The Experts below are selected from a list of 3201 Experts worldwide ranked by ideXlab platform
Olechristoffer Granmo - One of the best experts on this subject based on the ideXlab platform.
-
a novel multi step finite state Automaton for arbitrarily deterministic tsetlin machine Learning
International Conference on Innovative Techniques and Applications of Artificial Intelligence, 2020Co-Authors: Darshana K Abeyrathna, Olechristoffer Granmo, Alex Yakovlev, Adrian Wheeldon, Jie Lei, Rishad Shafik, Morten GoodwinAbstract:Due to the high energy consumption and scalability challenges of deep Learning, there is a critical need to shift research focus towards dealing with energy consumption constraints. Tsetlin Machines (TMs) are a recent approach to machine Learning that has demonstrated significantly reduced energy usage compared to neural networks alike, while performing competitively accuracy-wise on several benchmarks. However, TMs rely heavily on energy-costly random number generation to stochastically guide a team of Tsetlin Automata (TA) to a Nash Equilibrium of the TM game. In this paper, we propose a novel finite-state Learning Automaton that can replace the TA in TM Learning, for increased determinism. The new Automaton uses multi-step deterministic state jumps to reinforce sub-patterns. Simultaneously, flipping a coin to skip every d’th state update ensures diversification by randomization. The d-parameter thus allows the degree of randomization to be finely controlled. E.g., \(d=1\) makes every update random and \(d=\infty \) makes the Automaton completely deterministic. Our empirical results show that, overall, only substantial degrees of determinism reduces accuracy. Energy-wise, random number generation constitutes switching energy consumption of the TM, saving up to 11 mW power for larger datasets with high d values. We can thus use the new d-parameter to trade off accuracy against energy consumption, to facilitate low-energy machine Learning.
-
a novel multi step finite state Automaton for arbitrarily deterministic tsetlin machine Learning
arXiv: Learning, 2020Co-Authors: Darshana K Abeyrathna, Olechristoffer Granmo, Rishad A Shafik, Alex Yakovlev, Adrian Wheeldon, Jie Lei, Morten GoodwinAbstract:Due to the high energy consumption and scalability challenges of deep Learning, there is a critical need to shift research focus towards dealing with energy consumption constraints. Tsetlin Machines (TMs) are a recent approach to machine Learning that has demonstrated significantly reduced energy usage compared to neural networks alike, while performing competitively accuracy-wise on several benchmarks. However, TMs rely heavily on energy-costly random number generation to stochastically guide a team of Tsetlin Automata to a Nash Equilibrium of the TM game. In this paper, we propose a novel finite-state Learning Automaton that can replace the Tsetlin Automata in TM Learning, for increased determinism. The new Automaton uses multi-step deterministic state jumps to reinforce sub-patterns. Simultaneously, flipping a coin to skip every $d$'th state update ensures diversification by randomization. The $d$-parameter thus allows the degree of randomization to be finely controlled. E.g., $d=1$ makes every update random and $d=\infty$ makes the Automaton completely deterministic. Our empirical results show that, overall, only substantial degrees of determinism reduces accuracy. Energy-wise, random number generation constitutes switching energy consumption of the TM, saving up to 11 mW power for larger datasets with high $d$ values. We can thus use the new $d$-parameter to trade off accuracy against energy consumption, to facilitate low-energy machine Learning.
