Word differences in news media of lower and higher peace countries revealed by natural language processing and machine learning.

Language is both a cause and a consequence of the social processes that lead to conflict or peace. "Hate speech" can mobilize violence and destruction. What are the characteristics of "peace speech" that reflect and support the social processes that maintain peace? This study used existing peace indices, machine learning, and on-line, news media sources to identify the words most associated with lower-peace versus higher-peace countries. As each peace index measures different social properties, they can have different values for the same country. There is however greater consensus with these indices for the countries that are at the extremes of lower-peace and higher-peace. Therefore, a data driven approach was used to find the words most important in distinguishing lower-peace and higher-peace countries. Rather than assuming a theoretical framework that predicts which words are more likely in lower-peace and higher-peace countries, and then searching for those words in news media, in this study, natural language processing and machine learning were used to identify the words that most accurately classified a country as lower-peace or higher-peace. Once the machine learning model was trained on the word frequencies from the extreme lower-peace and higher-peace countries, that model was also used to compute a quantitative peace index for these and other intermediate-peace countries. The model successfully yielded a quantitative peace index for intermediate-peace countries that was in between that of the lower-peace and higher-peace, even though they were not in the training set. This study demonstrates how natural language processing and machine learning can help to generate new quantitative measures of social systems, which in this study, were linguistic differences resulting in a quantitative index of peace for countries at different levels of peacefulness.

Conflict Dynamics in Scale-Free Networks with Degree Correlations and Hierarchical Structure.

We present a study of the dynamic interactions between actors located on complex networks with scale-free and hierarchical scale-free topologies with assortative mixing, that is, correlations between the degree distributions of the actors. The actor's state evolves according to a model that considers its previous state, the inertia to change, and the influence of its neighborhood. We show that the time evolution of the system depends on the percentage of cooperative or competitive interactions. For scale-free networks, we find that the dispersion between actors is higher when all interactions are either cooperative or competitive, while a balanced presence of interactions leads to a lower separation. Moreover, positive assortative mixing leads to greater divergence between the states, while negative assortative mixing reduces this dispersion. We also find that hierarchical scale-free networks have both similarities and differences when compared with scale-free networks. Hierarchical scale-free networks, like scale-free networks, show the least divergence for an equal mix of cooperative and competitive interactions between actors. On the other hand, hierarchical scale-free networks, unlike scale-free networks, show much greater divergence when dominated by cooperative rather than competitive actors, and while the formation of a rich club (adding links between hubs) with cooperative interactions leads to greater divergence, the divergence is much less when they are fully competitive. Our findings highlight the importance of the topology where the interaction dynamics take place, and the fact that a balanced presence of cooperators and competitors makes the system more cohesive, compared to the case where one strategy dominates.

How to live in peace? Mapping the science of sustaining peace: A progress report.

Despite good faith attempts by countless citizens, civil society, governments, and the international community, living in a sustainably peaceful community continues to be an elusive dream in much of our world. Among the challenges to sustaining peace is the fact that few scholars have studied enduringly peaceful societies, or have examined only narrow aspects of them, leaving our understanding of the necessary conditions, processes and policies fragmented, and deficient. This article provides a work-in-progress overview of a multidisciplinary, multimethod initiative, which aims to provide a holistic, evidence-based understanding of how peace can be sustained in societies. The Sustaining Peace Project, launched in 2014, uses complexity science as an integrative platform for synthesizing knowledge across disciplines, sectors and communities. This article introduces the multiple components of the project and shares preliminary findings. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Recurrence Networks in Natural Languages.

We present a study of natural language using the recurrence network method. In our approach, the repetition of patterns of characters is evaluated without considering the word structure in written texts from different natural languages. Our dataset comprises 85 ebookseBooks written in 17 different European languages. The similarity between patterns of length m is determined by the Hamming distance and a value r is considered to define a matching between two patterns, i.e., a repetition is defined if the Hamming distance is equal or less than the given threshold value r. In this way, we calculate the adjacency matrix, where a connection between two nodes exists when a matching occurs. Next, the recurrence network is constructed for the texts and some representative network metrics are calculated. Our results show that average values of network density, clustering, and assortativity are larger than their corresponding shuffled versions, while for metrics like such as closeness, both original and random sequences exhibit similar values. Moreover, our calculations show similar average values for density among languages which that belong to the same linguistic family. In addition, the application of a linear discriminant analysis leads to well-separated clusters of family languages based on based on the network-density properties. Finally, we discuss our results in the context of the general characteristics of written texts.

The dynamic consequences of cooperation and competition in small-world networks.

We present a study of the social dynamics among cooperative and competitive actors interacting on a complex network that has a small-world topology. In this model, the state of each actor depends on its previous state in time, its inertia to change, and the influence of its neighboring actors. Using numerical simulations, we determine how the distribution of final states of the actors and measures of the distances between the values of the actors at local and global levels, depend on the number of cooperative to competitive actors and the connectivity of the actors in the network. We find that similar numbers of cooperative and competitive actors yield the lowest values for the local and global measures of the distances between the values of the actors. On the other hand, when the number of either cooperative or competitive actors dominate the system, then the divergence is largest between the values of the actors. Our findings make new testable predictions on how the dynamics of a conflict depends on the strategies chosen by groups of actors and also have implications for the evolution of behaviors.

