Past events

  • Compressed sensing concerns the recovery of signals and images from seemingly incomplete data sets. Introduced nearly a decade ago, it has since become an intensive area of research in applied mathematics, engineering and computer science. However, many practical problems in which compressed sensing is applied, e.g. imaging, are not fully explained by existing theory. In this talk I will present a new framework for compressed sensing that seeks to bridge this gap. This framework is based on replacing some standard principles of compressed sensing with new local notions; specifically, sparsity in levels, local coherence in levels and multilevel random subsampling. I will demonstrate a series of near-optimal recovery guarantees based on these local concepts that explains the effectiveness of compressed sensing in such applications. Moreover, this framework is not just useful in understanding existing compressed sensing approaches. In the final part of the talk I will demonstrate how leveraging local sparsity through appropriately-designed locally incoherent sensing matrices leads to substantially improved compressed sensing techniques in a range of other applications.

  • Abstract: The term “curse of dimensionality” refers to increases in the dimensionality of model spaces that result in undesirable increases in data sparsity, such that model parameters are no longer sufficiently constrained by the data. Although the term is usually employed in combinatorics, machine learning, and data mining, it is also directly relevant for many problems in exploration geophysics. The most obvious applications are 3D tomographic inversions, which typically include very large numbers of unknowns.
    There is a further “curse of dimensionality” and related data sparsity that may impede many geophysical investigations: 3D surveys typically involve the acquisition of data using only a 2D array of sensors distributed across the Earth’s surface. As a consequence, procedures for imaging the subsurface are missing data recorded in the third dimension, depth. Similar problems affect 2D inversions of (1D) profile data.
    Computational problems that need to be overcome in large-scale tomographic inversions are additional issues associated with the “curse of dimensionality.” In particular, the rapidly emerging field of realistic 3D full-waveform inversions of elastic and anisotroic data is hitting the limits of current computer facilities. Seemingly ever increasing computing power will undoubtedly be beneficial for such endeavors. Nevertheless, suitable model parameterizations that offer appropriate spatial resolution while keeping the inversion problem computationally tractable will continue to be critical elements of any high dimension inversion endeavor.
    Because of the large computational costs and the difficulties to cover extensive areas with geophysical sensors in complicated terrain, many land surveys continue to involve data acquisition along profiles. Such surveys will play a significant role for the foreseeable future. When solving the associated 2D inversion problems, the “curse of dimensionality” strikes again. The underlying 2D assumption that subsurface properties and topography do not change in the third dimension, that is, perpendicular to the tomographic plane, is often unjustified.
    The problem of data sparsity can be partially alleviated by employing optimized experimental design and optimized data parameterization approaches. These techniques identify experimental configurations and data representations that optimize data information content and resultant models in a cost-effective manner.
    In this lecture, I will illustrate the “curse of dimensionality” by means of several examples from near-surface geophysics. I will present a variety of options for addressing the related problems, including experimental design techniques and optimized model parameterization strategies. I will also discuss problems and remedies related to out-of-plane features in 2D elastic full‑waveform inversions.
  • Inaugural Full-Waveform Inversion Workshop, Brazil , Monday, August 31, 2015 (All day) - Tuesday, September 1, 2015 (All day)
    We invite our colleagues in Latin America to attend the Inaugural Full-Waveform Inversion workshop. We are pleased to present this workshop as part of the International Inversion Initiative and in collaboration with partners Universidade Federal do Rio Grande do Norte, SENAI CIMATEC Supercomputing Centre Bahia, Imperial College London, the University of British Columbia (Canada) and BG Brasil. Please see below for program information and abstracts.
    Liacir dos Santos Lucena, Universidade Federal do Rio Grande do Norte, Brazil
    Mike Warner, Imperial College London, UK
    Felix J. Herrmann, University of British Columbia, Canada
    Description: Full waveform inversion (FWI) is now a widely established theory of geophysical imaging, yet not fully applied in practise due to limitations of seismic acquisitions, seismic simulations, inversion algorithms, and computational restrictions. This workshop will explore theoretical and practical applications of FWI theory through a series of Tutorials using field datasets, and Lectures discussing current leading industry practise, the latest research developments on theoretical and computational fronts, and techniques of data preconditioning and computational optimization. Researchers, students and industry practitioners are invited to attend.
    For more information see website:
    Monday August 31: Tutorials
    Tuesday Sept 1: Public Symposium
    The event will be hosted at the Ocean Palace Resort, Ponta Negra, Natal:
    Please call +55 (84) 3220-4144 to reserve.
    Workshop Fees:
    Students = US$25,00 / R$85,00
    Academic = US$75,00 / R$250,00
    Professionals = US$150,00 / R$500,00
    To register, please contact:
  • MS-We-D-07-3
    Wavefield Reconstruction Inversion
    with convex constraints
    Herrmann, Felix (UBC-SLIM)
    Abstract: During thistalk, we discuss how to exploit the special structure of Wavefield Recons
    truction Inversion (WRI) to include convex bound and total-variation constr
    aints in a computationally feasible manner. The resulting method shows prom
    ising results on challenging models that include high-velocity high-contras
    t inclusions such as salt or basalt. This is joint work with Ernie Esser wh
    o will be dearly missed.

