Abstracts

ComplexNetworks07

Tel Aviv University October 24-25, 2007

Organizers

The Italian embassy in Israel and the Tel Aviv University

Scientific coordinators:

Prof. Vito Latora and Prof. Eshel Ben Jacob

latora@ct.infn.it              eshelbj@gmail.com

------------

Conference Email

complexnetworks07@gmail.com

Green House, Tel Aviv University, Israel

October 24-25, 2007

To the Green house web site:

http://www.gv.co.il/intro.html


--------------------------------------------------

Uri Alon

Simplicity in System Biology

Abstract:

Biological systems seem to be simpler then they could have been. The talk will highlight one of these simplifying features: Biological regulation networks appear to be built of a small set of recurring interaction patterns, called network motifs. Each network motif is able to carry out defined dynamical functions, as demonstrated by highresolution dynamical experiments on living cells. Evolution seems to have converged ("rediscovered") the same motifs again and again in different systems and organisms. We will see how complicated biological networks made of interconnected motifs can be understood in terms of the dynamics of individual network motifs. We will discuss the current theoretical and experimental challenges that face us in the understanding the dynamics of biological networks.

--------------------------------------------------

Irun Cohen  

The immune system computes the state of the body

Abstract:

The immune system maintains the body by computing its current state; this computation is effectively expressed when the immune system deploys and dynamically adjusts a multi-agent inflammatory process to the varying needs of the body. The immune system carries out this computation through the interactions of distributed cells and molecules without the programming and logic characteristic of human-made computers. 

 I shall present two examples of the many system elements involved in immune computation: the architecture of cytokine mediated cell connectivity and the immunological homunculus– a dynamic representation of body biomarkers inherent in one’s repertoire of auto-antibodies.

--------------------------------------------------

Yael Hanein

Engineering neural networks

Abstract;

Unraveling the underlying mechanisms involved in the activity of neural network systems have been, and still remain, one of the fundamental challenges in modern scientific research. Real neural networks consist of enormously elaborate networks made of neurons and glia cells interconnected through axons and dendrites. The processing of information in neuronal networks relies on their ability to generate temporal sequences of action potentials of the individual neurons and to modulate the level and patterns of activity at the network level by control of the cells excitability and of the synaptic connections. It is widely accepted that these networks are specially linked in a manner suited to their collective task performance and a strong relation exists between the network form and function.

Despite the above assertion, the inherent complexity of real neural networks hampers systematic investigation into the link between a network structure and its activity. To facilitate such an investigation we have devised a set of novel neural network engineering tools suited for studying function-from relationship in neural networks. In particular, we have developed a carbon nanotube based scheme to engineer neural networks with pre-defined geometries and to facilitate very high resolution electrical recordings and stimulations from these systems. Using these new tools we have been able to study the organization of the neural system into electrically viable compactly wired network. We argue that such networks could facilitate the development and verification of the neural network models.

--------------------------------------------------

Shlomo Havlin

Immunization methods of complex networks

Abstract:

An important question in epidemics is how to minimize the number of immunized nodes (people, computers etc) and still stop the spreading of epidemics. I will compare several efficient immunization methods recently developed with a new method that seems to be most efficient when tested on several model and real networks.

------------------------------------------------------------

Martin Kupiec

A dissection of the genetic network controlling telomere control

Abstract:

Telomeres protect the ends of eukaryotic chromosomes from being treated as broken DNA molecules. In multi-cellular organisms, telomere shortening acts as a tumor suppressive mechanism. Replenishing telomeres is one of the few essential steps that a normal human fibroblast cell must take on its way to become malignant. Length regulation is central to telomere function, and is determined by equilibrium between processes that elongate telomeres (through the activity of telomerase) and those that shorten telomeres (e.g., nucleases).

We carried out a systematic screen for telomere-length alterations using 4,800 haploid deletion mutants of the yeast Saccharomyces cerevisiae, and identified >170 genes that affect telomere length (TLM genes). In two-thirds of the identified mutants, short telomeres were observed; whereas in one-third, telomeres were lengthened. The genes identified are very diverse in their functions, but certain categories, including DNA and RNA metabolism, chromatin modification, and vacuolar traffic, are overrepresented.

