BEGIN:VCALENDAR VERSION:2.0 PRODID:-//wp-events-plugin.com//7.4.2//EN TZID:Asia/Jerusalem X-WR-TIMEZONE:Asia/Jerusalem BEGIN:VEVENT UID:64@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20230509T143000 DTEND;TZID=Asia/Jerusalem:20230509T153000 DTSTAMP:20250109T093503Z URL:https://math.technion.ac.il/events/gauss-legendre-features-for-scalabl e-gaussian-process-regression/ SUMMARY:Gauss-Legendre Features for Scalable Gaussian Process Regression DESCRIPTION:Lecturer:Paz Fink Shustin\n Location:Amado 814\n Gaussian proce sses provide a powerful probabilistic kernel learning framework\, which al lows high-quality nonparametric learning via methods such as Gaussian proc ess regression. Nevertheless\, its learning phase requires unrealistic mas sive computations for large datasets. In this talk\, I present a quadratur e-based approach for scaling up Gaussian process regression via a low-rank approximation of the kernel matrix. The low-rank structure is utilized to achieve effective hyperparameter learning\, training\, and prediction. Ou r Gauss-Legendre features method is inspired by the well-known random Four ier features approach\, which also builds low-rank approximations via nume rical integration. However\, our method is capable of generating high-qual ity kernel approximation using a number of features that is poly-logarithm ic in the number of training points\, while similar guarantees will requir e an amount that is at the very least linear in the number of training poi nts when using random Fourier features. The utility of our method for lear ning with low-dimensional datasets is demonstrated using numerical experim ents.\n\n \;\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:65@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20230517T143000 DTEND;TZID=Asia/Jerusalem:20230517T153000 DTSTAMP:20250109T093504Z URL:https://math.technion.ac.il/events/algebraic-geometric-problems-for-ce rtain-quantum-interaction-models/ SUMMARY:Algebraic/geometric problems for certain quantum interaction models DESCRIPTION:Lecturer:Masato Wakayama \n Location:Amado 719\n We focus spect ral structure of the quantum Rabi model (QRM) and its asymmetric version ( AQRM)\, which are widely studied as the most fundamental models of light-m atter interactions\, and their covering model. The latter model is called the non-commutative harmonic oscillator (NCHO) and has rich but mysterious arithmetic properties. Certain algebraic and geometric problems/conjectur es arose from the study of the degeneracy of the spectrum and hidden symme try for AQRM will be presented. The talk is based on the joint work with C id Reyes-Bustos and longtime collaboration with Kazufumi Kimoto.\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 719 END:VEVENT BEGIN:VEVENT UID:66@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20230523T143000 DTEND;TZID=Asia/Jerusalem:20230523T153000 DTSTAMP:20250109T093504Z URL:https://math.technion.ac.il/events/convexity-and-second-order-theory-i n-calculus-of-variations/ SUMMARY:Convexity and second order theory in calculus of variations DESCRIPTION:Lecturer:Alexander Ioffe\n Location:Amado 814\n The main result to be presented at the talk contains a second order necessary\ncondition for a strong minimum in the standard problem of calculus of variations.\nN ovelty of this result is emphasized by the fact that no idea of a possibil ity of such\na condition has ever appeared in the classical theory. A simp le example shows that\nthe theorem can work when other known necessary con ditions fail.\nOriginally the theorem was obtained as a consequence of a c orresponding result\nfor a sufficiently general class of optimal control p roblems (Calculus of Variations and\nPDEs (2020)\, 59:83) which will be br ie y explained in the second part of the talk.\nBut an independent proof o f the theorem is noticeably simpler and will be described\nin sufficient d etails. The key element of the proof is a version of the convexication\nth eorem proved in 1930 by N.N. Bogolyubov (which is really a fundamental res ult\nclosely connected with the classical Weierstrass necessary condition but\, strangely\,\ntypically not mentioned in courses on calculus of varia tions).\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:118@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20231107T133000 DTEND;TZID=Asia/Jerusalem:20231107T143000 DTSTAMP:20250109T093604Z URL:https://math.technion.ac.il/events/from-stochastic-to-deterministic-st ochastic-gradient-descent-in-non-convex-and-high-dimensional-problems/ SUMMARY:From Stochastic to Deterministic: Stochastic Gradient Descent in N on-convex and high Dimensional Problems DESCRIPTION:Lecturer:Inbar Seroussi\n Location:Amado 814 (and zoom)\n \n ht tps://technion.zoom.us/j/2646490052?pwd=UDUreVNSTnV1OHFzNWJsVXhQQW1Rdz09 Meeting ID: 264 649 0052 Passcode: inbar CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 (and zoom) END:VEVENT BEGIN:VEVENT UID:122@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20231114T160000 DTEND;TZID=Asia/Jerusalem:20231114T170000 DTSTAMP:20250109T093604Z URL:https://math.technion.ac.il/events/on-nonparametric-estimation-of-low- and-high-dimensional-data/ SUMMARY:On (non)parametric estimation of low and high-dimensional data DESCRIPTION:Lecturer:Shira Faigenbaum-Golovin\n Location:zoom\n \n https:// technion.zoom.us/j/97941552937?pwd=RUNuSmJuSDFJeEpBYUNXamVzUFRYZz09 Meeti ng ID: 979 4155 2937 Passcode: shira CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:zoom END:VEVENT BEGIN:VEVENT UID:149@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240116T133000 DTEND;TZID=Asia/Jerusalem:20240116T143000 DTSTAMP:20250109T093607Z URL:https://math.technion.ac.il/events/shaping-liquids-into-optical-compon ents/ SUMMARY:Shaping liquids into optical components DESCRIPTION:Lecturer:Valeri Frumkin\n Location:Amado 814\n In this talk I w ill present how interfacial phenomena in fluid mechanics can be applied to the fabrication of high-quality optical components over a wide range of s cales and types.