Behavioral and emotional dynamics of two people struggling to reach consensus about a topic on which they disagree.

We studied the behavioral and emotional dynamics displayed by two people trying to resolve a conflict. 59 groups of two people were asked to talk for 20 minutes to try to reach a consensus about a topic on which they disagreed. The topics were abortion, affirmative action, death penalty, and euthanasia. Behavior data were determined from audio recordings where each second of the conversation was assessed as proself, neutral, or prosocial. We determined the probability density function of the durations of time spent in each behavioral state. These durations were well fit by a stretched exponential distribution, [Formula: see text] with an exponent, [Formula: see text], of approximately 0.3. This indicates that the switching between behavioral states is not a random Markov process, but one where the probability to switch behavioral states decreases with the time already spent in that behavioral state. The degree of this "memory" was stronger in those groups who did not reach a consensus and where the conflict grew more destructive than in those that did. Emotion data were measured by having each person listen to the audio recording and moving a computer mouse to recall their negative or positive emotional valence at each moment in the conversation. We used the Hurst rescaled range analysis and power spectrum to determine the correlations in the fluctuations of the emotional valence. The emotional valence was well described by a random walk whose increments were uncorrelated. Thus, the behavior data demonstrated a "memory" of the duration already spent in a behavioral state while the emotion data fluctuated as a random walk whose steps did not have a "memory" of previous steps. This work demonstrates that statistical analysis, more commonly used to analyze physical phenomena, can also shed interesting light on the dynamics of processes in social psychology and conflict management.

A mathematical model of psychotherapy: an investigation using dynamic non-linear equations to model the therapeutic relationship.

Mathematical models, such as the one developed by Gottman et al. (1998, 2000, 2002) to understand the interaction between husbands and wives, can provide novel insights into the dynamics of the therapeutic relationship. A set of nonlinear equations were used to model the changing emotional state of a therapist and client. The results suggest: (1) The person that is most responsive to the other achieves the most positive state, (2) the emotional state of the client oscillates before reaching its final state, (3) therapy is least successful when the therapist starts from a negative state, and (4) there is an inverse relationship between models that change only the influence parameter and models that change only the inertia parameter, creating a series of four basic models to work with. These theoretical models require further, empirical investigation to test the derived parameters. If validated, or revised based on observations of therapist-client relationships in development, they could provide specific direction in creating successful therapeutic relationships for training clinicians and those already in practice.

Mathematical model of the dynamics of psychotherapy.

The success of psychotherapy depends on the nature of the therapeutic relationship between a therapist and a client. We use dynamical systems theory to model the dynamics of the emotional interaction between a therapist and client. We determine how the therapeutic endpoint and the dynamics of getting there depend on the parameters of the model. Previously Gottman et al. used a very similar approach (physical-sciences paradigm) for modeling and making predictions about husband-wife relationships. Given that this novel approach shed light on the dyadic interaction between couples, we have applied it to the study of the relationship between therapist and client. The results of our computations provide a new perspective on the therapeutic relationship and a number of useful insights. Our goal is to create a model that is capable of making solid predictions about the dynamics of psychotherapy with the ultimate intention of using it to better train therapists.

Dynamic stabilization in the PU1-GATA1 circuit using a model with time-dependent kinetic change.

The PU.1 and GATA1 genes play an important role in the differentiation of blood stem cells. The protein levels expressed by these genes are thought to be regulated by a self-excitatory feedback loop for each gene and a cross-inhibitory feedback loop between the two genes. A mathematical model that captures the dynamical interaction between these two genes reveals that constant levels of self-excitation and cross-inhibition allow the most self-exciting or cross-inhibiting gene to dominate the system. However, since biological systems rarely exist in an unchanging equilibrium, we modeled this gene circuit using discrete time-dependent changes in the parameters in lieu of steady state parameters. These time-dependent parameters lead to new phenomena, including the development of new limit cycles and basins of attraction. These phenomena are not present in models using constant parameter values. Our findings suggest that even small perturbations in the PU.1 and GATA1 feedback loops may substantially alter the gene expression and therefore the cell phenotype. These time-dependent parameter models may also have implications for other gene systems and provide new ways to understand the mechanisms of cellular differentiation.

Patterns of subnet usage reveal distinct scales of regulation in the transcriptional regulatory network of Escherichia coli.

The set of regulatory interactions between genes, mediated by transcription factors, forms a species' transcriptional regulatory network (TRN). By comparing this network with measured gene expression data, one can identify functional properties of the TRN and gain general insight into transcriptional control. We define the subnet of a node as the subgraph consisting of all nodes topologically downstream of the node, including itself. Using a large set of microarray expression data of the bacterium Escherichia coli, we find that the gene expression in different subnets exhibits a structured pattern in response to environmental changes and genotypic mutation. Subnets with fewer changes in their expression pattern have a higher fraction of feed-forward loop motifs and a lower fraction of small RNA targets within them. Our study implies that the TRN consists of several scales of regulatory organization: (1) subnets with more varying gene expression controlled by both transcription factors and post-transcriptional RNA regulation and (2) subnets with less varying gene expression having more feed-forward loops and less post-transcriptional RNA regulation.

