BEGIN:VCALENDAR VERSION:2.0 PRODID:researchseminars.org CALSCALE:GREGORIAN X-WR-CALNAME:researchseminars.org BEGIN:VEVENT SUMMARY:Konstantinos Kourliouros (Imperial College London) DTSTART:20261007T113000Z DTEND:20261007T120000Z DTSTAMP:20261006T031620Z UID:MathEngSeminars/1 DESCRIPTION:Title: Geometric Theory of Random Dynamical Systems\nby Konstantinos Kourliouros (Imperial College London) as part of Istanbul Technical Univer sity Mathematical Engineering Department Seminars\n\nLecture held in Fen-E debiyat Fakültesi B1-326.\n\nAbstract\nIn this talk\, I will present rece nt joint work with collaborators at Imperial College London on the geometr isation of the theory of random dynamical systems. Research on such system s has traditionally relied on stochastic and functional analytic methods\, particularly for differential equations driven by unbounded noise such as Brownian motion. Systems with bounded noise\, however\, generally possess multiple coexisting attractors\, each supporting a stationary ergodic pro bability measure that describes its long-term statistics.\n\nThe geometry\ , topology\, stability and bifurcations of these attractors have remained inaccessible to conventional differential-geometric tools\, partly because the hyperspace of compact subsets of the state space has no natural Banac h-space structure. I will explain how this problem can be addressed throug h a contact lift of the random dynamical system\, analogous to the Pontrya gin construction in geometric control theory. This produces a deterministi c contact dynamical system on the unit cotangent bundle\, called the bound ary system. Co-oriented boundaries of random attractors and other invarian t sets lift uniquely to invariant Legendrian submanifolds of this system.\ n\nUsing this correspondence principle\, I will discuss recent results on the stability of random attractors through normal hyperbolicity of the bou ndary system\, as well as ongoing work on local and global bifurcations th rough the associated bifurcations of Legendrian fronts.\n\nShort Bio:\n\nK onstantinos Kourliouros received his BSc and MSc degrees in applied mathem atics and mathematical physics from the University of Patras and completed further MSc studies in pure mathematics at the University of Nice. He obt ained his PhD from Imperial College London under the supervision of Profes sors Jeroen Lamb and Dima Turaev\, working on singularities in symplectic and volume-preserving geometry.\n\nHe subsequently held postdoctoral posit ions at the Technion and the University of São Paulo\, where he worked on singular and constrained Hamiltonian systems. He later returned to Imperi al College London as a Research Associate and group leader under two conse cutive EPSRC projects on the stability\, resilience and bifurcations of at tractors in random dynamical systems with bounded noise. He was recently a warded a three-year TÜBİTAK 2232-A International Fellowship for Outstand ing Researchers for his project "Geometric Methods for Singular Dynamical Systems".\n LOCATION:https://researchseminars.org/talk/MathEngSeminars/1/ END:VEVENT BEGIN:VEVENT SUMMARY:Deniz Eroğlu (Imperial College London & Kadir Has University) DTSTART:20261007T120000Z DTEND:20261007T123000Z DTSTAMP:20261006T031620Z UID:MathEngSeminars/2 DESCRIPTION:Title: Network-Oriented Dynamics and Data Science: A Mathematical View of Interacting Systems\nby Deniz Eroğlu (Imperial College London & Kadi r Has University) as part of Istanbul Technical University Mathematical En gineering Department Seminars\n\nLecture held in Fen-Edebiyat Fakültesi B 1-326.\n\nAbstract\nMany high-dimensional systems can be represented as ne tworks of interacting dynamical units. Their observable behaviour often de pends less on the individual components than on the collective organisatio n created by network topology\, heterogeneity and nonlinear coupling.