The large number of telomere-afecting genes uncovered in our screen represents a challenge. Our work integrates two complementary approaches: computational and experimental. As a first stage in deciphering the various pathways that regulate telomere length, we have subjected our panel of tlm mutants to an epistasis analysis aimed at identifying genetic interactions. The mutants were crossed to create double mutants. The telomere lengths of these strains are then compared to that of the single mutant parents. We aim at modeling the genetic interactions, and reconstructing signaling pathways that connect TLM genes to the telomerase machinery, thereby shedding light on their mechanisms of action. To this end, we are combining available protein-protein, protein-DNA and genetic interaction data to construct a combined network of interactions in yeast and map the TLM network.

--------------------------------------------------

Shimon Marom

Representation in neuronal networks

Abstract:

To be functional, biological networks must be capable of representing spatiotemporal features environmental object .his is not a simple task given their intrinsic dynamics and structural complexity .Measurements from large-scale recurrent networks of biological neurons provide a unique opportunity to observe and define issues pertaining to representation of environmental objects.

I will focus on principles underlying invariance of representations to time warping of network  responses, as well as on network mechanisms for novelty detection.

  --------------------------------------------------

 Yitzhak (Zachi) Pilpel

Coping with genetic and non-genetic perturbations

Abstract:

To provide robust fulfillment of regulatory gene expression programs cells rely on genetic circuits that allow them to respond more efficiently to their environment, and to control the internal stochastic fluctuations. We study how these goals are fulfilled by regulating genomic redundancies that offer genetic backup against external and internal variability.

I will discuss our explorations of regulatory systems probabilistically predicts environmental changes before they actually occur. I will introduce our investigations into the structure of redundancy-based circuits that provide cellular solution to genetic and non-genetic fluctuations.

--------------------------------------------------

Ron Meir

Optimal dynamic state estimation by neural networks based on recursive spike train decoding.

Abstract:

It is becoming increasingly evident that organisms acting in uncertain dynamical environments often employ exact or approximate Bayesian statistical calculations in order to continuously estimate the environmental state, integrate information from multiple sensory modalities, form predictions and choose actions.

What is less clear is how these putative computations are implemented by cortical neural networks. An additional level of complexity is introduced because these networks observe the world through spike trains received from primary sensory afferents, rather than directly. A recent line of research has described mechanisms by which such computations can be implemented using a network of neurons whose activity directly represents a probability distribution across the possible "world states''.

Much of this work, however, uses various approximations, which severely restrict the domain of applicability of these implementations. Here we make use of rigorous mathematical results from the theory of continuous time point process filtering, and show how optimal real-time state estimation and prediction may be implemented in a general setting using linear neural networks. We demonstrate the applicability of the approach with several examples, and relate the required network properties to the statistical nature of the environment, thereby quantifying the compatibility of a given network with its environment.

--------------------------------------------------

David Horn
Enzymatic Profiles of Metagenomes
Abstract:
Genomic studies of microbial communities have become an important research trend, displaying results that turn out to be characteristic to the environments in which these communities are found. Thus networks of organisms are represented by their overall genetic content. Metagenomics is expected to develop into metaproteomics, providing a direct measure of
the proteins available within the microbial community. Present studies provide lists of genes, i.e. putative proteins. We have developed a method for data mining of enzymes within such lists, based on our recent discovery that Specific Peptides (SPs) located on enzyme sequences, can predict the biochemical function of an enzyme (known as its EC number) with high accuracy. Applying this method to metagenomes, we find large numbers of putative enzymes, allowing us to construct an Enzymatic Profile for the metagenome in question, consisting of the numbers of different genes present in each
EC number. This may be regarded as a step toward capturing relevant metaproteomic information.

 
--------------------------------------------------

Yuval Shavit

Studying the Internet complex structure with DIMES

Abstract:

The Internet is possibly the largest manmade complex system.  Due to its distributed management and routing protocol, the only way to map its topology is by having measurement presence in almost every corner of the Internet. Managing thousands of measurement boxes is impractical, thus, we suggest instead a light-weight software measurement agent to be downloaded by volunteers around the world. The DIMES agent can be executed on every PC (and in the future even smaller devices, like PDAs) and enables us to map the Internet and track its evolution in time in several levels of granularity from the fine router level to the coarse Autonomous System (AS) level.  Currently, DIMES has over 12,000 installations in almost 100 nations around the world, which produce over 5 million measurements a day. Our database has over to 2.5 Billion measurements.

