Synchronization

Patterns of Synchronization of Non-verbal Cues and Speech in ECAs: Towards a More "Natural" Conversational Agent

by Nicla Rossini

In Esposito, A., Esposito, A. M., Martone, R. Mueller, V. C., Scarpetta, G. (Eds) Toward Autonomous, Adaptive, and Context-Aware Multimodal Interfaces: Theoretical and Practical Issues. Pp. 97-104. Springer-Verlag: Berlin.

N.B.: THIS IS A PRE-PUBLICATION EARLY DRAFT. IT MAY CONTAIN ERROS AND DIFFER SIGNIFICANTLY FROM THE PUBLISHED VERSION. FOR QUOTATION PURPOSES, PLEASE ASK ME A COPY OF THE PUBLISHED VERSION.

Please find it on Springer online: http://www.springerlink.com/content/l410238375431524/

This paper presents an analysis of the verbal and non-verbal cues of Conversational Agents, with a special focus on... more This paper presents an analysis of the verbal and non-verbal cues of Conversational Agents, with a special focus on REA and GRETA, in order to allow further research aimed at correcting some traits of their performance still considered unnatural by their final users. Despite the striking performance of new generation ECA, some important features make these conversational agents unreliable to the users, who usually prefer interacting with a classical computer for information retrieval. The users’ preference can be due to several factors, such as the quality of speech synthesis, or the inevitable unnaturalness of the graphics animating the avatar. Apart from the unavoidable traits that can render ECAs unnatural to the ultimate users, instances of poor synchronization between verbal and non-verbal behaviour may contribute to unfavourable results. An instance of synchronization patterns between non-verbal cues and speech is here analysed and re-applied to the basic architecture of an ECA in order to improve the ECA’s verbal and non-verbal synchronization. A proposal for future inquiry aimed at creating alternative model for the ultimate Mp4 output is also proposed, for further development in this field.

Ph.D Thesis : Neural dynamics of synchronous imitative interaction

by Guillaume Dumas

Supervisors: Jacqueline Nadel & Jacques Martinerie (formerly Line Garnero)
Defended the 3rd of October 2011

Since 2002, a new neuroimaging technique called hyperscanning allows to record several participants simultaneously and... more Since 2002, a new neuroimaging technique called hyperscanning allows to record several participants simultaneously and thus to study social interaction in a reciprocal and spontaneous social context. Meanwhile, neurodynamics and sensorimotor theories suggested to understand social interaction in a more holistic approach by considering the two interacting individuals as a single system, and giving equal importance to behavior and brain activity.

This thesis presents the study of pairs of participants recorded simultaneously during spontaneous imitation of the movement of their hands, by a dual-video combined with a hyperscanning-EEG setup. A fine grained video analysis identified the episodes of interactional synchrony and imitation, thus allowing the neurodynamic characterization of various aspects of the interaction, both at the inter- and intra-individual. The first study showed that episodes of interactional synchrony were accompanied by the emergence of inter-brain phase synchronizations in several frequency bands. The second study showed a neural differentiation between self- and other-attribution of action primacy, and found a signature of the co-ownership of the action in both partners during the spontaneous imitation. The third study validated the experimental measurements with biophysical simulations of pairs of human brains. It also showed the effects of anatomical connectivity on intra-individual neural dynamics and the facilitation of the inter-individual sensorimotor coupling.

Motor unit synchronization in FDI and biceps brachii muscles of strength-trained males

by Brett Fling

Motor unit (MU) synchronization is the simultaneous or near-simultaneous firing of two MUs which occurs more often... more Motor unit (MU) synchronization is the simultaneous or near-simultaneous firing of two MUs which occurs more often than would be expected by chance. The present study sought to investigate the effects of exercise training, muscle group, and force level, by comparing the magnitude of synchronization in the biceps brachii (BB) and first dorsal interosseous (FDI) muscles of untrained and strength-trained college-aged males at two force levels, 30% of maximal voluntary contraction (MVC) and 80% MVC. MU action potentials were recorded directly via an intramuscular needle electrode. The magnitude of synchronization was assessed through computation of k’, E and CIS synchronization indices. All indices were significantly higher in the FDI than in the BB. Greater synchronization was observed in the strength-trained group with CIS, but not with E or k’. Also, synchronization was significantly greater at 80% MVC than at 30% MVC with E, but only moderately greater with CIS and there was no force difference with k’. Synchronization prevalence was found to be greater in the BB (80.1%) than in the FDI (71.5%). Thus, although the effects of training and force level on MU synchronization remain equivocal, there is evidence that these factors affect synchronization magnitude.