\nIn the first part of this talk\, I will present the Flui dic Shaping method\, which relies on surface energy minimization under neu tral buoyancy conditions to shape liquid bodies into optical topographies. This method has enabled rapid fabrication of a wide range of freeform opt ical components with sub-nanometric surface quality\, and was recently use d to demonstrate the first fabrication of lenses on board the Internationa l Space Station. I will also discuss how the inherent scale invariance of the Fluidic Shaping method\, made it the underlying technology behind the Fluidic Telescope Experiment (FLUTE) at NASA – a project aimed at explor ing the use of fluids for creation of large space telescopes.\nIn the seco nd part of my talk\, I will introduce a method that leverages projected li ght patterns as a mechanism for freeform deformations of thin liquid films via the thermocapillary effect. After establishing a basic model for the film’s topography under the effect of thermocapillarity\, I will introdu ce our approach to solving the inverse problem\, which is finding the temp erature field required to achieve a desired topography. I will then demons trate how combining photo-curable films with this approach can be used to fabricate high-quality\, solid diffractive optical elements\, in less than five minutes.\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:157@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240130T133000 DTEND;TZID=Asia/Jerusalem:20240130T143000 DTSTAMP:20250109T093608Z URL:https://math.technion.ac.il/events/dual-formulation-of-the-multi-state -schrodinger-poisson-equation/ SUMMARY:Dual formulation of the multi-state Schrodinger Poisson Equation DESCRIPTION:Lecturer:Gershon Wolansky\n Location:Amado 814\n I'll review  the Schrodinger-Poisson system in the attractive case (Choquard equation) \, and extend its formulation to multi state (involving several eigenstate s). A dual formulation\, handy for  this case\, will be presented. Using the Lieb-Thirring   inequality\, a result concerning the existence of ste ady states in the sub-critical regime is introduced.\n\n \;\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:188@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240219T153000 DTEND;TZID=Asia/Jerusalem:20240219T163000 DTSTAMP:20250109T093656Z URL:https://math.technion.ac.il/events/symmetry-breaking-bifurcations-in-f luid-structure-interaction/ SUMMARY:Symmetry-Breaking Bifurcations in Fluid-Structure Interaction DESCRIPTION:Lecturer:Jun Zhang\n Location:Amado 814\n The study of fluid-st ructure interaction involves many disciplines: from fluid dynamics\, biolo gical locomotion\, mechanical engineering\, geophysical phenomena to appli ed mathematics. In this talk\, I will demonstrate a few examples when seem ingly symmetric systems bifurcate into asymmetric and unstable states as a result of close interaction between structure and fluid. In particular\, one example shows how a "symmetric bird" or a symmetric wing\, once flappe d up and down in a fluid environment\, could "take off" and sustain a unid irectional flight. Another example shows that a model-continent\, when int eracting with a convective mantle\, results in oscillatory or trapped stat es. Further\, I will discuss a recent experiment that was inspired by the superrotation of the solid core of the Earth\, as a symmetric structure ta kes a spontaneous spin in a symmetric cavity that contains fluid undergoin g turbulent thermal convection. These phenomena were first found experimen tally\, and the results have been investigated by further experiments and numerical studies.\n\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:199@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240305T133000 DTEND;TZID=Asia/Jerusalem:20240305T143000 DTSTAMP:20250109T093657Z URL:https://math.technion.ac.il/events/completing-large-matrices-with-only -few-observed-entries-a-provable-one-line-algorithm/ SUMMARY:Completing large matrices with only few observed entries: A provabl e one-line algorithm DESCRIPTION:Lecturer:Pini Zilber\n Location:Amado 814\n Suppose you observe very few entries from a large matrix. In many applications\, you can safe ly assume the matrix is (approximately) low rank. How can you use this kno wledge to complete the missing entries? In this talk\, I will present a no vel yet surprisingly simple method to solve the problem. Our method is sho wn to recover the matrix in the challenging settings of very few observati ons and/or ill-conditioning\, where many popular methods fail. We also pro ve strong theoretical guarantees for our method\, some of which improve up on the state of the art. Finally\, I will discuss how one can complete a m atrix by incorporating extra knowledge beyond its observed entries\, a pro blem known as inductive matrix completion. The talk is based on two joint papers with Boaz Nadler\, published in SIMODS and ICML.