In the talk, I’ll discuss the rational behind DIMES design and present some new results obtained from the data DIMES collected.

--------------------------------------------------

Sorin Solomon

Distinguishing the Tree of Knowledge from the Network Forrest

Abstract:

Currently we are facing an explosive development and discovery of vast self-organized information processing systems such as WWW, Wikipedia, peer-to-peer, biological and cognitive networks. These newly emerged entities have started to transformed science from a world thirsty of sheer information into a world desperate for creating meaning out of it. Sometimes this feels like searching for an infinitesimal needle in a vast haystack.

Yet the solution may come from a very close and familiar place. I will illustrate how understanding the fundamental principles of complex networks might play crucial role in the effort to extract meaning from information.

References:
Self-Emergence of Knowledge Trees: Extraction of the Wikipedia Hidden Hierarchies ;  Lev Muchnik, Royi Itzhak, Sorin Solomon, and Yoram Louzoun; http://shum.huji.ac.il/~sorin/ccs/Self_Emergence_of_Knowledge_Trees.doc 

Distributive immunization of networks against viruses using the 'honey-pot' architecture Goldenberg J., Shavitt  Y., Shir  E.& Solomon  S. Nature (/Physics) 1. 184 - 188 (2005). http://shum.cc.huji.ac.il/~sorin/ccs/isi/goldenberg%20antiviral%20Nat%20Phys%2012-05.pdf 

Copying nodes versus editing links: the source of the difference between genetic regulatory networks and the WWW, Yoram. Louzoun, Lev Muchnick and Sorin Solomon Bioinformatics (Oxford University Press 2006) 22(5):581-588;
http://shum.cc.huji.ac.il/~sorin/ccs/bio-info-yoram.pdf

 

---------------------------------------------------------

Ginestra Bianconi (ICTP, Trieste)

Entropy of randomized network ensembles 

Abstract:

Randomized network ensembles are the null models of real networks and are extensively used to compare a real system to a null hypothesis. In this talk we will present randomized network ensembles with the same degree distribution, the same degree-correlations or the same community structure of any given real network. We characterize these randomized network ensembles by their entropy, i.e. the normalized logarithm of the total number of networks which are part of these ensembles. We estimate the entropy of randomized ensembles starting from a large set of real directed and undirected networks. We propose entropy as an indicator to assess the role of each structural feature in a given real network. We observe that the ensembles with fixed scale-free degree distribution have smaller entropy than the ensembles with homogeneous degree distribution indicating a higher level of order in scale-free networks.

---------------------------------------------------------

Stefano Boccaletti

Links and Synch(ronization) 

Abstract:

---------------------------------------------------------

Vittoria Colizza  (ISI Foundation, Torino)

Computational approach to the forecast of emerging diseases

Abstract:

Epidemic spread is inevitably entangled with human behavior, social contacts, and population flows among different geographical regions. The analysis of datasets which trace the activities and interactions of individuals, social patterns, and travel fluxes, have unveiled the presence of connectivity patterns characterized by complex features encoded in large-scale heterogeneities and unbounded statistical fluctuations. Presenting a large-scale computational approach for the study of global epidemics, we study the impact of these features on the modeled spreading pattern and analyze the effect of complex real world transportation networks in realistic models for the forecast of the global spread of emerging diseases. Case studies for risk assessment analysis and comparison with historical epidemics are presented.

--------------------------------------------------------- 

Luigi Fortuna  (DEES, University of Catania)

Emerging properties of dynamical networks 

Abstract:

We focus on synchronization of dynamical networks of multiplexed chaotic systems with smooth nonlinearities. The strategy to establish if such synchronization is achievable is based on the Master Stability Function approach and on the optimization of the coupling parameters. With this  approach we are able to show that systems formed by three independent
canonical chaotic circuits (i.e. a Lorenz system, a Rossler oscillator and a Chua's circuit) can be synchronized through a unique scalar signal

---------------------------------------------------------

Santo Fortunato (ISI Foundation, Torino)

Three puzzles in community detection

Abstract:
Detecting communities in networks is crucial to uncover relationships between nodes of complex networks. In spite of the sizeable literature on the topic, the elements of the problem are not clearly defined. I discuss the three main aspects of the problem, which in some respect still pose a challenge to scholars: the definition(s) of community, the evaluation of partitions and the issue of hierarchies.