-
A Conclusive Analysis of the Finite-Time Behavior of the Discretized Pursuit Learning Automaton
IEEE Transactions on Neural Networks and Learning Systems, 2020Co-Authors: Xuan Zhang, Lei Jiao, John B. Oommen, Olechristoffer GranmoAbstract:This paper deals with the finite-time analysis (FTA) of Learning automata (LA), which is a topic for which very little work has been reported in the literature. This is as opposed to the asymptotic steady-state analysis for which there are, probably, scores of papers. As clarified later, unarguably, the FTA of Markov chains, in general, and of LA, in particular, is far more complex than the asymptotic steady-state analysis. Such an FTA provides rigid bounds for the time required for the LA to attain to a given convergence accuracy. We concentrate on the FTA of the Discretized Pursuit Automaton (DPA), which is probably one of the fastest and most accurate reported LA. Although such an analysis was carried out many years ago, we record that the previous work is flawed. More specifically, in all brevity, the flaw lies in the wrongly “derived” monotonic behavior of the LA after a certain number of iterations. Rather, we claim that the property should be invoked is the submartingale property. This renders the proof to be much more involved and deep. In this paper, we rectify the flaw and reestablish the FTA based on such a submartingale phenomenon. More importantly, from the derived analysis, we are able to discover and clarify, for the first time, the underlying dilemma between the DPA's exploitation and exploration properties. We also nontrivially confirm the existence of the optimal Learning rate, which yields a better comprehension of the DPA itself.
-
Generalized Bayesian Pursuit: A Novel Scheme for Multi-Armed Bernoulli Bandit Problems
2011Co-Authors: Xuan Zhang, B. Oommen, Olechristoffer GranmoAbstract:In the last decades, a myriad of approaches to the multi-armed bandit problem have appeared in several different fields. The current top performing algorithms from the field of Learning Automata reside in the Pursuit family, while UCB-Tuned and the ε-greedy class of algorithms can be seen as state-of-the-art regret minimizing algorithms. Recently, however, the Bayesian Learning Automaton (BLA) outperformed all of these, and other schemes, in a wide range of experiments. Although seemingly incompatible, in this paper we integrate the foundational Learning principles motivating the design of the BLA, with the principles of the so-called Generalized Pursuit algorithm (GPST), leading to the Generalized Bayesian Pursuit algorithm (GBPST). As in the BLA, the estimates are truly Bayesian in nature, however, instead of basing exploration upon direct sampling from the estimates, GBPST explores by means of the arm selection probability vector of GPST. Further, as in the GPST, in the interest of higher rates of Learning, a set of arms that are currently perceived as being optimal is pursued to minimize the probability of pursuing a wrong arm. It turns out that GBPST is superior to GPST and that it even performs better than the BLA by controlling the Learning speed of GBPST. We thus believe that GBPST constitutes a new avenue of research, in which the performance benefits of the GPST and the BLA are mutually augmented, opening up for improved performance in a number of applications, currently being tested.
-
Learning Automaton based on line discovery and tracking of spatio temporal event patterns
Pacific Rim International Conference on Artificial Intelligence, 2010Co-Authors: Anis Yazidi, Olechristoffer Granmo, John B. Oommen, Min Lin, Xifeng Wen, Martin Gerdes, Frank ReichertAbstract:Discovering and tracking of spatio-temporal patterns in noisy sequences of events is a difficult task that has become increasingly pertinent due to recent advances in ubiquitous computing, such as community-based social networking applications. The core activities for applications of this class include the sharing and notification of events, and the importance and usefulness of these functionalites increases as event-sharing expands into larger areas of one's life. Ironically, instead of being helpful, an excessive number of event notifications can quickly render the functionality of event-sharing to be obtrusive. Rather, any notification of events that provides redundant information to the application/user can be seen to be an unnecessary distraction. In this paper, we introduce a new scheme for discovering and tracking noisy spatio-temporal event patterns, with the purpose of suppressing reoccurring patterns, while discerning novel events. Our scheme is based on maintaining a collection of hypotheses, each one conjecturing a specific spatio-temporal event pattern. A dedicated Learning Automaton (LA) - the Spatio-Temporal Pattern LA (STPLA) - is associated with each hypothesis. By processing events as they unfold, we attempt to infer the correctness of each hypothesis through a real-time guided random walk. Consequently, the scheme we present is computationally efficient, with a minimal memory footprint. Furthermore, it is ergodic, allowing adaptation. Empirical results involving extensive simulations demonstrate the STPLA's superior convergence and adaptation speed, as well as an ability to operate successfully with noise, including both the erroneous inclusion and omission of events. Additionally, the results included, which involve a so-called "Presence Sharing" application, are both promising and in our opinion, impressive. It is thus our opinion that the proposed STPLA scheme is, in general, ideal for improving the usefulness of event notification and sharing systems, since it is capable of significantly, robustly and adaptively suppressing redundant information.