\n\nT his talk will provide a broad overview of the mathematical research progra mme developed within the Network-Oriented Dynamics and Data Science group. Starting from deterministic and random dynamical systems\, I will discuss how interaction structure shapes synchronisation\, stability\, emergence and bifurcations\; how high-dimensional network problems can be reduced to analytically and computationally tractable descriptions\; and how governi ng dynamics and interaction networks can be reconstructed from data. Parti cular attention will be given to identifying loss of stability and critica l transitions from partial or noisy observations.\n\nExamples from neural and genetic networks\, climate dynamics\, power grids and transport system s will illustrate the common mathematical questions behind these applicati ons. Rather than focusing on a single result\, the talk will trace the dev elopment of this programme from the mathematical foundations of collective dynamics to data-driven reconstruction\, model reduction and the detectio n of critical transitions.\n\nShort Bio:\n\nDeniz Eroğlu received his PhD in theoretical physics from Humboldt University Berlin. Following postdoc toral research positions in the Department of Mathematics at Imperial Coll ege London and at Northwestern University\, he joined Kadir Has University in 2019. He returned to Imperial College London as a Research Fellow from 2024 to 2026\, where he worked on random dynamical systems and their appl ications and contributed to teaching and doctoral supervision.\n\nHis rese arch lies at the intersection of dynamical systems\, complex networks and data-driven mathematics. He leads the interdisciplinary Network-Oriented D ynamics and Data Science group\, which studies collective behaviour\, netw ork reconstruction\, mathematical reduction and critical transitions in bi ological\, physical and engineered systems. His research has been supporte d by TÜBİTAK and European Union programmes. He received the Science Acad emy’s 2021 Young Scientist Award in Physics and will join the University of Western Australia as a Senior Lecturer in 2027.\n LOCATION:https://researchseminars.org/talk/MathEngSeminars/2/ END:VEVENT BEGIN:VEVENT SUMMARY:Nazlı Doğan (Fatih Sultan Mehmet Vakıf University) DTSTART:20261014T113000Z DTEND:20261014T123000Z DTSTAMP:20261006T031620Z UID:MathEngSeminars/3 DESCRIPTION:Title: Toeplitz Operators: From Hardy and Bergman Spaces to Köthe Spaces \nby Nazlı Doğan (Fatih Sultan Mehmet Vakıf University) as part of Istanbul Technical University Mathematical Engineering Department Seminars \n\nLecture held in Fen-Edebiyat Fakültesi B1-326.\n\nAbstract\nToeplitz operators constitute a classical and extensively studied class of operator s arising\nnaturally in the theory of Hilbert spaces\, particularly in the setting of Hardy and Bergman\nspace. Their classical theory begins with H ardy spaces and extends naturally to Bergman\nspaces. In this talk\, we fi rst introduce Toeplitz operators on Hardy spaces and then examine their de finition and properties on Bergman spaces. We discuss compactness criteria for\nToeplitz operators in this setting and present some results obtained jointly with Şahutoğlu. \n\nMore recently\, the study of Toeplitz opera tors has been extended beyond the classical\nHilbert space setting to broa der classes of topological vector spaces. Jasiczak introduced\nand charact erized the class of Toeplitz operators on the Fréchet space of entire fun ctions. \nThis naturally leads us to consider Toeplitz operators on Köthe spaces\, which provide a broader locally convex framework containing impo rtant spaces of holomorphic functions.\nWe discuss their compactness and i nvestigate several of their dynamical properties\, including topologizabil ity\, power boundedness\, and mean ergodicity.\n\nShort Bio:\n\nAssist. Pr of. Dr. Nazlı DOĞAN received her B.Sc. degree from the Department of Mat hematics at Istanbul University in 2011. She completed her M.Sc. studies i n 2013 at the Department of Mathematics and Computer Science\, Istanbul K ültür University\, under the supervision of Prof. Dr. Mert ÇAĞLAR\, an d earned her Ph.D. degree in 2020 from the Department of Mathematical Engi neering at Istanbul Technical University under the supervision of Prof. Dr . Kamil ORUÇOĞLU and Prof. Dr. Aydın AYTUNA.