Social capital designs the ensemble of resources which are accessible to a social actor through i... more Social capital designs the ensemble of resources which are accessible to a social actor through its relationship with other actors [1]. As such, social capital is always embedded in a social network, where nodes represent actors (eg individuals, political groups, associations), and links represent relationships (eg, friendship, common interests, hostility)[2]. Social networks are inherently dynamic: over time, new actors enter the scene while old ones leave, and relationships among them change continuously. Intuitively, the dynamics of the ...
ABSTRACT Task partitioning consists in dividing a task into sub-tasks that can be tackled separat... more ABSTRACT Task partitioning consists in dividing a task into sub-tasks that can be tackled separately. Partitioning a task might have both positive and negative effects: On the one hand, partitioning might reduce physical interference between workers, enhance exploitation of specialization, and increase efficiency. On the other hand, partitioning may introduce overheads due to coordination requirements. As a result, whether partitioning is advantageous or not has to be evaluated on a case-by-case basis. In this paper we consider the case in which a swarm of robots must decide whether to complete a given task as an unpartitioned task, or utilize task partitioning and tackle it as a sequence of two sub-tasks. We show that the problem of selecting between the two options can be formulated as a multi-armed bandit problem and tackled with algorithms that have been proposed in the reinforcement learning literature. Additionally, we study the implications of using explicit communication between the robots to tackle the studied task partitioning problem. We consider a foraging scenario as a testbed and we perform simulation-based experiments to evaluate the behavior of the system. The results confirm that existing multi-armed bandit algorithms can be employed in the context of task partitioning. The use of communication can result in better performance, but in may also hinder the flexibility of the system.
30th Benelux Meeting on Systems and Control—Book of Abstracts, 2011
Skip to main content. VUB Artificial Intelligence Lab. Search form. Search. You are here. Home. R... more Skip to main content. VUB Artificial Intelligence Lab. Search form. Search. You are here. Home. Reinforcement learning for repetitive systems with discrete sensors. Title, Reinforcement learning for repetitive systems with discrete sensors. Publication Type, Conference Paper. Year of Publication, 2011. ...
Abstract This paper presents an overview of model-based (Iterative Learning Control, Model Predic... more Abstract This paper presents an overview of model-based (Iterative Learning Control, Model Predictive Control and Iterative Optimization) and non-model-based (Genetic-based Machine Learning and Reinforcement Learning) learning strategies for the control of wet clutches. Based on theoretical considerations and a validation on an experimental test bench containing wet clutches, the benefits and drawbacks of the different strategies are compared. Although after convergence a good engagement quality can be obtained by all ...
Social capital designs the ensemble of resources which are accessible to a social actor through i... more Social capital designs the ensemble of resources which are accessible to a social actor through its relationship with other actors [1]. As such, social capital is always embedded in a social network, where nodes represent actors (eg individuals, political groups, associations), and links represent relationships (eg, friendship, common interests, hostility)[2]. Social networks are inherently dynamic: over time, new actors enter the scene while old ones leave, and relationships among them change continuously. Intuitively, the dynamics of the ...
Social capital is naturally embedded in social networks. In his famous work on the causal relatio... more Social capital is naturally embedded in social networks. In his famous work on the causal relationship between “bridging” social capital (eg, associational life), trust, and civic behavior, Putnam [1993] did not investigate the structural aspect of such networks. Recently, the relationship between associational life and civicness of ethnic minority groups in Europe has been investigated [Fennema and Tillie, 2001; Jacobs et al., 2004; Vermeulen and Berger, 2008], without reaching uniform conclusions.
Abstract A common approach when applying reinforcement learning to address control problems is th... more Abstract A common approach when applying reinforcement learning to address control problems is that of first learning a policy based on an approximated model of the plant, whose behavior can be quickly and safely explored in simulation; and then implementing the obtained policy to control the actual plant. Here we follow this approach to learn to engage a transmission clutch, with the aim of obtaining a rapid and smooth engagement, with a small torque loss.
Abstract This thesis presents methods for minimizing the computational effort of problem solving.... more Abstract This thesis presents methods for minimizing the computational effort of problem solving. Rather than looking at a particular algorithm, we consider the issue of computational complexity at a higher level, and propose techniques that, given a set of candidate algorithms, of unknown performance, learn to use these algorithms while solving a sequence of problem instances, with the aim of solving all instances in a minimum time.
Abstract In most existing motion control algorithms, a reference trajectory is tracked, based on ... more Abstract In most existing motion control algorithms, a reference trajectory is tracked, based on a continuous measurement of the system's response. In many industrial applications, however, it is either not possible or too expensive to install sensors which measure the system's output over the complete stroke: instead, the motion can only be detected at certain discrete positions.
Social capital designs the ensemble of resources which are accessible to a social actor through i... more Social capital designs the ensemble of resources which are accessible to a social actor through its relationship with other actors [1]. As such, social capital is always embedded in a social network, where nodes represent actors (eg individuals, political groups, associations), and links represent relationships (eg, friendship, common interests, hostility)[2]. Social networks are inherently dynamic: over time, new actors enter the scene while old ones leave, and relationships among them change continuously. Intuitively, the dynamics of the ...
