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.

Utorak, 03.12.2013. u 14:15, Sala 301f, MI SANU:
doc.dr Zoran Sevarac, Fakultet organizacionih nauka

NEUROPH - SOFTVER ZA RAZVOJ NEURONSKIH MREZA U JAVA OKRUZENJU

Rezime: Predavanje predstavlja softver otvorenog koda za razvoj neuronskih mreza Neuroph, koji se razvija na FON-u u okviru Laboratorije za vestacku inteligenciju, a koji je kompanija Oracle nagradila kao jedan od najinovativnijih projekata na Java platformi. Opisan je razvoj ovog softvera od ranih pocetaka u vidu studentskog projekta, do najpopularnijeg svetskog softvera u oblasti neuronskih mreza. Predavanje obuhvata i kratku demonstraciju kreiranja neuronskih mreza pomocu ovog softvera, njegovu primenu u nastavi i za prepoznavanje slika i slova.

Utorak, 10.12.2013. u 14:15h, Sala 301f, MI SANU:

Marija Milojevic Jevric, Matematicki institut SANU
OPTIMIZACIONI PROBLEMI KOD MASINSKIH ELEMENATA I SKLOPOVA

Rezime: Analiza dostupne literature o dizajnu masinskih elemenata i sklopava ukazala je na pojavu mnogih optimizacionih problema, koji su najcesce kontinualnog tipa. Na ove probleme istrazivaci su primenjivali metodu konacnih elemanata, obicne algebarske transformacije, Bajesove i vestacke neuronske mreze (za predikciju), hibridne evolutivne algoritme, genetske algoritme. Dolazi se do zakljucka da simulacija sto realnije situacije sa svim uticajnim parametrima moze biti vremenski i/ili memorijski zahtevna. Zbog toga su se aproksimativne metode (tipa neuralnih mreza ili genetskih algoritama) pokazale prikladnijima. One omogucavaju dobijanje preciznijih rezultata za znatno krace vreme rada.
Kao primer analizira se optimizacija faktora raspodele opterecenja duz linije kontakta spregnutog cilindricnog zupcastog para. Matematicki model spregnutog cilindricnog zupcastog para je utvdjen u skladu sa ISO stadardima. Funkcija optimizacije ima 12 direktnih ulaznih vrednosti, dok se prostale vrednosti dobijaju jednostavnim algebarskim proracunom ili po potrebi Njutn-Rapsonovim metodom, nelinearnom interpolacijom... Za optimizaciju su korisceni genetski algoritmi integrisani u MatLab softver. Rezultati su pokazali da je najuticajni parametar na raspodelu opterecenja duz linije dodira nagibni ugao zupca kod zupcanika sa kosim zupcima, a zatim i koeficijenti pomeranja profila osnovne zupcaste letve. Uticajni parametri pripadaju grupi geometrijsko-konstrukcionih parametara. Optimizacijom je utvrdjeno da se teorijski moze postici ravnomerna raspodela opterecenja za bilo koji broj zubaca iz opsega (18 - 54) i za bilo koji prenosni odnos (iz opsega 1-5).

Utorak, 17.12.2013. u 14:15h, Sala 301f, MI SANU:

Dragan Bjelica, Fakultet organizacionih nauka
UTICAJ WEB BAZIRANIH ALATA ZA UPRAVLJANJE PROJEKTIMA NA TEHNICKE KOMPETENCIJE PROJEKTNIH PORTFOLIO MENADZERA

Rezime: U radu je prikazan uticaj modula za upravljanje projektima u web baziranom okruzenju na tehnicke kompetencije projektnih portfolio menadzera. Standardni moduli u web baziranim alatima za upravljanje projektima su: resursni i projektni centar, upravljanje finansijama, upravljanje vremenom i aktivnostima, modul za kolaboraciju, upravljanje rizikom i spornim pitanjima, poslovna inteligencija i izvestavanje, portfolio selekcija i analitika (modul strategije). U skladu sa IPMA Competence Baseline, sprovedeno je istrazivanje na uzorku od 51 projektnog portfolio menadzera iz 51 organizacije iz profitnog, neprofitnog i javnog sektora. Instrument za prikupljanje podataka bio je upitnik. Doprinosi istrazivanja se mogu svrstati u cetiri segmenta: 1. tehnicke kompetencije iz organizacionog ugla gledista; 2. tehnicke kompetencije iz projektnog ugla gledista; 3. kriticni faktori uspeha iz organizacionog, projektnog i licnog ugla gledista; 4. primena web baziranih modula za upravljanje projektima.

Ponedeljak, 23.12.2013. u 18:00, Sala 301f, MI SANU:

Nemanja Djuric, DABI, Temple University, USA
BIG DATA ALGORITHMS FOR VISUALIZATION AND SUPERVISED LEARNING

Abstract: Explosive growth in data size, data complexity, and data rates, triggered by emergence of high-throughput technologies such as remote sensing, crowd-sourcing, social networks, or computational advertising, in recent years has led to an increasing availability of data sets of unprecedented scales, with billions of high-dimensional data examples stored on hundreds of terabytes of memory. As a result, there exists an evident need for development of novel, scalable algorithms for big data.
This presentation addresses these important problems, and propose both supervised and unsupervised tools for handling large-scale data. First, we consider unsupervised approach to big data analysis, and explore scalable, e.cient visualization method that allows fast knowledge extraction. Next, we consider supervised learning setting and propose algorithms for fast training of accurate classi.cation models on large data sets, capable of learning classi.ers on data sets with millions of examples and features within minutes. Experimental evaluation of the proposed methods shows state-of-the-art performance on a number of synthetic and real-world data sets, further paving a way for e.cient and e.ective knowledge extraction from big data problems.