Mohammad Reza Meybodi - One of the best experts on this subject based on the ideXlab platform.
-
adaptive petri net based on irregular cellular Learning automata with an application to vertex coloring problem
Applied Intelligence, 2017Co-Authors: Mehdi S Vahidipour, Mohammad Reza Meybodi, Mehdi EsnaashariAbstract:An adaptive Petri net, called APN-LA, that has been recently introduced, uses a set of Learning automata for controlling possible conflicts among the transitions in a Petri net (PN). Each Learning Automaton (LA) in APN-LA acts independently from the others, but there could be situations, where the operation of a LA affects the operation of another LA by possibly enabling or disabling some of the transitions within the control of that LA. In such situations, it is more appropriate to let the Learning automata within the APN-LA, cooperate with each other, instead of operating independently. In this paper, an adaptive Petri net system based on Irregular Cellular Learning Automata (ICLA), in which a number of Learning automata cooperate with each other, is proposed. The proposed adaptive system, called APN-ICLA, consists of two layers: PN-layer and an ICLA-layer. The PN-layer is a Petri net, in which conflicting transitions are partitioned into several clusters. There should be a controller in each cluster to control the possible conflicts among the transitions in that cluster. The ICLA-layer in APN-ICLA provides the required controllers for the PN-layer. The ICLA-layer is indeed an ICLA, in which each cell corresponds to a cluster in the PN-layer. The LA resides in a particular cell in the ICLA-layer and acts as the controller of the corresponding cluster in the PN-layer. To evaluate the efficiency of the proposed system, several algorithms, based on the APN-ICLA for vertex coloring problem, are designed. Simulation results justify the effectiveness of the proposed APN-ICLA.
-
a new memetic algorithm based on cellular Learning automata for solving the vertex coloring problem
Memetic Computing, 2016Co-Authors: Mehdi Rezapoor Mirsaleh, Mohammad Reza MeybodiAbstract:Vertex coloring problem is a combinatorial optimization problem in graph theory in which a color is assigned to each vertex of graph such that no two adjacent vertices have the same color. In this paper a new hybrid algorithm which is obtained from combination of cellular Learning automata (CLA) and memetic algorithm (MA) is proposed for solving the vertex coloring problem. CLA is an effective probabilistic Learning model combining cellular automata and Learning Automaton (LA). Irregular CLA (ICLA) is a generalization of CLA in which the restriction of rectangular grid structure in CLA is removed. The proposed algorithm is based on the irregular open CLA (IOCLA) that is an extension of ICLA in which the evolution of CLA is influenced by both local and global environments. Similar to other IOCLA-based algorithms, in the proposed algorithm, local environment is constituted by neighboring LAs of any cell and the global environment consists of a pool of memes in which each meme corresponds to a certain local search method. Each meme is represented by a set of LAs from which the history of the corresponding local search method can be extracted. To show the superiority of the proposed algorithm over some well-known algorithms, several computer experiments have been conducted. The results show that the proposed algorithm performs better than other methods in terms of running time of algorithm and the required number of colors.