\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:204@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240312T133000 DTEND;TZID=Asia/Jerusalem:20240312T143000 DTSTAMP:20250109T093658Z URL:https://math.technion.ac.il/events/revisiting-the-exclusion-principle- in-epidemiology-at-its-ultimate-limit/ SUMMARY:Revisiting the exclusion principle in epidemiology at its ultimate limit  DESCRIPTION:Lecturer:Nir Gavish\n Location:Amado 814\n  \nThe competitive exclusion principle in epidemiology implies that more transmissible strain s will tend to outcompete other strains\, leading to the extinction of les s competitive strains over time.  This principle holds true when the reco very by one strain provides complete immunity to other strains' infections .  Nevertheless\, studies have shown that there is coexistence among comp eting strains in epidemic systems with various interactions between the st rains\, but these studies focus only on cases where there are slight diffe rences in transmissibility\, resulting in limited competition between the strains.  One can expect that if one strain has a substantial competitive advantage over others\, the competitive exclusion principle would come in to play\, overriding coexistence mechanisms.\n \nIn this work\, we use as ymptotic methods to test the validity of the exclusion principle at an ult imate limit in which one strain has a vast competitive advantage over the other strains.   We show that in the case of partial cross-immunity\, su ch systems have a stable multi-species endemic equilibrium. Thus\, the com petitive exclusion principle does not hold beyond the established case of complete immunity.   Furthermore\, we have provided examples of cases wh ere the less transmissible strain is driven to extinction when the other s train is moderately more transmissible\, but not when the other strain is significantly more transmissible.\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:210@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240319T133000 DTEND;TZID=Asia/Jerusalem:20240319T143000 DTSTAMP:20250109T093659Z URL:https://math.technion.ac.il/events/recovering-evolutionary-trees-via-s pectral-graph-theory/ SUMMARY:Recovering evolutionary trees via spectral graph theory DESCRIPTION:Lecturer:Ariel Jaffe\n Location:Amado 814\n Recovering a tree m odel that accurately represents the developmental process of high-dimensio nal data is a key challenge in multiple domains. A common setting is the l atent tree model\, where the task is to infer the tree structure given onl y observations of its terminal nodes. For example\, in phylogenetics\, the evolutionary history of a set of organisms is inferred by their nucleotid e or protein sequences.\n\nIn our work\, we incorporate results from spect ral graph theory to develop novel methods for recovering latent tree model s. We show that the tree structure is strongly related to the spectral pro perties of a fully connected graph defined over the terminal nodes of the tree. This relation forms the theoretical basis of two new methods: (i) sp ectral neighbor-joining\, where subsets of nodes are iteratively merged to form the full tree\, and (ii) spectral top-down recovery\, where the term inal nodes are iteratively partitioned into smaller subsets. Comparing our approach to several competing methods\, we show that in many settings\, s pectral methods have stronger theoretical guarantees and work better in pr actice.\n\n \;\n\nThe talk is based on the following two papers:\n\n "Spectral neighbor joining for reconstruction of latent tree models"\, A J affe\, N Amsel\, Y Aizenbud\, B Nadler\, JT Chang\, Y Kluger\nSIAM journal on mathematics of data science (2021)\n\n \;\n\n "Spectral top-down recovery of latent tree models"\, Y Aizenbud\, A Jaffe\, M Wang\, A Hu\, N Amsel\, B Nadler\, JT Chang\, Yuval Kluger\n\nInformation and Inference: A Journal of the IMA (2023)\n\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:219@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240402T133000 DTEND;TZID=Asia/Jerusalem:20240402T143000 DTSTAMP:20250109T093700Z URL:https://math.technion.ac.il/events/on-landis-conjecture-for-schrodinge r-operators-on-graphs/ SUMMARY:On Landis conjecture for Schrödinger operators on graphs DESCRIPTION:Lecturer:Matthias Keller\n Location:Amado 814\n Landis conjectu re is concerned with growth bounds for harmonic functions of Schrödinger operators whose potential in absolute value is bounded by one. We study th is problem for real potentials on general graphs and in the Euclidean latt ice in particular. (This is joint work with Ujjal Das and Yehuda Pinchover .)