-----------------------------------------------------

Andrea Giansanti  (Department of Physics , University Roma La Sapienza)

Inner hydrophobic networks in proteins

Abstract:

The metaphor of the network is here used to represent the logic of a protein molecule, structural and functional, encoded in its sequence. In general, it is assumed that hydrophobic effects drive a protein into its native state. We use here hydrophobicity scales to reconstruct in various ways a network, using the information contained in protein sequences. Various quantities, derived from the language of networks, can be used to classify proteins and to predict effects, induced by point mutations. In particular, we discuss how the network metaphor can shed some light on the problem of the stability and aggregation properties of acyl-phossphateses and of a class of viral neuraminidases.

---------------------------------------------------------

Mattia Frasca (DEES , University of Catania)

Emerging properties of dynamical networks of mobile agents

Abstract:

 Complex systems made of mobile agents are recently gaining increasing interest,  since they can be used to model several complex phenomena arising in social and engineered systems. Epidemic spreading models based on agents traveling along
the nodes of a complex networks and distributed control of teams of robots are only two examples. When mobile agents are let to communicate with other agents within a given neighborhood, the interaction network that can be used to characterize
the system is a dynamical one, i.e., a network in which links do evolve in time.  In this work, we study different aspects of systems of mobile agents under the perspective of the analysis of the underlying dynamical networks. In particular,  we study how local interaction rules can be used to obtain a form of coordinated behavior and how motion can influence epidemic spreading in such systems. We link the emerging behavior to the properties of the underlying interaction network.

 ---------------------------------------------------------

Jesus Gomez-Gardenes (Scuola Superiore di Catania)

Scale-Free topologies and Excitatory-Inhibitory interactions

Abstract:

One of the current theoretical challenges to the explanatory powers of Evolutionary Theory is the understanding of the observed evolutionary survival of cooperative behavior when selfish actions provide higher fitness (reproductive success). In unstructured populations )social panmixia) natural selection drives cooperative strategy to extinction. However, when individuals are allowed to interact only with their neighbors, specified by a graph of social contacts cooperation promoting mechanisms (known as lattice reciprocity (offer to cooperation the opportunity of evolutionary survival. Recent numerical works on the evolution of Prisoner's Dilemma in complex networks settings have revealed that graph heterogeneity enhances dramatically the lattice reciprocity. Here we show that, when far from social panmixia, i.e. in highly heterogeneous populations, the fixation of a strategy in the whole population is in general an impossible event, for there is an asymptotic partition of the population in three subsets, one of which experiences cycles of invasion by the competing strategies, while each strategy reaches respectively fixation at each of the two other subsets. We show how this partition correlates with the partition into connectivity classes and characterize the temporal fluctuations, unveiling the mechanisms stabilizing cooperation in macroscopic scale-free structured societies.

---------------------------------------------------------

 Matteo Lo Presti (STMicroelectronics)

STMicroelectronics for Embedded System

Abstract:

We introduce examples of STMicroelectronics activities on the Embedded Systems: energy saving in home appliances, efficient solutions in factory automation, new sensors for medical image diagnosis, high scale integration and remote monitoring on cardiac applications, new solutions for industrial and service robotics. One of the Main Product Group of STMicroelectronics is the Industrial & Multisegment Sector, which maintains the highest ranking in Power Conversion Market (iSupply 2006).  A team of IMS, Systems Lab, is active on exploitation of know-how on ST products and on solutions to the all the industrial customers.  STMicroelectronics is a global independent semiconductor company, leader in developing and delivering semiconductor solutions across the spectrum of microelectronics applications. ST is one of the world's largest semiconductor companies. In 2006, ST's net revenues were US$9.85 billion and net earnings were US$782 million

---------------------------------------------------------

Rosario Nunzio Mantegna (Department of Physics, University of Palermo)

Correlation based networks in finance 

Abstract:

We review some results our research group has obtained in the characterization and quantitative description of correlation based networks observed in finance. We discuss the most important networks used so far which include the minimum spanning tree [1] and the planar maximally filtered graph [2]. We also briefly discuss the recently introduced average linkage spanning tree. We also discuss about the nature and stability of correlation based networks and we present a method based on bootstrap able to quantify the statistical reliability of links detected in the considered network [3]. We also show how to quantify the goodness of the filtering procedures associated with the correlation based network procedures by using the Kullback-Leibler distance as a measure of the information filtered from multivariate data [4].