Anatomical Connectivity Influences both Intra- and Inter-Brain Synchronizations

by Guillaume Dumas

Dumas G, Chavez M, Nadel J, Martinerie J (2012) PLoS ONE 7(5): e36414. doi:10.1371/journal.pone.0036414

Recent development in diffusion spectrum brain imaging combined to functional simulation has the potential to further... more Recent development in diffusion spectrum brain imaging combined to functional simulation has the potential to further our understanding of how structure and dynamics are intertwined in the human brain. At the intra-individual scale, neurocomputational models have already started to uncover how the human connectome constrains the coordination of brain activity across distributed brain regions. In parallel, at the inter-individual scale, nascent social neuroscience provides a new dynamical vista of the coupling between two embodied cognitive agents. Using EEG hyperscanning to record simultaneously the brain activities of subjects during their ongoing interaction, we have previously demonstrated that behavioral synchrony correlates with the emergence of inter-brain synchronization. However, the functional meaning of such synchronization remains to be specified. Here, we use a biophysical model to quantify to what extent inter-brain synchronizations are related to the anatomical and functional similarity of the two brains in interaction. Pairs of interacting brains were numerically simulated and compared to real data. Results show a potential dynamical property of the human connectome to facilitate inter-individual synchronizations and thus may partly account for our propensity to generate dynamical couplings with others.

A Discussion of Special Relativity

by Galina Weinstein

Five topics: A rigid body does not exist in the special theory of relativity; distant simultaneity defined with... more Five topics: A rigid body does not exist in the special theory of relativity; distant simultaneity defined with respect to a given frame of reference without any reference to synchronized clocks; challenges on Einstein's connection of synchronization and contraction; a theory of relativity without light, composition of relative velocities and space of relative velocities

Decentralised reinforcement learning for energy-efficient scheduling in wireless sensor networks

by Mihail Mihaylov

Mihaylov, M., Le Borgne, Y-A., Tuyls, K. and Nowé, A. (2012) ‘Decentralised reinforcement learning for energy-efficient scheduling in wireless sensor networks’, International Journal of Communication Networks and Distributed Systems, Vol. 9, Nos. 3/4, pp.207–224.

We present a self-organising reinforcement learning (RL) approach for scheduling the wake-up cycles of nodes in a... more We present a self-organising reinforcement learning (RL) approach for scheduling the wake-up cycles of nodes in a wireless sensor network. The approach is fully decentralised, and allows sensor nodes to schedule their active periods based only on their interactions with neighbouring nodes. Compared to standard scheduling mechanisms such as SMAC, the benefits of the proposed approach are twofold. First, the nodes do not need to synchronise explicitly, since synchronisation is achieved by the successful exchange of data messages in the data collection process. Second, the learning process allows nodes competing for the radio channel to desynchronise in such a way that radio interferences and therefore packet collisions are significantly reduced. This results in shorter communication schedules, allowing to not only reduce energy consumption by reducing the wake-up cycles of sensor nodes, but also to decrease the data retrieval latency. We implement this RL approach in the OMNET++ sensor network simulator, and illustrate how sensor nodes arranged in line, mesh and grid topologies autonomously uncover schedules that favour the successful delivery of messages along a routing tree while avoiding interferences.

Synchronization of coupled oscillators

by Paulo Maurício

MsC Thesis

In this work we begin by introducing the Kuramoto model, constructing its solutions in the thermodynamic limit and... more In this work we begin by introducing the Kuramoto model, constructing its solutions in the thermodynamic limit and showing the close connection between statistical physics and
dynamical systems that lead to the main theoretical insights. The systematic study of a finite population of self sustained oscillators began in the first decade of this century.
Unlike most of the papers we have found, we are not interested in the synchronization transition in itself but rather in phase locked patterns and their relation with frequency distribution among oscillators.
The problem of stability, as we have already mentioned, experienced great advances in recent years. In a brief discussion we only address the problem of stability of the simplest solution allowed by the Kuramoto model: the incoherent solution. After that we introduce Chimera states, First noticed by Kuramoto and his colleagues in which the introduction of a non local coupling gives origin to a split in a region with synchronised oscillators and other with asynchronous one.
Then we proceed by exploring the literature and the results with a fnite number of oscillators, model explored with persistence only since mainly 2004. But here we are yet in Kuramoto framework which is abandoned, in a rigorous terminology, when we pursuit structured and not all-to-all coupling. Although we could introduce the same mean models quantities if well defined in each situation, this did not help us in making sense of the
results and is not an help in any analytical work.
In our analysis of a ring of coupled oscillators we construct a space that allows us to relate the stable solutions with the eigenvectors of the laplacian of the graph in which we work.
work.