\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:241@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240528T133000 DTEND;TZID=Asia/Jerusalem:20240528T143000 DTSTAMP:20250109T093742Z URL:https://math.technion.ac.il/events/oscillation-and-asymptotic-properti es-of-second-order-delay-equations-new-trends/ SUMMARY:Oscillation and Asymptotic Properties of Second Order Delay Equatio ns: New Trends DESCRIPTION:Lecturer:Alexander Domoshnitsky\, Ariel University\n Location:A mado 819\n \n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 819 END:VEVENT BEGIN:VEVENT UID:247@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240604T133000 DTEND;TZID=Asia/Jerusalem:20240604T143000 DTSTAMP:20250109T093742Z URL:https://math.technion.ac.il/events/regularity-theory-and-fully-nonline ar-free-boundary-problems/ SUMMARY:Regularity theory and fully nonlinear free boundary problems DESCRIPTION:Lecturer:David Stolnicki\n Location:Amado 814\n Our study delve s into the existence and regularity of  transmission problems represented by the equation $\\sigma_i (|Du|) F(D^2u)=f$\, concerning regions where $ u >\; 0$ and $u <\; 0$. This work solves  the question on existence o f solutions for the problem by utilizing a comparison principle for an aux iliary problem and Perron's Method.  Furthermore\, we establish\n$C^1$ re gularity estimates for the interior as a subsequent step in our analysis.\ n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:242@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240618T133000 DTEND;TZID=Asia/Jerusalem:20240618T141500 DTSTAMP:20250109T093743Z URL:https://math.technion.ac.il/events/network-modeling-of-porous-media-tr ansport/ SUMMARY:Network modeling of porous media transport DESCRIPTION:Lecturer:Lou Kondic\, New Jersey Institute of Technology\n Loca tion:Amado 814\n \n\n\nWe model porous medium as a random pore network and focus on the influence of the medium internal structure on its flow and a dsorptive behavior\, see Figure 1.  A particular focus is modeling suspe nsion flow\, where the particles adsorb on the pore walls.  We first for mulate the governing equations of fluid flow on a general network. Then\, we model adsorption by imposing an advection equation with a sink term on each pore and study the influence of network parameters on the flow and tr ansport\; see the papers below for examples of our recent works. The talk will focus on developing better understanding of the connection between th e topology of the medium (pore network) and the flow properties. The chall enging aspect of understanding and quantifying evolving pore network topol ogy is addressed by using topological methods that allow for simplified ne twork descriptions\, both regarding their static and their dynamic propert ies. For this purpose\, we use tools based on persistent homology. These t ools enable us to connect topology\, transport\, and adsorption as the ess ential step toward designing porous media of desired properties.\n\nRefere nces: Gu\, Kondic\, Cummings\, SIAM J. Appl. Math. 82\, 950 (2022)\; J. Me m. Sci. 657\, 120668 (2022)\; Phys. Rev. Fluids 8\, 044502 (2023)\n\n\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:260@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240625T133000 DTEND;TZID=Asia/Jerusalem:20240625T143000 DTSTAMP:20250109T093744Z URL:https://math.technion.ac.il/events/wavelets-are-all-you-need-for-autor egressive-image-generation/ SUMMARY:Wavelets are all you need for autoregressive image generation DESCRIPTION:Lecturer:Shai Dekel\n Location:Amado 814\n \n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:270@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240702T133000 DTEND;TZID=Asia/Jerusalem:20240702T143000 DTSTAMP:20250109T093745Z URL:https://math.technion.ac.il/events/the-g-invariant-graph-laplacian/ SUMMARY:The G-invariant graph Laplacian DESCRIPTION:Lecturer:Eitan Rosen\n Location:Amado 814\n Graph Laplacian met hods are a central tool for analyzing data lying on a manifold. In this ta lk\, we introduce the G-invariant graph Laplacian (G-GL)\, designed for ma nifold data closed under the action of a known matrix (Lie) group G. We pr esent the G-GLs key properties and show that exploiting the symmetries in the data has significant theoretical and computational advantages. We then show how to use the G-GLs eigendecomposition to construct G-equivariant a nd G-invariant embeddings of the data\, related to certain random walks ov er the data. Finally\, we discuss applications to the cryo-electron micros copy class averaging problem.\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:295@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240730T133000 DTEND;TZID=Asia/Jerusalem:20240730T143000 DTSTAMP:20250109T093816Z URL:https://math.technion.ac.il/events/eigenfunction-expansion-and-subordi nacy-theory-of-jacobi-operators-on-z/ SUMMARY:Eigenfunction expansion and subordinacy theory of Jacobi operators on Z DESCRIPTION:Lecturer:Netanel Levi\n Location:Amado 814\n If T is a self-adj oint operator on a finite-dimensional Hilbert space H\, then any vector v in H can be\nwritten as a linear combination of eigenfunctions of T. The c oncept of eigenfunction expansion is the\ngeneralization of this fact to i nfinite-dimensional Hilbert spaces. In this talk\, we introduce this conce pt\nin the specific case of Jacobi operators on \\ell^2(\\mathbb{Z}). In p articular\, we will present the connection between\neigenfunction expansio n and subordinacy theory\, which relates asymptotic properties of solution s to\nthe eigenvalue equation to singularity properties of spectral measur es.