-
A Cellular Learning Automata-based Algorithm for Solving the Coverage and Connectivity Problem in Wireless Sensor Networks
Ad Hoc & Sensor Wireless Networks, 2014Co-Authors: Reza Ghaderi, Mehdi Esnaashari, Mohammad Reza MeybodiAbstract:Presence of redundant nodes is common in wireless sensor networks because of various reasons such as high probability of failures and necessity of long lifetime. When such redundancy exists, some distributed algorithms are needed for selecting minimal subset of nodes as active nodes in a manner that network area is covered entirely with the selected active nodes. In this paper, a distributed algorithm is proposed which attempts to minimize the number of active nodes in the network using cellular Learning automata in such a way that the following two conditions are met: 1. network area is covered entirely, and 2. network of selected active nodes is connected. In the proposed algorithm, each node is equipped with a Learning Automaton which locally decides for the node to be active or not based on the remaining energy of the node and its neighbors’ situations. To ensure the network connectivity, we analytically determine the radio transmission range of sensor nodes according to their sensing range so that complete coverage of the network area guarantees the connectivity of active nodes. The time and space costs of the proposed algorithm are analytically determined and compared with those of similar existing algorithms such as PEAS and PECAS. Simulation results in J-Sim simulator environment specify the efficiency of the proposed algorithm over existing algorithms such as PEAS and PECAS—especially against high ratio of unexpected failures and nodes’ energy depletion.
-
maximizing lifetime of target coverage in wireless sensor networks using Learning automata
Wireless Personal Communications, 2013Co-Authors: Habib Mostafaei, Mohammad Reza MeybodiAbstract:In wireless sensor networks, when each target is covered by multiple sensors, we can schedule sensor nodes to monitor deployed targets in order to improve lifetime of network. In this paper, we propose an efficient scheduling method based on Learning automata, in which each node is equipped with a Learning Automaton, which helps the node to select its proper state (active or sleep), at any given time. To study the performance of the proposed method, computer simulations are conducted. Results of these simulations show that the proposed scheduling method can better prolong the lifetime of the network in comparison to similar existing methods.
-
a cellular Learning automata based deployment strategy for mobile wireless sensor networks
Journal of Parallel and Distributed Computing, 2011Co-Authors: Mehdi Esnaashari, Mohammad Reza MeybodiAbstract:One important problem which may arise in designing a deployment strategy for a wireless sensor network is how to deploy a specific number of sensor nodes throughout an unknown network area so that the covered section of the area is maximized. In a mobile sensor network, this problem can be addressed by first deploying sensor nodes randomly in some initial positions within the area of the network, and then letting sensor nodes to move around and find their best positions according to the positions of their neighboring nodes. The problem becomes more complicated if sensor nodes have no information about their positions or even their relative distances to each other. In this paper, we propose a cellular Learning automata-based deployment strategy which guides the movements of sensor nodes within the area of the network without any sensor to know its position or its relative distance to other sensors. In the proposed algorithm, the Learning Automaton in each node in cooperation with the Learning automata in the neighboring nodes controls the movements of the node in order to attain high coverage. Experimental results have shown that in noise-free environments, the proposed algorithm can compete with the existing algorithms such as PF, DSSA, IDCA, and VEC in terms of network coverage. It has also been shown that in noisy environments, where utilized location estimation techniques such as GPS-based devices and localization algorithms experience inaccuracies in their measurements, or the movements of sensor nodes are not perfect and follow a probabilistic motion model, the proposed algorithm outperforms the existing algorithms in terms of network coverage.
John A Copeland - One of the best experts on this subject based on the ideXlab platform.
-
reinforcement Learning for repeated power control game in cognitive radio networks
IEEE Journal on Selected Areas in Communications, 2012Co-Authors: Pan Zhou, Yusun Chang, John A CopelandAbstract:Cognitive radio (CR) users are expected to be uncoordinated users that opportunistically seek the spectrum resource from primary users (PUs) in a competitive way. In most existing works, however, CR users are required to share the interference channel information and power strategies to conduct the game with pricing mechanisms that incur the frequent exchange of information. The requirement of significant communication overheads among CR users impedes fully distributed solutions for the deployment of CR networks, which is a challenging problem in the research communities. In this paper, a robust distributed power control algorithm is designed with low implementation complexity for CR networks through reinforcement Learning, which does not require the interference channel and power strategy information among CR users (and from CR users to PUs). To the best of our knowledge, this research provides the solution for the first time for the incomplete-information power control game in CR networks. During the repeated game, CR users can control their power strategies by observing the interference from the feedback signals of PUs and transmission rates obtained in the previous step. This procedure allows achieving high spectrum efficiency while conforming to the interference constraint of PUs. This constrained repeated stochastic game with Learning Automaton is proved to be asymptotically equivalence to the traditional game with complete information. The properties of existence, diagonal concavity and uniqueness for the game are studied. A Bush-Mosteller reinforcement Learning procedure is designed for the power control algorithm, and the properties of convergence and Learning rate of the algorithm are analyzed. The performance of the Learning-based power control algorithm is thoroughly investigated with simulation results, which demonstrates the effectiveness of the proposed algorithm in solving variety of practical CR network problems for real-world applications.