ABSTRACT Task partitioning consists in dividing a task into sub-tasks that can be tackled separat... more ABSTRACT Task partitioning consists in dividing a task into sub-tasks that can be tackled separately. Partitioning a task might have both positive and negative effects: On the one hand, partitioning might reduce physical interference between workers, enhance exploitation of specialization, and increase efficiency. On the other hand, partitioning may introduce overheads due to coordination requirements. As a result, whether partitioning is advantageous or not has to be evaluated on a case-by-case basis. In this paper we consider the case in which a swarm of robots must decide whether to complete a given task as an unpartitioned task, or utilize task partitioning and tackle it as a sequence of two sub-tasks. We show that the problem of selecting between the two options can be formulated as a multi-armed bandit problem and tackled with algorithms that have been proposed in the reinforcement learning literature. Additionally, we study the implications of using explicit communication between the robots to tackle the studied task partitioning problem. We consider a foraging scenario as a testbed and we perform simulation-based experiments to evaluate the behavior of the system. The results confirm that existing multi-armed bandit algorithms can be employed in the context of task partitioning. The use of communication can result in better performance, but in may also hinder the flexibility of the system.
30th Benelux Meeting on Systems and Control—Book of Abstracts, 2011
Skip to main content. VUB Artificial Intelligence Lab. Search form. Search. You are here. Home. R... more Skip to main content. VUB Artificial Intelligence Lab. Search form. Search. You are here. Home. Reinforcement learning for repetitive systems with discrete sensors. Title, Reinforcement learning for repetitive systems with discrete sensors. Publication Type, Conference Paper. Year of Publication, 2011. ...
Abstract This paper presents an overview of model-based (Iterative Learning Control, Model Predic... more Abstract This paper presents an overview of model-based (Iterative Learning Control, Model Predictive Control and Iterative Optimization) and non-model-based (Genetic-based Machine Learning and Reinforcement Learning) learning strategies for the control of wet clutches. Based on theoretical considerations and a validation on an experimental test bench containing wet clutches, the benefits and drawbacks of the different strategies are compared. Although after convergence a good engagement quality can be obtained by all ...
Social capital designs the ensemble of resources which are accessible to a social actor through i... more Social capital designs the ensemble of resources which are accessible to a social actor through its relationship with other actors [1]. As such, social capital is always embedded in a social network, where nodes represent actors (eg individuals, political groups, associations), and links represent relationships (eg, friendship, common interests, hostility)[2]. Social networks are inherently dynamic: over time, new actors enter the scene while old ones leave, and relationships among them change continuously. Intuitively, the dynamics of the ...
Social capital is naturally embedded in social networks. In his famous work on the causal relatio... more Social capital is naturally embedded in social networks. In his famous work on the causal relationship between “bridging” social capital (eg, associational life), trust, and civic behavior, Putnam [1993] did not investigate the structural aspect of such networks. Recently, the relationship between associational life and civicness of ethnic minority groups in Europe has been investigated [Fennema and Tillie, 2001; Jacobs et al., 2004; Vermeulen and Berger, 2008], without reaching uniform conclusions.
Abstract A common approach when applying reinforcement learning to address control problems is th... more Abstract A common approach when applying reinforcement learning to address control problems is that of first learning a policy based on an approximated model of the plant, whose behavior can be quickly and safely explored in simulation; and then implementing the obtained policy to control the actual plant. Here we follow this approach to learn to engage a transmission clutch, with the aim of obtaining a rapid and smooth engagement, with a small torque loss.
Abstract This thesis presents methods for minimizing the computational effort of problem solving.... more Abstract This thesis presents methods for minimizing the computational effort of problem solving. Rather than looking at a particular algorithm, we consider the issue of computational complexity at a higher level, and propose techniques that, given a set of candidate algorithms, of unknown performance, learn to use these algorithms while solving a sequence of problem instances, with the aim of solving all instances in a minimum time.
Abstract In most existing motion control algorithms, a reference trajectory is tracked, based on ... more Abstract In most existing motion control algorithms, a reference trajectory is tracked, based on a continuous measurement of the system's response. In many industrial applications, however, it is either not possible or too expensive to install sensors which measure the system's output over the complete stroke: instead, the motion can only be detected at certain discrete positions.
Workshop ‘Digital Media, Power, and Democracy in Election Campaigns’', Washington, DC, Jul 2, 2015
The case of the May 2014 Belgian elections offers a unique opportunity to investigate
campaign b... more The case of the May 2014 Belgian elections offers a unique opportunity to investigate
campaign behaviors. We explore interactions between candidates on Twitter in the run up to
the elections of three different legislative bodies: the federal, regional and European
assemblies. We define interactions based on the flow of retweets and conversations (@)
initiated by candidates. In so doing we capture dynamic interactions or networks, as opposed
to a more static definition based on followers. These multilayer elections enable us to map the
positions of candidates, explore whether Twitter based networks go beyond the language
divide (French/Flemish) and if ideological boundaries characterize candidates ‘interactions.
Interactions are analysed using Ucinet. Finally, we explore the extent to which visibility
within the network of candidates is fully explained by candidates’ offline profiles.
Uploads
campaign behaviors. We explore interactions between candidates on Twitter in the run up to
the elections of three different legislative bodies: the federal, regional and European
assemblies. We define interactions based on the flow of retweets and conversations (@)
initiated by candidates. In so doing we capture dynamic interactions or networks, as opposed
to a more static definition based on followers. These multilayer elections enable us to map the
positions of candidates, explore whether Twitter based networks go beyond the language
divide (French/Flemish) and if ideological boundaries characterize candidates ‘interactions.
Interactions are analysed using Ucinet. Finally, we explore the extent to which visibility
within the network of candidates is fully explained by candidates’ offline profiles.