Thursday 7.4 at 16.15, auditorium CK112

• Tuomas Hytönen (University of Helsinki): Quantitative Differentiation
• The basic question that I discuss is at the foundations of mathematical analysis: Given epsilon, how large can you make your delta in the epsilon-delta definition of differentiability? In fact, we would like to find a good value of delta that works not just for one function but a family of functions. This innocent-looking questions leads to a surprisingly rich theory. This is recent joint work with Sean Li (Chicago) and Assaf Naor (Princeton).
• Tuomas Hytönen was educated at the Helsinki University of Technology in the turn of the millennium. After postdoctoral positions at the Delft University of Technology (Netherlands) and the University of Turku, he has been based at the University of Helsinki since 2006. He is the leader of the Harmonic Analysis Research Group and was appointed Professor of Mathematics in 2015.
  
  

Wednesday 9.12 at 16.15, auditorium CK112

• Camilla Hollanti (Aalto University): Matroids Meet Cloud Storage
• Matroids are handy combinatorial objects that characterize independence in abstract terms, much in the same way as one can define linear independence for matrices. In big data centres such as those of Google and Facebook, there are thousands or even millions of distributed and interlinked servers storing huge amounts of data. In this talk, I will explain what matroids and distributed storage systems are, and how we can optimize certain aspects of storage systems by using matroid theory. The talk requires no other prerequisites than linear algebra.
• Camilla Hollanti is Associate Professor in mathematics at Aalto University, and is leading a research group in Algebra, Number Theory, and Applications. The research topics of her group vary from highly applied research related to wireless communications and cloud storage to purely algebraic and number theoretic research. Currently, she is leading three Academy of Finland projects, including a Finland Distinguished Professor (FiDiPro) project with Prof. Marcus Greferath.

Thursday 1.10 at 16.15, auditorium B123

Jukka Corander (University of Helsinki): Voodoota vai vakavaa päättelyä? Approksimatiivinen inferenssi kohtaa koneoppimisen

• Moderni tilastollinen mallinnus johtaa enenevässä määrin todennäköisyyskuvauksiin joille ei voida määritellä uskottavuusfunktiota analyyttisesti tai numeerisesti. Tilanne on tyypillinen sekä mikro- että makrotason mallinnuksessa. Ekosysteemien, ilmaston ja perimän muuttumisen dynamiikan mallit ovat esimerkkejä tämänkaltaisista haasteista. Malleja voidaan kuitenkin implementoida tietokoneohjelmiksi joiden avulla tuotetaan keinotekoisia havaintoja. Vertaamalla niitä oikeisiin havaintoihin voidaan pyrkiä päättelemään mallien parametreille järkeviä arvoja. Tämänkaltainen approksimatiivinen päättely on yksi viime vuosien kuumimmista tutkimusaloista tilastotieteessä. Sen aiheuttamat laskennalliset haasteet ovat kuitenkin valtavia ja rajoittavat lähestymistavan laajempaa käyttöä. Tässä esitelmässä perehdytään COIN-huippuyksikössä äskettäin kehitettyihin tietojenkäsittelytieteen menetelmiin joiden avulla päättelyongelma ratkaistaan tilastollisen koneoppimisen keinoin. Nämä menetelmät antavat tarkkoja ratkaisuja 1000-10000 kertaa nopeammin kuin perinteisen tilastotieteen alalla aiemmin kehitetyt algoritmit ja mahdollistavat siten uusien sovellusten syntymistä.
• Jukka Corander on tilastotieteen professori, COIN-huippuyksikön varajohtaja sekä HIITin tutkimusohjelmajohtaja. Coranderin ryhmän kehittämillä laskennallisen päättelyn menetelmillä on saavutettu monia laajaa kansainvälistä huomiota saaneita tutkimustuloksia liittyen bakteerien ja virusten evoluutioon sekä leviämiseen. Vuonna 2014 Jukka Coranderille myönnettiin ensimmäisenä suomalaisena tutkijana Cozarelli-tiedejulkaisupalkinto.

Wednesday  22.4.2015 at 16.15, auditorium CK112

        Samuli Siltanen (University of Helsinki): 3D X-ray Vision, or the Mathematics of Tomography 

• Tomographic X-ray imaging was invented by Cormack and Hounsfield in the 1960’s, and they received a Nobel Prize for it. Now Computerized Tomography (CT) is daily practice in hospitals all over the world. This talk describes the basic mathematics behind classical tomography and shows some recent developments into low-dose and dynamic X-ray    imaging.
• Samuli Siltanen is a Professor of Industrial Mathematics at the University of Helsinki. His research interests include inverse problems of medical imaging.

Thursday 12.2.2015 at 15.15, auditorium CK112

        Kari Vilonen (Northwestern University, University of Helsinki): What is Langlands?   

• In this lecture I will explain, in as elementary terms as possible, what the Langlands program is about.
• Kari Vilonen is Professor of mathematics at Northwestern University and Helsinki University. He specializes in (algebraic) geometry, representation theory, and the Langlands       program.

Wednesday 5.11.2014 at 16.15, auditorium CK112

Boban Velickovic (University of Paris 7): How to prove unprovable statements

• Boban Velickovic is a Professor at the University of Paris 7. His research interests include combinatorial set theory and forcing, games and infinitary languages, classification     problems and definable equivalence relations, and theoretical computer science.

Wednesday 22.10.2014 at 16.15, auditorium CK112

        Antti Kupiainen (University of Helsinki): What is Random Geometry?       

• Antti Kupiainen is an Academy Professor at the University of Helsinki and leader of the Mathematical physics group. His background is in constructive quantum field theory, and  research interests include foundations of non-equilibrium statistical mechanics and the problem of deriving diffusion and transport from first principles.