\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:302@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20240806T133000 DTEND;TZID=Asia/Jerusalem:20240806T143000 DTSTAMP:20250109T093817Z URL:https://math.technion.ac.il/events/landis/ SUMMARY:On the Landis conjecture DESCRIPTION:Lecturer:Ujjal Das\n Location:Amado 814\n \n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:314@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20241203T133000 DTEND;TZID=Asia/Jerusalem:20241203T143000 DTSTAMP:20250109T093839Z URL:https://math.technion.ac.il/events/novel-patterns-in-a-class-of-klein- gordon-equations/ SUMMARY:Novel Patterns in a Class of Klein-Gordon Equations DESCRIPTION:Lecturer:Philip Rosenau\n Location:Amado 814\n \n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:354@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20241210T133000 DTEND;TZID=Asia/Jerusalem:20241210T143000 DTSTAMP:20250109T093840Z URL:https://math.technion.ac.il/events/deep-learning-on-sets-and-distribut ions-via-euclidean-embeddings/ SUMMARY:Deep Learning on Sets and Distributions via Euclidean Embeddings DESCRIPTION:Lecturer:Tal Amir\n Location:Amado 814\n Machine learning probl ems involving non-Euclidean data such as sets\, point clouds\, and graphs\ , require theoretical tools and practical architectures that respect the i nherent symmetries of the data. Efficient methods to embed such data into Euclidean space are thus valuable tools\, as they allow the application of neural architectures to these non-Euclidean domains.\n\nIn this talk\, I will present two novel methods for embedding sets and distributions over R ^d into Euclidean space: one based on moments of shallow MLPs with analyti c activations\, and the other on Fourier sampling of projected quantile fu nctions. I will discuss their theoretical properties (injectivity\, bi-Lip schitzness) and demonstrate their efficient application in practical learn ing tasks. Additionally\, I will present two impossibility results: (1) em beddings based on moments cannot be bi-Lipschitz on multisets\, and (2) no embedding can be bi-Lipschitz on distributions.\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:353@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem;VALUE=DATE:20241223 DTEND;TZID=Asia/Jerusalem;VALUE=DATE:20241224 DTSTAMP:20250109T093841Z URL:https://math.technion.ac.il/events/%d7%9b%d7%a0%d7%a1-%d7%90%d7%91%d7% a8%d7%91%d7%a0%d7%90%d7%9c-%d7%9c%d7%9e%d7%aa%d7%9e%d7%98%d7%99%d7%a7%d7%9 4-%d7%a9%d7%99%d7%9e%d7%95%d7%a9%d7%99%d7%aa-%d7%95%d7%97%d7%99%d7%a9%d7%9 5%d7%91%d7%99/ SUMMARY:כנס אברבנאל למתמטיקה שימושית וחישוב ית DESCRIPTION:Lecturer:\n Location:\n כנס אברבנאל למתמטיקה שימושית וחישובית הוא הכנס המרכזי בתחום ב ישראל\, ומתקיים אחת לשנתיים באוניברסיטת תל אביב. השנה ייערך הכנס ביום שני 23 בדצמב ר 2024\, ואנו מזמינים אתכם להשתתף בכנס!\nזהו אירוע מיוחד המאפשר להיחשף להרצאות מגוונ ות בנושאים עדכניים במתמטיקה שימושית\, שי ועברו על ידי חברי סגל מובילים מאוניברסיט אות שונות בארץ. בנוסף\, הכנס מהווה הזדמנו ת מצוינת לפגוש סטודנטים וסטודנטיות ממוס דות אקדמיים אחרים.\nחלק ייחודי בכנס השנה הוא הרצאות בזק (Flash Talks) שבהן סטודנטים יצי גו את נושאי המחקר שלהם בפורמט קצר של 3–5 ד קות. פורמט זה\, הנפוץ בכנסים בינלאומיים\, מאפשר שיתוף רעיונות בצורה מרוכזת ומעורר ת עניין.\nתכנית הכנס המלאה זמינה בקישור:\nh ttps://sites.google.com/view/2024abarbanelworkshop/program\nהכנס לל א תשלום\, אך נדרשת הרשמה מראש. להרשמה ולצי ון עניין במתן הרצאת בזק\, היכנסו לקישור ה בא\n: https://sites.google.com/view/2024abarbanelworkshop/registration\n \n CATEGORIES:PDE and Applied Mathematics,Seminars END:VEVENT BEGIN:VEVENT UID:324@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20250121T133000 DTEND;TZID=Asia/Jerusalem:20250121T143000 DTSTAMP:20250109T092954Z URL:https://math.technion.ac.il/events/so-what-are-stresses-some-mathemati cal-aspects-of-stress-theory/ SUMMARY:So\, what are Stresses? Some Mathematical Aspects of Stress Theory DESCRIPTION:Lecturer:Prof. Reuven Segev\, Department of Mechanical Engineer ing\, BGU\n Location:Amado 814\n Stress tensors are used in strength analy sis of structures\, fluid dynamics\, electromagnetism\, and general relati vity. Yet\, from the theoretical point of view\, the stress tensor object is not a primitive one. It is derived on the basis of some physically moti vated mathematical assumptions. Following a short introduction to the fund amentals of the classical theory\, we will present a formulation of stress theory on general differentiable manifolds\, devoid of any particular Rie mannian metric. The basic mathematical object is the configuration space-- -the Banach manifold of k-times continuously differentiable sections of a fiber bundle over a compact base manifold---the material body. The choice of topology is natural so that the set of embeddings of the body manifold in a space manifold is open in the manifold of all mappings. Forces are de fined to be elements of the cotangent bundle of the configuration space an d their action on virtual velocities---elements of the tangent bundle---is interpreted physically as virtual power. Stresses and hyper-stresses emer ge naturally from a representation theorem as measures\, valued in the dua l bundles of some jet bundles\, which represent forces.