-
Learning through reinforcement for repeated power control game in cognitive radio networks
Global Communications Conference, 2010Co-Authors: Pan Zhou, Yusun Chang, John A CopelandAbstract:This paper studies the repeated power control game in cognitive radio (CR) networks through reinforcement Learning without channel and power strategy information exchange among CR users. Unlike traditional game-theoretical approaches on CR power control, this research solves the incomplete information power control problems for selfish and autonomous CR users for the first time. Each CR user in the problem only knows its own channel and power strategy while the information of primary users (PUs) and other different types of CR users are unknown. The formulated power control problem is a constrained repeated stochastic game with Learning Automaton. The objective of this repeated game is to maximize the average utility of each CR user under the interference power constraints of PUs. At each time step, the CR user only knows its own utility and the interference functions after the play but no further information. This power control game is proved to be asymptotically equivalent to the traditional game theory approaches. The properties of existence, diagonal concavity and uniqueness for this game are illustrated in detail. A Bush-Mosteller reinforcement Learning procedure is designed for the power control algorithm. Finally, the Learning based power control algorithm is implemented, and the simulation results with detailed analysis are shown to enforce the effectiveness of the proposed algorithms.
Branka Vucetic - One of the best experts on this subject based on the ideXlab platform.
-
Socially aware caching strategy in device-to-device communication networks
IEEE Transactions on Vehicular Technology, 2018Co-Authors: Chuan Ma, Zihuai Lin, Guoqiang Mao, He Chen, Ying-chang Liang, Ming Ding, Branka VuceticAbstract:As a response to the challenge of data traffic explosion in wireless networks, content caching in device-to-device (D2D) communication networks has emerged as a promising solution. However, in practical deployment, D2D content caching has its own problems. In particular, not all of the user devices are willing to share the content with others due to numerous concerns, such as security, battery life, and social relationship. In this paper, we consider the factor of social relationship in the deployment of D2D content caching. First, we apply stochastic geometry theory to derive an analytical expression of downloading performance for the D2D caching network. Specifically, a social relationship model with respect to the physical distance is adopted in our analysis to obtain the average download delay performance using random and deterministic caching strategies. Second, to achieve a better performance in more practical and specific scenarios, we develop a socially aware distributed caching strategy based on a decentralized Learning Automaton, to optimize the cache placement operation in D2D networks. Different from the existing caching schemes, the proposed algorithm not only considers the file request probability and the closeness of devices as measured by their physical distance but also takes into account the social relationship between D2D users. Our simulation results show that the proposed algorithm can converge quickly and outperforms the random and deterministic caching strategies. With these results, our work sheds insights on the design of D2D caching in the practical deployment of 5G networks.
Mehdi Esnaashari - One of the best experts on this subject based on the ideXlab platform.