\n\nFrom 2015 to 2019\, she worked as a research assistant at the Department of Mathematics\, Istanbu l Technical University. Between 2020 and 2022\, she served as an assistant professor at Nişantaşı University. Since 2022\, she has been an assist ant professor in the Department of Artificial Intelligence and Data Engine ering\, Faculty of Engineering\, Fatih Sultan Mehmet Vakıf University. Du ring 2023–2024\, she conducted postdoctoral research at the University o f Toledo with Prof. Dr. Sönmez ŞAHUTOĞLU under the TÜBİTAK 2219 Inter national Postdoctoral Research Fellowship Program.\n\nHer research focuses on the structural theory of Fréchet spaces\, operator theory on Fréchet and Hilbert spaces\, particularly Hankel and Toeplitz operators\, and pro blems at the intersection of functional analysis and several complex varia bles.\n LOCATION:https://researchseminars.org/talk/MathEngSeminars/3/ END:VEVENT BEGIN:VEVENT SUMMARY:Ahmet Göncü (Istanbul Technical University) DTSTART:20261021T113000Z DTEND:20261021T123000Z DTSTAMP:20261006T031620Z UID:MathEngSeminars/4 DESCRIPTION:Title: Real-Time Market Regime Identification in Borsa İstanbul: Machine Learning versus Simple Trend Rules\nby Ahmet Göncü (Istanbul Techni cal University) as part of Istanbul Technical University Mathematical Engi neering Department Seminars\n\nLecture held in Fen-Edebiyat Fakültesi B1- 326.\n\nAbstract\nThis study examines whether up- and down-regimes of the BIST-30 index can be identified in real time without look-ahead bias. Regi mes are defined by peaks and troughs of the smoothed price series\, with l abels recomputed at each model update from past prices only. Machine-learn ing classifiers and a Markov-switching model are evaluated in a walk-forwa rd design over a 20-year out-of-sample period (2006–2026)\, against buy- and-hold\, Turkish lira deposit rates and simple trend rules\, net of tran saction costs. A long-only regime filter matches buy-and-hold returns with about 35% lower volatility and cuts the maximum drawdown from 64% to abou t 25%\, although excess returns are not statistically significant. Machine learning performs on par with a simple moving-average rule while trading less\, and long-short strategies and take-profit exits do not improve perf ormance. The results point to regime filters as a tool for improving risk- adjusted performance: their lower volatility allows leveraged implementati on through index futures\, where the same risk budget translates into high er returns.\n\nShort Bio:\n\nAhmet Göncü is an Associate Professor of Fi nance in the Department of Management Engineering at Istanbul Technical Un iversity. He received his PhD in Financial Mathematics from Florida State University in 2009 and spent the 17 years working abroad\, including 12 ye ars in China at Shandong University and Xi'an Jiaotong-Liverpool Universit y\, where he supervised four PhD students to completion. His research cove rs quantitative finance\, asset pricing\, Monte Carlo simulations\, risk m anagement and algorithmic trading. He has published 36 articles in interna tional journals\, such as Quantitative Finance\, Annals of Operations Rese arch\, International Review of Financial Analysis\, and is the author of b ooks on quantitative finance and portfolio management. Alongside his acade mic work\, Dr. Göncü has advised hedge funds and designed algorithmic tr ading strategies for China's commodity futures markets. He still works on developing trading strategies and software solutions for fund management i ndustry in Türkiye.\n LOCATION:https://researchseminars.org/talk/MathEngSeminars/4/ END:VEVENT BEGIN:VEVENT SUMMARY:Tuğçe Sezer (Istanbul Technical University) DTSTART:20261104T113000Z DTEND:20261104T123000Z DTSTAMP:20261006T031620Z UID:MathEngSeminars/5 DESCRIPTION:by Tuğçe Sezer (Istanbul Technical University) as part of Is tanbul Technical University Mathematical Engineering Department Seminars\n \nLecture held in Fen-Edebiyat Fakültesi B1-326.