Upravni odbor Matematickog instituta SANU je na nedavnoj sednici doneo odluku da se dosadasnji Seminar za primenjenu matematiku, sada nazove Seminar za racunarstvo i primenjenu matematiku, a u cilju potenciranja znacaja racunarstva kao jedne od oblasti delatnosti Instituta. Istovremeno, Upravni odbor doneo je odluku o osnivanju Odeljenja za racunarstvo i primenjenu matematiku i vezao rad novog odeljenja za rad Seminara za racunarstvo i primenjenu matematiku.

Petak, 8.05.2015. u 14:00h, sala 301f, MI SANU
!!!OBRATITE PAZNJU NA TERMIN!!!
Zajednicki sastanak sa Odeljenjem za matematiku, Odeljenjem za mehaniku, i Seminarom za istoriju i filozofiju matematike i mehanike
Katica R. (Stevanovic) Hedrih, Matematicki institut SANU
PETROVIC'S ELEMENTS OF MATHEMATICAL PHENOMENOLOGY AND PHENOMENOLOGICAL MAPPINS: THEORY AND APPLICATIONS

Abstract: Lecture starts with short description of Element of Mathematical Phenomenology and Phenomenological Mappings published in Petrovic's theory. The biographical data of Mihailo Petrovic (1868-1943) is presented. Petrovic was a famous Serbian mathematician, one of three Henrei Poincare's doctoral students. Next it is a description of abstraction of real system to the physical, chemical or biological and mathematical model.

Some of basic elements of mathematical phenomenology are elements of non-linear-functional transformations of coordinates from one to other functional curvilinear coordinate system. Some of these elements, as it is basic vectors of tangent space of kinetic point vector position and their changes (velocity of their magnitude extensions and component angular velocities of rotations), are presented in different functional coordinate systems.

Mihailo Petrovic's theory contains two types of analogies: mathematical and qualitative, and in this lecture third type - structural analogy is described. Taking into account large possibility for applications of all three types of analogies, numerous original examples are presented using, between other, fractional system dynamics with one degree of freedom, finite number of degrees of freedom as well as multi-body discrete continuum hybrid fractional order system dynamics.

Mathematical analogies between vector models in local area of stress state, strain stare of the point in stressed and deformed deformable body as well as with vector model of the mass inertia moment state at point of rigid body, used mass inertia moment vectors coupled for pole and axis, are presented, also.

Using discrete continuum method, fractional order mode analysis in hybrid system dynamics is presented. For a class of fractional order system dynamics with finite number of degrees of freedom, independent eigen main fractional order modes are determined with corresponding eigen main coordinates of the system and presented by Tables. A number of theorems of energy fractional order dissipation presented in corresponding Tables, also. It is shown that applications of qualitative, structural and mathematical analogies in analysis of fractional order modes appear in analogous mechanical, electrical and biological fractional order chains, and that is very power, suitable and useful tools to reduce research models to corresponding minimal numbers, and, in same time, develop power of analysis use phenomenological mappings between local and global phenomena and properties.

An analogy between kinetic parameters of collision of two rigid body in translator motions and collision of two rolling billiards' balls is presented and corresponding new theorems are defined.

Phenomenological approximate mappings on nonlinear phenomena, in local area around stationary points or stationary states, are presented. Corresponding kinetic parameters of model of nonlinear dynamics of real system behavior are presented, also. For obtaining approximate differential equations and approximate solutions in local area around singular points, linear and non-liner approximations are used. Method of local analysis based on phenomenological approximate mappings between local linear as well as nonlinear phenomena is power to obtain information of all local nonlinear phenomena in the nonlinear dynamics of the system for completing kinetic elements for global analysis of the system nonlinear dynamics and stability and to use different analogies.