Synchronization of coupled oscillators is a game

by Huibing Yin

Self-Organizing Synchronicity and Desynchronicity using Reinforcement Learning

by Mihail Mihaylov

M. Mihaylov, Y.-A. Le Borgne, K. Tuyls, and A. Nowé, "Self-Organizing Synchronicity and Desynchronicity using Reinforcement Learning," in Proceedings of the 3rd International Conference on Agents and Artificial Intelligence (ICAART 2011), Rome, Italy, 2011, pp. 94-103.

We present a self-organizing reinforcement learning (RL) approach for coordinating the wake-up cycles of nodes in a... more We present a self-organizing reinforcement learning (RL) approach for coordinating the wake-up cycles of nodes in a wireless sensor network in a decentralized manner. To the best of our knowledge we are the first to demonstrate how global synchronicity and desynchronicity can emerge through local interactions alone without the need of central mediator or any form of explicit coordination. We apply this RL approach to wireless sensor nodes arranged in different topologies and study how agents, starting with a random policy, are able to self-adapt their behavior based only on their interaction with neighboring nodes. Each agent independently learns to which nodes it should synchronize to improve message throughput and at the same with whom to desynchronize in order to reduce communication interference. The obtained results show how simple and computationally bounded sensor nodes are able to coordinate their wake-up cycles in a distributed way in order to improve the global system performance through (de)synchronicity.

Inter-individual performance differences in younger and older adults differentially relate to amplitude modulations and phase stability of oscillations controlling working memory contents.

by Markus Werkle-Bergner

Phase synchronization analysis of event-related potentials in language processing

by Carsten Allefeld

PhD thesis, University of Potsdam, 2004.

The topic of synchronization forms a link between nonlinear dynamics and neuroscience. On the one hand,... more The topic of synchronization forms a link between nonlinear dynamics and neuroscience. On the one hand, neurobiological research has shown that the synchronization of neuronal activity is an essential aspect of the working principle of the brain. On the other hand, recent advances in the physical theory have led to the discovery of the phenomenon of phase synchronization. A method of data analysis that is motivated by this finding - phase synchronization analysis - has already been successfully applied to empirical data.
The present doctoral thesis ties up to these converging lines of research. Its subject are methodical contributions to the further development of phase synchronization analysis, as well as its application to event-related potentials, a form of EEG data that is especially important in the cognitive sciences.
The methodical contributions of this work consist firstly in a number of specialized statistical tests for a difference in the synchronization strength in two different states of a system of two oscillators. Secondly, in regard of the many-channel character of EEG data an approach to multivariate phase synchronization analysis is presented.
For the empirical investigation of neuronal synchronization a classic experiment on language processing was replicated, comparing the effect of a semantic violation in a sentence context with that of the manipulation of physical stimulus properties (font color). Here phase synchronization analysis detects a decrease of global synchronization for the semantic violation as well as an increase for the physical manipulation. In the latter case, by means of the multivariate analysis the global synchronization effect can be traced back to an interaction of symmetrically located brain areas.
The findings presented show that the method of phase synchronization analysis motivated by physics is able to provide a relevant contribution to the investigation of event-related potentials in the cognitive sciences.