  • 2:00 pm, ICCSX836, 2366 Main, UBC Campus. The theory of c
    ompressive sensing (CS) has opened up new opportunities in
    the field of op
    tical imaging. However, its implementation is often not
    and involves design approaches that differ significantly
    from the conventi
    onal ones. Here we try to overview systematically the main
    challenges invo
    lved in the practical application of CS theory for imaging
    and present som
    e solution to overcome them. The overview is supplemented
    with examples ofCS techniques that we have developed for progressive
    imaging acquisition
    , motion tracking, spectrometry, hyperspectral imaging
    and for digital h
    olography. We conclude with a list of open questions. The
    questions range
    from theoretical to computational ones. Answers to these
    questions will en
    rich significantly the optical designer's toolkit.

    Adrian Stern (M'09)received the B.Sc., M. Sc. (cum laude) and Ph.D. degrees
    from Ben-GurionUniversity of the Negev, Israel, in 1988, 1997, and 2002
    , all in electrical and computer engineering. He is an Associate
    r at Electro-Optical Engineering department at Ben-Gurion University
    in Is
    rael. During the years 2002-2004 he was a postdoc fellow at University
    Connecticut. During 2007-2008 he served as senior research and algorithm
    pecialist for GE Molecular Imaging, Israel. Currently, during his
    ical leave, he is holding a visiting teaching and scholar position at
    sachusetts Institute of Technology. His current research interests
    include3D imaging, compressive imaging and computational imaging. Dr. Stern
    a Fellow of SPIE, member of IEEE and OSA. He serves as an associate
    r for Optics Express journal, and has served as guest Editor for
    IEEE/OSAJournal on Display Technology.

  • 2015 SEG Summer Research Workshop “FWI Applications from Imagin
    g to Reservoir Characterization”
    eetings/houston2015 Abstract: Most approaches to full-waveform inversion r
    ely on eliminating PDE-constraints. While this leads to computationally fea
    sible implementations, it increases reliance on starting models. Differen
    t extensions of the search space have been proposed that aim to overcome th
    is problem by fitting observations even for poor starting models. Examples
    ninclude extended velocity models and Wiener filters, which when combined
    with WEMVA-like focusing, are less
    susceptible to cycle skipping. Wavefie
    ld Reconstruction Inversion, which alternates between minimizing its objec
    tive w.r.t. the wavefields and the velocity model, by design also fits thedata but differs because it uses the wave-equation itself in combination w
    ith convex constraints to update the model parameters. The resulting formal
    ism is intuitive and yields encouraging results on complex models where con
    ventional FWI may fail.