Dynamical response of multi-patch, flux-based models to the input of infected people: Epidemic response to initiated events.

The time course of an epidemic can be modeled using the differential equations that describe the spread of disease and by dividing people into "patches" of different sizes with the migration of people between these patches. We used these multi-patch, flux-based models to determine how the time course of infected and susceptible populations depends on the disease parameters, the geometry of the migrations between the patches, and the addition of infected people into a patch. We found that there are significantly longer lived transients and additional "ancillary" epidemics when the reproductive rate R is closer to 1, as would be typical of SARS (Severe Acute Respiratory Syndrome) and bird flu, than when R is closer to 10, as would be typical of measles. In addition we show, both analytical and numerical, how the time delay between the injection of infected people into a patch and the corresponding initial epidemic that it produces depends on R.

Developing combinatorial multi-component therapies (CMCT) of drugs that are more specific and have fewer side effects than traditional one drug therapies.

Drugs designed for a specific target are always found to have multiple effects. Rather than hope that one bullet can be designed to hit only one target, nonlinear interactions across genomic and proteomic networks could be used to design Combinatorial Multi-Component Therapies (CMCT) that are more targeted with fewer side effects. We show here how computational approaches can be used to predict which combinations of drugs would produce the best effects. Using a nonlinear model of how the output effect depends on multiple input drugs, we show that an artificial neural network can accurately predict the effect of all 215 = 32,768 combinations of drug inputs using only the limited data of the output effect of the drugs presented one-at-a-time and pairs-at-a-time.

Information flow dynamics and timing patterns in the arrival of email viruses.

Analysis of the timing of the arrival of email viruses at different computers provides a way of probing the structural and dynamical properties of the Internet. We found that the intervals t between the arrival of four different strains of email viruses have a power law distribution proportional to t(-d), where 1.5

Temporal patterns of atrial arrhythmia recurrences in patients with implantable defibrillators: implications for assessing antiarrhythmic therapies.

INTRODUCTION: The statistical measures commonly used to assess therapies for recurrent atrial arrhythmias (such as time to first recurrence) often assume a uniformly random pattern of arrhythmic events over time. However, the true temporal pattern of atrial arrhythmia recurrences is unknown. The aim of this study was to use linear and nonlinear analyses to characterize the temporal pattern of atrial arrhythmia recurrences in patients with implantable cardioverter defibrillators. METHODS AND RESULTS: The time and date of atrial tachyarrhythmias recorded in 65 patients with combined atrial and ventricular defibrillators were used to construct a probability density function (PDF) and a model of a Poisson distribution of arrhythmic events for each patient. Average patient age was 66 +/- 10 years and follow-up was 7.8 +/- 4.8 months. A total of 10,759 episodes of atrial tachyarrhythmias were analyzed (range 43 to 618 episodes per patient). The PDF fit a power law distribution for all 65 patients, with an average r2 = 0.89 +/- 0.08. The PDF distribution differed significantly from the model Poisson distribution in 47 of 65 patients (P = 0.0002). Differences from the Poisson distribution were noted for patients both taking (30/43 patients; P < or = 0.015) and not taking (17/22 patients; P < or = 0.017) antiarrhythmic drugs. Median time between atrial arrhythmia detections for all 65 patients was 10.8 minutes. CONCLUSION: In implantable cardioverter defibrillator patients, the temporal pattern of frequent recurrences of atrial tachyarrhythmias usually is characterized by a power law distribution. The unique statistical properties of this type of distribution should be considered in designing outcome measures for treatment of atrial tachyarrhythmias.

Correlation in stimulated respiratory neural noise.

Noise in spontaneous respiratory neural activity of the neonatal rat isolated brainstem-spinal cord preparation stimulated with acetylcholine (ACh) exhibits positive correlation. Neural activity from the C4 (phrenic) ventral spinal rootlet, integrated and corrected for slowly changing trend, is interpreted as a fractal record in time by rescaled range, relative dispersional, and power spectral analyses. The Hurst exponent H measured from time series of 64 consecutive signal levels recorded at 2 s intervals during perfusion of the preparation with artificial cerebrospinal fluid containing ACh at concentrations 62.5 to 1000 &mgr;M increases to a maximum of 0.875+/-0.087 (SD) at 250 &mgr;M ACh and decreases with higher ACh concentration. Corrections for bias in measurement of H were made using two different kinds of simulated fractional Gaussian noise. Within limits of experimental procedure and short data series, we conclude that in the presence of added ACh of concentration 250 to 500 &mgr;M, noise which occurs in spontaneous respiratory-related neural activity in the isolated brainstem-spinal cord preparation observed at uniform time intervals exhibits positive correlation. (c) 1995 American Institute of Physics.