Utorak, 24.12.2013. u 14:15, Sala 301f, MI SANU:

Milan Drazic, Matematicki fakultet Univerziteta u Beogradu, Rade Lazovic, Vera Kovacevic-Vujcic, Fakultet organizacionih nauka
SPARSITY PRESERVING PRECONDITIONERS FOR LINEAR SYSTEMS IN INTERIOR POINT METHODS

Apstract: Systems of normal equations arising in interior point methods for linear programming in the case when the optimal face is degenerate have highly ill-conditioned coefficient matrices. In 2004, Monteiro, Oeal and Tsuchiya proposed preconditioners which guarantee uniform well-conditionedness. However, the proposed preconditioners may lead to considerable loss of sparsity. Our approach is directed towards a generalization of the proposed preconditioners which make a balance between a sparsity and well-conditionedness. Experimental results on OR-Lib instances show the effects of the new approach.
Nakon predavanja, uspesno zavrsena kalendarska godina u okviru Seminara bice svecano obelezena prigodnim koktelom. Koktel povodom kraja godine bice organizovan u Institutu i 31.12.2013. godine u podne. Koristimo ovu priliku da sve ucesnike Seminara pozovemo i na ovo druzenje.

Ponedeljak, 30.12.2012. u 14:15, soba 301f, MI SANU:

Dejan Milojicic, HP Labs, USA, 2014 IEEE Computer Society President
IEEE COMPUTER SOCIETY 2022 REPORT

Abstract: Predicting future is hard and risky. Predicting future in computer industry is even harder and riskier due to dramatic changes in technology and limitless challenges to innovation - to bringing new technology into a broad use. Only a small fraction of innovations are truly disrupting the state of the art. Some are not practical or cost-effective, others are ahead of its time, yet others do not have market. Nine technical leaders in IEEE Computer Society have spearheaded writing a Technical Report titled IEEE CS 2022, symbolically surveying 22 potential technologies that can change the landscape of computer science and industry by the year 2022. In particular, they have surveyed the following technologies: 3D printing, big data and analytics, open intellectual property movement, massive online open courses, security cross-cutting issues, universal memory, 3D integrated circuits, photonics, cloud computing, computational biology and bioinformatics, device and nanotechnology, sustainability, high performance computing, the Internet of things, life sciences, machine learning and intelligent systems, natural user interfaces, networking and interconnectivity, quantum computing, software defined networks, multicore, and robotics. For each of the 22 technologies, a common approach has been taken: summary of the state-of-the-art, challenges, where we think the technology will go, and what is disruption. We have tied these technologies into a common scenario that we call seamless intelligence. Together they represent a common view of the future. Independently, we have surveyed a few thousand IEEE members on the technology drivers and disruptors. They have confirmed some of our predictions and provided another perspective on the future of technology advancements. Finally, we have endeavored to predict what kind of future society is needed for our profession, for professionals that will be learning, practicing and putting into use the technologies we presented in this paper.This presentation was intended for computer science professionals, students and professors, as well as laymen interested in technology and technology use. It is also targeted to members of computer society and similar societies around the world, daring to predict what kind of a future professional society will be best suited to take discussed technologies to the next level through its publications, conferences, communities, standards, courses, and future artifacts in support of our profession and humanity.

About the speaker: Dr Dejan Milojicic is a senior researcher and senior manager at HP Labs, Palo Alto, CA, working in the technical areas of systems software, distributed systems, high performance computing and service management. He is IEEE Computer Society 2014 President. He is a founding Editor-in-Chief of IEEE ComputingNow (2008-2012) and past chair of IEEE Technical Committee on Operating Systems (2000). He has been on many program committees of ACM, IEEE and USENIX conferences (ICDCS, CLOUD, ICWS, EDOC, AAMAS, ICAC, Middleware, HotCloud, IC2E, etc.) and on journal editorial boards (IEEE Internet Computing and IEEE Transactions on Cloud Computing). He has been a member of IEEE CS, ACM, and USENIX for over 20 years. He worked in OSF Research Institute, Cambridge, MA [1994-1998], and Institute "Mihajlo Pupin", Belgrade, Serbia [1983-1991]. He is teaching a class on Cloud Management at SJSU, San Jose CA. He received his PhD from University of Kaiserslautern, Germany (1993); and MSc/BSc from Belgrade University, Serbia (1983/86). He has been on 6 thesis committees (GaTech, UIUC, Monash, USP) and has guided over 40 interns. Dejan is an IEEE Fellow, ACM Distinguished Engineer, and USENIX member. Dejan has published over 130 papers and 2 books; he has 12 patents and 25 patent applications.

RUKOVODIOCI SEMINARA

MI SANU
Vera Kovačević-Vujčić
Milan Dražić

FON
Zorica Bogdanovic
Marijana Despotovic-Zrakic

IEEE
Bozidar Radenkovic