0:00:04i mean value engineering or not
0:00:14zero
0:00:16the task and complex then it explain emergent macroscopic properties from the interactions of microscopic
0:00:22constituents
0:00:24considered by calibrating it would be most fascinating to understand how our ability to process
0:00:29story information images from the interaction of individual neurons
0:00:33moreover it's kind of knowledge might enable us one day to design a technologies with
0:00:38similar image and capacity
0:00:41when it comes to explain why engineering units
0:00:43one method of choice is to describe the complex system a dynamical network
0:00:48the network to the relation between emergent behaviour and exactly unit translate to the relation
0:00:54between collected dynamics structure
0:00:57in many cases explain the dynamics of phonetic required to consider the global structure however
0:01:03some of the greatest clusters in which finds have reviewed that already local structural properties
0:01:08such as the nodes degree may have to think implications for the local dynamics
0:01:13recently that the scale probably come and focus of network signs
0:01:17coding questions are can considering a multiscale structure such as a single row already and
0:01:22all emotions are properties of the network as a whole
0:01:25can we extract implications independent of embedding wing and which has a structure should be
0:01:31considered
0:01:32in our paper we present an analytical framework which indicates that the answers i
0:01:39and in any case this metric
0:01:42our approach you don't and structure we can be studied in terms of matrix vector
0:01:47okay so that i mean relevant matrix that we matrix was actually contains information about
0:01:52the system's going to behavior
0:01:54okay so the structure the relevant matrix the adjacency matrix spectrum contains information about topological
0:02:00properties
0:02:02recently discovered i think are encoded
0:02:04spectral adjacency matrices have an interesting property
0:02:08not contain several databases symmetry i just adjacency matrix well localised on the top graph
0:02:14sense that all entries corresponding nodes of type of crowd the danish
0:02:20in our paper we shall estimator makes no i'm of the adjacency matrix to look
0:02:26like i was trickling matrix
0:02:28the localized modes i sensitive to the apology on hyperspectral
0:02:32assumption that local authorities at green and contains some matrix of can be analyzed without
0:02:40considering the quality of the embedding that way
0:02:43in practice this means that given a dynamical model number of symmetric topological building blocks
0:02:50we can calculate locations used by each of the building blocks separately adjusting that the
0:02:55low computational cost
0:02:57the results can then be used to approximate or explain about location diagram anymore realisation
0:03:03that contains the symmetric building blocks
0:03:08you know people we illustrate the idea but not exactly like delay is
0:03:12for start compare this morning something like that contains only once in the tree building
0:03:16block the triangle
0:03:19first we calculate and identically the local that is associated to the balloon got about
0:03:24the result to run the location however
0:03:28exactly the location that from a larger network plot of the number of positive eigenvalues
0:03:33of your cooking change colours indicate location
0:03:39the agreement between both not contain that the characteristic somebody used by the semantics of
0:03:44raw indeed explain part of the behaviour of the network
0:03:48okay now with several subjective goals
0:03:52and that's what containing that triangle directed hexagon standard multi is expected to display the
0:03:58egg got their characteristic into the t
0:04:01in the comparison with indirectly contain patient confidence at the characteristic at and used by
0:04:07the difference in the trees are superimposed without interfering
0:04:12right
0:04:14logical way at the nodes we don't a symmetric rows
0:04:18the ability to underline economical application what is our understanding of the robot okay and
0:04:23the functioning of the respective complex systems
0:04:26we see the ability to handle combine topological being well defined dynamically implications they constitute
0:04:32a design future technical systems that exploit intelligence context