\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:334@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20250128T133000 DTEND;TZID=Asia/Jerusalem:20250128T143000 DTSTAMP:20250109T092954Z URL:https://math.technion.ac.il/events/surface-diffusion-some-new-results- and-approaches/ SUMMARY:Surface Diffusion: Some new results and approaches DESCRIPTION:Lecturer:Amy Novick-Cohen\n Location:Amado 814\n \n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:411@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20250401T133000 DTEND;TZID=Asia/Jerusalem:20250401T143000 DTSTAMP:20250327T061753Z URL:https://math.technion.ac.il/events/dynamical-and-spectral-defects-in-a pproximations-of-substitution-systems/ SUMMARY:Dynamical and spectral defects in approximations of substitution sy stems DESCRIPTION:Lecturer:Lior Tenenbaum\n Location:Amado 814\n abstract is atta ched as a file\, as well as a zoom link.\n\nzoom passcode: 081307\n https: //technion.zoom.us/j/94515734619?pwd=u2tRcV2OHCIdcwR9S3k1Mv5dkHaK8X.1 CATEGORIES:Graduation Seminars,PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:555@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20251125T133000 DTEND;TZID=Asia/Jerusalem:20251125T143000 DTSTAMP:20251119T114538Z URL:https://math.technion.ac.il/events/mathematical-models-for-dense-field -estimation-from-sparse-sensing/ SUMMARY:Mathematical models for dense field estimation from sparse sensing DESCRIPTION:Lecturer:Alon Feldman\n Location:Amado 814\n \nEnvironmental se nsing and modeling occupy a critical intersection of applied mathematics\, data science\, and sustainability. Capturing the spatial and temporal dyn amics of processes such as air pollution\, soil contamination\, and water quality requires transforming sparse and noisy measurements into continuou s\, interpretable fields. Despite notable advances in sensing technologies \, environmental monitoring remains fundamentally limited by sensor covera ge\, economic constraints\, and the nonlinear\, heterogeneous nature of th ese systems.\n\n\n\n\n\n\n\nThis thesis introduces mathematical frameworks for data-driven environmental inference\, addressing the challenge of rec onstructing continuous environmental fields from incomplete observations. It comprises three integrated studies that span the sensing–inference co ntinuum. The first study presents a novel vision-based approach to air qua lity estimation\, demonstrating that physical quantities such as $CO_2$ an d $PM_{2.5}$ can be indirectly inferred from visual cues in roadside image ry using deep convolutional neural networks. The second study develops the Ridiculously Simple Interpolation Method (RSIM)\, a mathematically princi pled yet computationally efficient technique that combines machine learnin g with physics-based digital twin simulations to reconstruct dense polluti on maps from sparse sensor data. The third study generalizes the RSIM fram ework to soil and groundwater contamination\, showing that the same formul ation effectively captures spatial dependencies in subsurface transport de spite discontinuities and heterogeneity in porous media.\n\n\n\n\n\n\n\nTo gether\, these contributions illustrate how mathematical reasoning and dat a-driven modeling can jointly overcome sensing limitations\, offering a sc alable path toward high-resolution\, cross-domain environmental monitoring .\n\n\n\n\n\n\n\n\n\n \;\n\n\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:572@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20251209T133000 DTEND;TZID=Asia/Jerusalem:20251209T143000 DTSTAMP:20251130T122655Z URL:https://math.technion.ac.il/events/singular-structures-in-nematic-liqu id-crystals/ SUMMARY:Singular structures in nematic liquid crystals DESCRIPTION:Lecturer:Dmitry Golovaty (University of Akron)\n Location:Amado 814\n I will focus on recent results on behavior of minimizers of some co mmon models for liquid crystals stemming from the seminal work of Brezis\, Bethuel and Helein on the Ginzburg-Landau theory. In particular\, I will describe a recent result for Ginzburg-Landau theory posed on a singular su rface (a cone) to explore the role of a singularity on a surface in determ ining vortex location. I will also discuss a model for nematic/isotropic p hase transitions in which there emerge both line defects and singular phas e boundaries.\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:584@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20251230T133000 DTEND;TZID=Asia/Jerusalem:20251230T143000 DTSTAMP:20251221T071015Z URL:https://math.technion.ac.il/events/theoretical-guarantees-for-global-s tructure-from-motion/ SUMMARY:Theoretical Guarantees for Global Structure from Motion DESCRIPTION:Lecturer:Gilad Lerman (University of Minnesota)\n Location:Amad o 814\n Recovering 3D geometry from 2D images is a foundational problem in computer vision. Part of the common pipeline in the field has depended on the 45-year-old Random Sample Consensus (RANSAC) algorithm\, which relies on minimal subset sampling—effectively ignoring the bulk of the data du ring model generation. I will present a robust\, global\, and complementar y alternative: the Subspace-constrained Tyler’s Estimator (STE). I will analyze the success of STE by assuming a generalized weak inlier-outlier m odel\, pushing the outlier fraction beyond the threshold where the problem is traditionally proven to be computationally hard. I will show that in t his setting\, provided the initialization satisfies a specific condition\, STE can recover the underlying subspace with high probability. I will als o review additional mathematical ideas to infer global camera poses from r elative ones—specifically the recent Cycle-Sync method—and outline fut ure directions for robust geometric perception.