 

[1] R. N. Mantegna, Hierarchical Structure in Financial Markets, Eur. Phys. J. B 11, 193-197 (1999).

[2] M. Tumminello, T. Aste, T. Di Matteo, R. N. Mantegna, A tool for filtering information in complex systems , Proc Natl. Acad. Sci. U.S.A 102, 10421-10426 (2005). [3] M. Tumminello, C. Coronnello, F. Lillo, S. Miccichè, R. N. Mantegna, Spanning Trees and bootstrap reliability estimation in correlation based networks, International Journal of Bifurcation and Chaos, 17  2319-2329 (July 2007).

[4] M. Tumminello, F. Lillo, R. N. Mantegna, Kullback-Leibler distance as a measure of the information filtered from multivariate data, Physical Review E (in press 2007)

---------------------------------------------------------

Massimo Marchiori (Department of Computer Science , University of Padova +W3Consortium)

Social Search Engines

Abstract:

On the one hand, the web landscape has undergone massive changes in the last years. On the other hand, search engine technology hasn't quite kept the same pace. In this talk we look at the current scenarios, and argue how social flows can be used to make up for a better generation of search engines. We consider how society and technological progress somehow changed the rules of the game, introducing good but also bad components and see how this situation could be modeled by search engines. Along this line of thinking, we show how the real components of interest are not just web pages, but flows of information of any kind, that need to be merged: this opens up for a wide range of improvements and far-looking developments, towards a new horizon of social search.

---------------------------------------------------------

Remo Pareschi  (Department of Computer Science , University of  Molise)

Networks as Interpretation of Networks 

Abstract:

This is a  contribution to the study of networks from the point of view of their interpretation and representation. In particular, we try to cope with the fact that dynamic networks, which are the type of networks that can be used to model relevant phenomena in nature and society, are characterized by implicit roles and models, such as hubs and small worlds.
The point here is whether real networks with these characteristics can be represented through more abstract networks where we do away from the single individual nodes and replace them with node types which make explicit the role relationships of the original network. This could be particularly useful if we want to reason about the general behavior of very large dynamic networks in structural terms in addition to the insight provided by their statistical properties. But this approach allows us also to go the other way around, by designing abstract dynamic networks with the desired characteristics and then making them concrete by populating them with individual nodes and letting them evolve according to their structural properties. Or we might want to model the situation where two networks come into contact and effectively compose into a larger network, as it may happen, for instance, when two enterprises are merged and the underlying communities of people become a single one; the resulting community may then be itself characterized by totally new emergent properties but nevertheless inherits initially also the properties of the parent communities, an aspect which must be suitably accounted for. In order to provide an implementation of this approach we introduce a representation framework based on logic, in its modern interpretation as the study of concurrent networks of proof objects. This theoretical background comes as particularly handy in dealing with aspects of network composition, as it translates them directly in the different possibilities of connecting networks through  varieties of logical connectives.


---------------------------------------------------------

Vittorio Rosato (ENEA, Roma + Ylichron)

Networks as Interpretation of Networks 

Abstract:

Infrastructures are excellent metaphores of Complex Systems: they are large assembly of technological elements, born and grown in  unsupervised regime, which must sustain complex flows of different matters (data, electrical power, vehicles etc.) managed by complex rules, whose behavior affects life and wealth of modern societies.  As complex systems,  critical infrastructures also display effects which are seldom predictible on the basis of their normal behavior, particularly during their pathological states. As a further element of complexity and a major source of technological problems, critical infrastructures are also strongly interdepedent, as the operability of one of them depends on the correct operability of the others. Complexity science may help technology to sustain the intriguing behavior of these objects. Indeed, much has been done to correlate technological vulnerability with topological properties of the networks representing their structures.  However, the behavior of an infrastructure cannot be totally decoupled by the transport mechanisms
acting on them. Complexity science models should thus be focussed at reproducing both structure and transport on critical infrastructures. The talk will review some results obtained in this field, with a specific focus on the electrical power transmission networks and the Internet.