Detecting synchronization clusters in multivariate time series via coarse-graining of Markov chains

by Carsten Allefeld

C. Allefeld and S. Bialonski. Detecting synchronization clusters in multivariate time series via coarse-graining of Markov chains. Physical Review E, 76:066207, 2007.
The uploaded version is from arXiv:0707.2479

Synchronization cluster analysis is an approach to the detection of underlying structures in data sets of multivariate... more Synchronization cluster analysis is an approach to the detection of underlying structures in data sets of multivariate time series, starting from a matrix R of bivariate synchronization indices. A previous method utilized the eigenvectors of R for cluster identification, analogous to several recent attempts at group identification using eigenvectors of the correlation matrix. All of these approaches assumed a one-to-one correspondence of dominant eigenvectors and clusters, which has however been shown to be wrong in important cases. We clarify the usefulness of eigenvalue decomposition for synchronization cluster analysis by translating the problem into the language of stochastic processes, and derive an enhanced clustering method harnessing recent insights from the coarse-graining of finite-state Markov processes. We illustrate the operation of our method using a simulated system of coupled Lorenz oscillators, and we demonstrate its superior performance over the previous approach. Finally we investigate the question of robustness of the algorithm against small sample size, which is important with regard to field applications.

Quantifying the dynamics of coupled networks of switches and oscillators

by Elana Fertig

Matthew R Francis and EJ Fertig* (2012) PLoS One, 7:e29497. *Corresponding author

Complex network dynamics have been analyzed with models of systems of coupled switches or systems of coupled... more Complex network dynamics have been analyzed with models of systems of coupled switches or systems of coupled oscillators. However, many complex systems are composed of components with diverse dynamics whose interactions drive the system's evolution. We, therefore, introduce a new modeling framework that describes the dynamics of networks composed of both oscillators and switches. Both oscillator synchronization and switch stability are preserved in these heterogeneous, coupled networks. Furthermore, this model recapitulates the qualitative dynamics for the yeast cell cycle consistent with the hypothesized dynamics resulting from decomposition of the regulatory network into dynamic motifs. Introducing feedback into the cell-cycle network induces qualitative dynamics analogous to limitless replicative potential that is a hallmark of cancer. As a result, the proposed model of switch and oscillator coupling provides the ability to incorporate mechanisms that underlie the synchronized stimulus response ubiquitous in biochemical systems.

Synchronization by Nonlinear Frequency Pulling

by Ron Lifshitz

M.C. Cross, A. Zumdieck, Ron Lifshitz, and J.L. Rogers. Phys. Rev. Lett. 93 (2004) 224101.

We analyze a model for the synchronization of nonlinear oscillators due to reactive coupling and nonlinear frequency... more We analyze a model for the synchronization of nonlinear oscillators due to reactive coupling and nonlinear frequency pulling motivated by the physics of arrays of nanoscale oscillators. We study the model for the mean field case of all-to-all coupling, deriving results for the onset of synchronization as the coupling or nonlinearity increase, and the fully locked state when all the oscillators evolve with the same frequency.

Emergence of synchronous EEG spindles from asynchronous MEG spindles

by Nima Dehghani

Magnetoencephalography demonstrates multiple asynchronous generators during human sleep spindles

by Nima Dehghani

Management of synchronized network activity by highly active neurons

by Vladislav Volman

Towards a second cybernetics model for cognitive systems

by Hans H. Diebner

Chaos, Solitons & Fractals 13/7, 1465-1474, (2002). Coauthors: Axel A. Hoff, Adolf Mathias, Horst Prehn, Marco Rohrbach and Sven Sahle.

A realtime adaptive system for dynamics recognition

by Hans H. Diebner

Chaos, Solitons & Fractals, 13/4, 781-786 (2002). Coauthors: Sven Sahle and Axel A. Hoff

A simulating cognitive system with adaptive capability

by Hans H. Diebner

BioSystems, 64/1-3, 141-147 (2002).

Dedicated to the memory of Michael Conrad, this paper builds on his seminal ideas expressed in his famous book... more Dedicated to the memory of Michael Conrad, this paper builds on his seminal ideas expressed in his famous book Adaptability, as well as in his later works. We investigate a recently published adaptive system for the instantaneous recognition of dynamics with respect to its adaptability to the Lorenz system. The system consists of a pool of internal dynamical elements. These elements are defined through a set of parameter values that encode for a specific dynamics behavior. If the system is now faced with an unknown external dynamics--unknown with respect to the parameter--it is capable not only to recognize the dynamics but also to adapt to the correct dynamics, which in turn leads to a simulation capability. The system impressively quickly follows the sudden qualitative changes of the external dynamics. The adaptation works even quicker when the correct dynamics are already represented within the internal pool. This leads to the idea of memorizing the represented dynamics within the pool, whereby the elements that correspond to rarely externally presented dynamics can be given free for the adaptation and memorization of more frequently presented dynamics.