\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:598@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20260127T133000 DTEND;TZID=Asia/Jerusalem:20260127T143000 DTSTAMP:20260108T125433Z URL:https://math.technion.ac.il/events/eigenvalues-and-approximate-eigenva lues-of-the-biharmonic-operator/ SUMMARY:Eigenvalues and approximate eigenvalues of the biharmonic operator. DESCRIPTION:Lecturer:Matania Ben-Artzi\n Location:Amado 814\n The talk will deal mostly with the biharmonic operator on an interval and the optimal a pproximation of its eigenvalues. \nTo this end a suitable discrete ellipt ic calculus is presented \, allowing the use of classical elliptic results (e.g.\, coercivity).\nThe fact that there is no "algorithm" for the first eigenvalue of the biharmonic operator in a square motivates this work.\n\ nBased in part on a joint work with Guy Katriel\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:622@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20260330T133000 DTEND;TZID=Asia/Jerusalem:20260330T143000 DTSTAMP:20260326T171852Z URL:https://math.technion.ac.il/events/mullins-nonlinear-grooving-solution s/ SUMMARY:Mullins’ nonlinear grooving solutions DESCRIPTION:Lecturer:Rawan Tarabeh\n Location:zoom\n The problem of thermal grooving was first proposed by Mullins in 1957. Mullins\nconsidered the m orphological evolution of thin solid surfaces\, namely solid films\,\nand focused primarily on the exterior surface motion with surface diffusion as the\ndominant mass transport process. By assuming a small slope approxima tion\, the\nproblem can be reduced to a linear surface diffusion problem. Mullins showed\nthat there exist self-similar solutions for the linearized problem. In the present\nstudy\, we prove the existence and uniqueness of a self-similar solution for the\noriginal nonlinear Mullins’ problem fo r sufficiently small values of the contact\nangle β. We do so by studying a reduced linear ODE system\, which allows us\nto understand the asymptot ic behavior of the solutions for the corresponding\nlinear homogeneous ope rator\, and then we prove existence of a unique solution\nto the nonlinear problem in closed form by iterations in an appropriate Banach\nspace.\nAd visor: Prof. Amy Novick-Cohen\n\n \;\n\n \;\n https://technion.zoo m.us/j/98140816947 CATEGORIES:Graduation Seminars,PDE and Applied Mathematics,Seminars LOCATION:zoom END:VEVENT BEGIN:VEVENT UID:587@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20260331T133000 DTEND;TZID=Asia/Jerusalem:20260331T143000 DTSTAMP:20251221T074258Z URL:https://math.technion.ac.il/events/the-willmore-energy-and-curvature-c oncentration/ SUMMARY:The Willmore energy and curvature concentration DESCRIPTION:Lecturer: David Padilla-garza \n Location:Amado 814\n We study isometric immersions of a two-dimensional Riemannian surface into three-di mensional Euclidean space. We consider their bending energy\,  i.e.\, the square-integral of their second fundamental form\, which is equivalent to the Willmore functional. We obtain two new lower bounds for this energy\, one in terms of the Gaussian curvature of the surface\, and the other in terms of a Burgers vector---a measure of non-flatness connected to torsion . These new estimates provide optimal blowup rates of the energy when the curvature is concentrated (e.g.\, in a conical geometry). In the more subt le case of dipoles of concentrated curvature\,  we use the Burgers vector estimates to obtain an optimal blowup rate in terms of the size of the sy stem. Our motivation comes from non-Euclidean elasticity\, in which cones and curvature-dipoles play a central role. The lower bounds derived in thi s work directly yield lower bounds for the elastic energy of thin elastic sheets. The derivation of the curvature-based lower bound involves an isop erimetric inequality for framed loops\, which we believe to be of independ ent interest.\n\n \;\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:651@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20260526T133000 DTEND;TZID=Asia/Jerusalem:20260526T143000 DTSTAMP:20260519T075329Z URL:https://math.technion.ac.il/events/from-kinetic-pdes-to-incompressible -fluids/ SUMMARY:From kinetic PDEs to incompressible fluids DESCRIPTION:Lecturer: Immanuel Ben-Porath\n Location:Amado 814\n The quasi- neutral limit is a hydrodynamical limit offering a derivation of a PDE of incompressible fluid from a kinetic PDE . In the specific context of the 2 D Vlasov-Poisson equation\, the quasi-neutral limit leads to the 2D incomp ressible Euler equation. The purpose of this talk would be to concentrate on the modulated energy method\, as developed by Brenier\, in order to rig orously justify this limit. Well-posedness of the underlying equations wil l be discussed as well. The talk will be mostly expositional in nature\, w ith some indication towards contemporary research themes.