
This article surveys the use of ML techniques for routing optimization in SDN based on three core categories (i.e. supervised learning, unsupervised learning, and reinforcement learning).  ...  In conventional networks, there was a tight bond between the control plane and the data plane.  ...  action, the state that the agent enters after taking that action, and the next action the agent chooses in its new state. 4) DEEP REINFORCEMENT LEARNING (DRL) DRL [103] is a subtype or subclass of  ... 
doi:10.1109/access.2021.3099092 fatcat:flp25cn2mbhohjxvuxgfupflny

Scalable Traffic Signal Controls using Fog-Cloud Based Multiagent Reinforcement Learning [article]

Paul Ha, Sikai Chen, Runjia Du, Samuel Labi
2021 arXiv   pre-print
This is achieved using graph attention networks (GATs) to serve as the neural network for deep reinforcement learning, which aids in maintaining the graph topology of the traffic network while disregarding  ...  Fortunately, recent studies have recognized the potential of exploiting advancements in deep and reinforcement learning to address this problem, and some preliminary successes have been achieved in this  ...  Deep reinforcement learning (DRL) combines RL with deep learning, which allows for end-toend training of multilayer models that can solve complex problems.  ... 
arXiv:2110.05564v1 fatcat:h6m3xsowo5hcbjmfl2hln3qo6a

Cognitive Caching at the Edges for Mobile Social Community Networks: A Multi-Agent Deep Reinforcement Learning Approach

Milena Radenkovic, Vu San Ha Huynh
2020 IEEE Access  
We propose a novel multi-agent deep reinforcement learning approach, CognitiveCache, in which edges adaptively learn their best caching policies while collaborating with other neighbouring edges to better  ...  INDEX TERMS Mobile social community networks, edge cloud and fog networks, content caching, deep reinforcement learning, multilayer spatial-temporal locality  ...  COGNITIVECACHE -A MULTI-AGENT DEEP REINFORCEMENT LEARNING BASED CACHING FRAMEWORK A.  ... 
doi:10.1109/access.2020.3027707 fatcat:d5oiifmwnjf4roeifsy7oortn4

Selective network discovery via deep reinforcement learning on embedded spaces

Peter Morales, Rajmonda Sulo Caceres, Tina Eliassi-Rad
2021 Applied Network Science  
We propose a framework, called network actor critic (NAC), which learns a policy and notion of future reward in an offline setting via a deep reinforcement learning algorithm.  ...  Finally, we outline learning regimes where planning is critical in addressing sparse and changing reward signals.  ...  Figure 1 gives a simple illustration of how this cumulative reward is computed over a network topology. In the next section, we describe in detail our deep reinforcement learning algorithm.  ... 
doi:10.1007/s41109-021-00365-8 fatcat:bh5bze3ykreddbtzn5vbu4jene

Feature Engineering for Deep Reinforcement Learning Based Routing

Jose Suarez-Varela, Albert Mestres, Junlin Yu, Li Kuang, Haoyu Feng, Pere Barlet-Ros, Albert Cabellos-Aparicio
2019 ICC 2019 - 2019 IEEE International Conference on Communications (ICC)  
Recent advances in Deep Reinforcement Learning (DRL) techniques are providing a dramatic improvement in decision-making and automated control problems.  ...  We test our representation in two different scenarios: (i) routing in optical transport networks and (ii) QoS-aware routing in IP networks.  ...  DEEP REINFORCEMENT LEARNING IN ROUTING The objective of Deep Reinforcement Learning is to learn the policy that leads to a maximum cumulative reward.  ... 
doi:10.1109/icc.2019.8761276 dblp:conf/icc/Suarez-VarelaMY19 fatcat:cuz2b3eupranldafs5k3qta6tq

Deep Learning based Wireless Resource Allocation with Application to Vehicular Networks [article]

Le Liang, Hao Ye, Guanding Yu, Geoffrey Ye Li
2019 arXiv   pre-print
We then highlight the deep reinforcement learning approach to address resource allocation problems that are difficult to handle in the traditional optimization framework.  ...  Meanwhile, deep learning, with many success stories in various disciplines, represents a promising alternative due to its remarkable power to leverage data for problem solving.  ...  Deep Q-learning, REINFORCE, actor-critic deep deterministic policy gradient (DDPG) algorithms have been leveraged to train the RL agent and all of them outperform the benchmark WMMSE [52] and FP algorithms  ... 
arXiv:1907.03289v2 fatcat:stvlo3uhqnde3icvovjsrggdwu
« Previous Showing results 1 — 15 out of 2,955 results