Utorak, 12.05.2015. u 14:15h, Sala 301f, MI SANU:
Jelena Jovanovic, doktorant, Matematicki fakultet, Univerzitet u Beogradu
OPTIMALAN USLOV ZA POLU-DISTRIBUTIVNOST MREZE KONGRUENCIJA

Rezime: Izlozicemo istrazivanje karakterizacije svojstva polu-distributivnosti mreze kongruencija u konacnim algebrama (i sire, u lokalno konacnim varijetetima algebri) pomocu sintaksnih uslova, odnosno term-identiteta. Kao rezultat smo dobili uslov koji je jednostavniji od do sada postojecih i takodje najbolji moguc (tj. optimalan) . Posebnu paznju posveticemo delu istrazivanja koji se odnosi na racunarsku pretragu algebri polimorfizama malih digrafa. Predstavicemo primenjene metode i analizu vremenske slozenosti.
Saradnici na ovim istrazivanjima bili su P. Markovic, R. McKenzie i M. Moore.

Utorak, 19.05.2015. u 14:15h, Sala 301f, MI SANU:
Ivan Zaklanovic, Cisco Srbija
INTERNET OF EVERYTHING

Apstrakt: Internet of everything, odnosno Internet svega, treba da omogu.i povezivanje ljudi, mesta, procesa, stvari i podataka. Omogucen je razvojem tehnologija kao sto su cloud computing, tehnologije mobilnog racunarstva, i 3D stampe. Njegov razvoj se narocito vezuje za primenu Internet of Things tehnologija i inovacije u brojnim oblastima tehnike, od bezicnih senzora do nanotehnologije. U narednim godinama ocekuju se intenzivnije povezivanje fizi.kog i virtuelnog sveta, razvoj novih poslovnih modela i promene u nacinu zivota. Internet of everything se sve cesce razmatra kao tehnoloska oblast koja ce u znatnoj meri odrediti buducnost racunarstva i komunikacija i koja predstavljava osnovu infrastrukture za pametniju planetu. U ovom predavanju bice napravljen pregled aktuelnih Internet of everything tehnologija kao i prikaz ocekivanih pravaca razvoja i primene.

Utorak, 26.05.2015. u 14:15h, Sala 301f, MI SANU:
Nebojsa Ikodinovic, Matematicki fakultet, Univerzitet u Beogradu
NEMONOTONO ZAKLJUCIVANJE

Rezime: Uopsteno govoreci, nemonotono zakljucivanje odrazava uobicajeno svojstvo ljudskog misljenja - promenu misljenja. Pri monotonom zakljucivanju, zakljucci izvedeni iz nekih premisa mogu biti izvedeni i iz bilo kog veceg skupa premisa. Pri nemonotonom zakljucivanju, zakljucci su privremeni, jer mogu biti povuceni kada se pojave neke nove informacije.
Na predavanju ce biti prikazani najznacajniji pristupi proucavanju nemonotonog zakljucivanja, sa naglaskom na motivacionim primerima i primenama.

Sreda, 27.05.2015. u 12:15h, Sala 301f, MI SANU:
Nemanja Djuric, Yahoo Labs, USA
MACHINE LEARNING FOR COMPUTATIONAL ADVERTISING

Abstract: Recent advent of big data applications and platforms has led to a renaissance in many areas of machine learning and data mining. Computational advertising has particularly benefited, and the industry has observed a steady two-digit growth in the past few years. In order to maintain and improve upon this positive trend, researchers in academia and industry alike are faced with numerous theoretical and practical questions. In this talk we discuss the main challenges in the research focus, as well as current state-of-the-art and future directions in the fields of Ad Targeting, User Modeling, Recommender Systems, and related areas in the era of Big Data.

This is a joint work with Vladan Radosavljevic and Mihajlo Grbovic.

RUKOVODIOCI SEMINARA

MI SANU
Vera Kovačević-Vujčić
Milan Dražić

FON
Zorica Bogdanovic
Marijana Despotovic-Zrakic

IEEE
Bozidar Radenkovic