\nAbstract: TBA\n LOCATION:https://researchseminars.org/talk/MathEngSeminars/5/ END:VEVENT BEGIN:VEVENT SUMMARY:Selcen Sayıcı (Independent Researcher) DTSTART:20261111T113000Z DTEND:20261111T123000Z DTSTAMP:20261006T031620Z UID:MathEngSeminars/6 DESCRIPTION:Title: New Linear Codes via Restricted Evaluation Codes\nby Selcen Sa yıcı (Independent Researcher) as part of Istanbul Technical University M athematical Engineering Department Seminars\n\nLecture held in Fen-Edebiya t Fakültesi B1-326.\n\nAbstract\nGalindo et al. introduced a class of cod es which are obtained by evaluation of polynomials at the roots of a trace map. Via subfield subcodes\, this construction yields new linear codes\nw ith good parameters. Here\, we extend this construction to allow evaluatio n at the roots of any\npolynomial which splits in the field of evaluation. Our proof relies on Galois-closedness of codes in\nconsideration. Moreove r\, we introduce a lengthening process that preserves Galois-closed proper ty\nof restricted evaluation codes. Subfield subcodes of such lengthened c odes yield further good linear\ncodes. In total\, we obtain 17 linear code s over $\\mathbb{F}_4$ and $\\mathbb{F}_5$ which improve the best known li near code parameters. Moreover\, we give a construction for two linear cod es which have the best known parameters\, but for which no construction wa s known before.\n\nShort Bio:\n\nSelcen Sayıcı received her B.Sc. from I stanbul Bilgi University in Mathematics. Then\, she got her Ph.D. in Mathe matics\nfrom Sabancı University under supervision of Cem Güneri at Augus t 2020. She was postdoctoral fellow in Middle East Technical University\, in charge of Ferruh Özbudak between July 2021 and April 2023. Her researc h area includes coding theory\, curves over finite fields.\n LOCATION:https://researchseminars.org/talk/MathEngSeminars/6/ END:VEVENT BEGIN:VEVENT SUMMARY:Ilmar Gahramanov (Boğaziçi University) DTSTART:20261125T113000Z DTEND:20261125T123000Z DTSTAMP:20261006T031620Z UID:MathEngSeminars/7 DESCRIPTION:by Ilmar Gahramanov (Boğaziçi University) as part of Istanbu l Technical University Mathematical Engineering Department Seminars\n\nLec ture held in Fen-Edebiyat Fakültesi B1-326.\nAbstract: TBA\n LOCATION:https://researchseminars.org/talk/MathEngSeminars/7/ END:VEVENT BEGIN:VEVENT SUMMARY:Ali Demirci (Istanbul Technical University) DTSTART:20261202T113000Z DTEND:20261202T123000Z DTSTAMP:20261006T031620Z UID:MathEngSeminars/8 DESCRIPTION:by Ali Demirci (Istanbul Technical University) as part of Ista nbul Technical University Mathematical Engineering Department Seminars\n\n Lecture held in Fen-Edebiyat Fakültesi B1-326.\nAbstract: TBA\n LOCATION:https://researchseminars.org/talk/MathEngSeminars/8/ END:VEVENT BEGIN:VEVENT SUMMARY:Ferruh İlhan (Turkish-German University) DTSTART:20261209T113000Z DTEND:20261209T123000Z DTSTAMP:20261006T031620Z UID:MathEngSeminars/9 DESCRIPTION:by Ferruh İlhan (Turkish-German University) as part of Istanb ul Technical University Mathematical Engineering Department Seminars\n\nLe cture held in Fen-Edebiyat Fakültesi B1-326.\nAbstract: TBA\n LOCATION:https://researchseminars.org/talk/MathEngSeminars/9/ END:VEVENT BEGIN:VEVENT SUMMARY:Saliha Demirbüken (Gebze Technical University) DTSTART:20261216T113000Z DTEND:20261216T123000Z DTSTAMP:20261006T031620Z UID:MathEngSeminars/10 DESCRIPTION:by Saliha Demirbüken (Gebze Technical University) as part of Istanbul Technical University Mathematical Engineering Department Seminars \n\nLecture held in Fen-Edebiyat Fakültesi B1-326.\nAbstract: TBA\n LOCATION:https://researchseminars.org/talk/MathEngSeminars/10/ END:VEVENT BEGIN:VEVENT SUMMARY:Fatma Zürnacı Yetiş (Istanbul Technical University) DTSTART:20261223T113000Z DTEND:20261223T123000Z DTSTAMP:20261006T031620Z UID:MathEngSeminars/11 DESCRIPTION:by Fatma Zürnacı Yetiş (Istanbul Technical University) as p art of Istanbul Technical University Mathematical Engineering Department S eminars\n\nLecture held in Fen-Edebiyat Fakültesi B1-326.\nAbstract: TBA\ n LOCATION:https://researchseminars.org/talk/MathEngSeminars/11/ END:VEVENT END:VCALENDAR