-
adaptive petri net based on irregular cellular Learning automata with an application to vertex coloring problem
Applied Intelligence, 2017Co-Authors: Mehdi S Vahidipour, Mohammad Reza Meybodi, Mehdi EsnaashariAbstract:An adaptive Petri net, called APN-LA, that has been recently introduced, uses a set of Learning automata for controlling possible conflicts among the transitions in a Petri net (PN). Each Learning Automaton (LA) in APN-LA acts independently from the others, but there could be situations, where the operation of a LA affects the operation of another LA by possibly enabling or disabling some of the transitions within the control of that LA. In such situations, it is more appropriate to let the Learning automata within the APN-LA, cooperate with each other, instead of operating independently. In this paper, an adaptive Petri net system based on Irregular Cellular Learning Automata (ICLA), in which a number of Learning automata cooperate with each other, is proposed. The proposed adaptive system, called APN-ICLA, consists of two layers: PN-layer and an ICLA-layer. The PN-layer is a Petri net, in which conflicting transitions are partitioned into several clusters. There should be a controller in each cluster to control the possible conflicts among the transitions in that cluster. The ICLA-layer in APN-ICLA provides the required controllers for the PN-layer. The ICLA-layer is indeed an ICLA, in which each cell corresponds to a cluster in the PN-layer. The LA resides in a particular cell in the ICLA-layer and acts as the controller of the corresponding cluster in the PN-layer. To evaluate the efficiency of the proposed system, several algorithms, based on the APN-ICLA for vertex coloring problem, are designed. Simulation results justify the effectiveness of the proposed APN-ICLA.
-
A Cellular Learning Automata-based Algorithm for Solving the Coverage and Connectivity Problem in Wireless Sensor Networks
Ad Hoc & Sensor Wireless Networks, 2014Co-Authors: Reza Ghaderi, Mehdi Esnaashari, Mohammad Reza MeybodiAbstract:Presence of redundant nodes is common in wireless sensor networks because of various reasons such as high probability of failures and necessity of long lifetime. When such redundancy exists, some distributed algorithms are needed for selecting minimal subset of nodes as active nodes in a manner that network area is covered entirely with the selected active nodes. In this paper, a distributed algorithm is proposed which attempts to minimize the number of active nodes in the network using cellular Learning automata in such a way that the following two conditions are met: 1. network area is covered entirely, and 2. network of selected active nodes is connected. In the proposed algorithm, each node is equipped with a Learning Automaton which locally decides for the node to be active or not based on the remaining energy of the node and its neighbors’ situations. To ensure the network connectivity, we analytically determine the radio transmission range of sensor nodes according to their sensing range so that complete coverage of the network area guarantees the connectivity of active nodes. The time and space costs of the proposed algorithm are analytically determined and compared with those of similar existing algorithms such as PEAS and PECAS. Simulation results in J-Sim simulator environment specify the efficiency of the proposed algorithm over existing algorithms such as PEAS and PECAS—especially against high ratio of unexpected failures and nodes’ energy depletion.
-
a cellular Learning automata based deployment strategy for mobile wireless sensor networks
Journal of Parallel and Distributed Computing, 2011Co-Authors: Mehdi Esnaashari, Mohammad Reza MeybodiAbstract:One important problem which may arise in designing a deployment strategy for a wireless sensor network is how to deploy a specific number of sensor nodes throughout an unknown network area so that the covered section of the area is maximized. In a mobile sensor network, this problem can be addressed by first deploying sensor nodes randomly in some initial positions within the area of the network, and then letting sensor nodes to move around and find their best positions according to the positions of their neighboring nodes. The problem becomes more complicated if sensor nodes have no information about their positions or even their relative distances to each other. In this paper, we propose a cellular Learning automata-based deployment strategy which guides the movements of sensor nodes within the area of the network without any sensor to know its position or its relative distance to other sensors. In the proposed algorithm, the Learning Automaton in each node in cooperation with the Learning automata in the neighboring nodes controls the movements of the node in order to attain high coverage. Experimental results have shown that in noise-free environments, the proposed algorithm can compete with the existing algorithms such as PF, DSSA, IDCA, and VEC in terms of network coverage. It has also been shown that in noisy environments, where utilized location estimation techniques such as GPS-based devices and localization algorithms experience inaccuracies in their measurements, or the movements of sensor nodes are not perfect and follow a probabilistic motion model, the proposed algorithm outperforms the existing algorithms in terms of network coverage.