\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:620@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20260609T133000 DTEND;TZID=Asia/Jerusalem:20260609T143000 DTSTAMP:20260420T064838Z URL:https://math.technion.ac.il/events/numerical-solvers-for-wave-propagat ion-problems-reduction-based-multigrid-methods/ SUMMARY:Numerical solvers for wave propagation problems: reduction-based mu ltigrid methods DESCRIPTION:Lecturer:Rachel Yovel\n Location:Amado 814\n The Helmholtz equa tions are the frequency-domain wave equations. Both the acoustic and the e lastic versions of this equation are challenging numerically: the resultin g linear system is indefinite\, which leads to failure of many iterative s olvers due to the oscillatory nature of the solutions. In some application s\, such as inverse problems and underground imaging\, the elastic version is needed. The elastic Helmholtz operator is known as harder to solve and fewer solvers address it. We show that\, in fact\, the difficulty in solv ing the elastic Helmholtz equation stems from the acoustic one.\nWe develo p a reduction-based approximate-commutator preconditioner for the elastic Helmholtz equation. By utilizing the commutation relations of the underlyi ng continuous differential operators\, we suggest a block-triangular preco nditioner whose diagonal blocks are acoustic operators. Thus\, we enable t he solution of the elastic version using virtually any existing solver for the acoustic version. We prove a sufficient condition for the convergence of our method\, exploring the spectral behavior of the underlying operato rs.  We show that the resulting reduction achieves wavenumber independent convergence. However\, solving the acoustic Helmholtz equation efficientl y is necessary for efficient application of this reduction. Standard multi grid methods fail to converge for high frequency Helmholtz problems\, and the performance of related preconditioners strongly depends on the wavenum ber. Using novel symbol analysis aimed at minimizing numerical dispersion\ , we develop a multigrid method for the acoustic Helmholtz equation\, whic h achieves wavenumber independent convergence\, even in the high-frequency regime. Overall\, our methods form fast and scalable solvers for high-fre quency waves in highly heterogeneous media in 2D and in 3D.\n\nThe results presented in this lecture were obtained during the phd thesis of the spea ker\, under the supervision of Eran Treister\n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:Amado 814 END:VEVENT BEGIN:VEVENT UID:650@math.technion.ac.il DTSTART;TZID=Asia/Jerusalem:20260616T133000 DTEND;TZID=Asia/Jerusalem:20260616T143000 DTSTAMP:20260514T100518Z URL:https://math.technion.ac.il/events/oscillations-in-strongly-2-cooperat ive-systems-and-their-applications-in-systems-biology/ SUMMARY:Oscillations in strongly 2-cooperative systems and their applicatio ns in systems biology DESCRIPTION:Lecturer:Ram Kats\n Location:amado 814\n The emergence of susta ined oscillations (via convergence to periodic orbits) in high-dimensional nonlinear dynamical systems is a highly non-trivial question with importa nt applications in systems biology\, including the understanding of bio-mo lecular oscillators ruling cell life-cycle and metabolism\, as well as cir cadian rhythms in hormone secretion\, body temperature and metabolic funct ions. In systems biology\, the mechanism underlying such widespread oscill atory biological motifs is still not fully understood. From a mathematical perspective\, the study of sustained oscillations is comprised of two par ts: (i) showing that at least one periodic orbit exists and (ii) studying the stability of periodic orbits and/or characterising the initial conditi ons which yield solutions that converge to periodic trajectories. In this talk we will focus on a specific class of nonlinear dynamical systems that are strongly 2-cooperative. Employing results from the theory of cones of rank k\, the spectral theory of totally positive matrices and Perron-Frob enius theory\, we will show that strongly 2-cooperative systems admit an e xplicit set of initial conditions of positive measure\, such that every so lution emanating from this set converges to a periodic orbit. We will furt her demonstrate our results using the n-dimensional Goodwin oscillator and a 4-dimensional biological oscillator based on RNA–mediated regulation. \n CATEGORIES:PDE and Applied Mathematics,Seminars LOCATION:amado 814 END:VEVENT BEGIN:VTIMEZONE TZID:Asia/Jerusalem X-LIC-LOCATION:Asia/Jerusalem BEGIN:DAYLIGHT DTSTART:20230324T030000 TZOFFSETFROM:+0200 TZOFFSETTO:+0300 TZNAME:IDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20231029T010000 TZOFFSETFROM:+0300 TZOFFSETTO:+0200 TZNAME:IST END:STANDARD BEGIN:DAYLIGHT DTSTART:20240329T030000 TZOFFSETFROM:+0200 TZOFFSETTO:+0300 TZNAME:IDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20241027T010000 TZOFFSETFROM:+0300 TZOFFSETTO:+0200 TZNAME:IST END:STANDARD BEGIN:DAYLIGHT DTSTART:20250328T030000 TZOFFSETFROM:+0200 TZOFFSETTO:+0300 TZNAME:IDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20251026T010000 TZOFFSETFROM:+0300 TZOFFSETTO:+0200 TZNAME:IST END:STANDARD BEGIN:DAYLIGHT DTSTART:20260327T030000 TZOFFSETFROM:+0200 TZOFFSETTO:+0300 TZNAME:IDT END:DAYLIGHT END:VTIMEZONE END:VCALENDAR