Tuesday, September 28, 2010

CS: LVA 2010 on Twitter, More ECG, CS on Manifolds, Bayesian PCA, Job Postings at SLIM

Prasad Sudhakar just sent me the following:
Dear Igor,

I am writing to inform you that the 9th international conference on latent variable analysis and signal separation is kicking off today at St. Malo, France.

Here is the conference website: http://lva2010.inria.fr/

There are two sessions on sparsity: Wednesday morning and afternoon.

Also, I will try to tweet the highlights of the proceedings (mostly the sparsity sessions and plenary talks) at http://twitter.com/lva2010

Could you please mention this on your blog?

Best regards,
Prasad Sudhakar


Thanks Prasad, I am now following @LVA2010 on the the Twitter. @LVA2010 is also, now, on the compressed-sensing list on Twitter as well (let me know if you want to be added to it).

Following up on the The EPFL Real-Time Compressed Sensing-Based Personal Electrocardiogram Monitoring System here is: Implementation of Compressed Sensing in Telecardiology Sensor Networks by Eduardo Correia Pinheiro , Octavian Adrian Postolache, and Pedro Silva Girão. The abstract reads:
Mobile solutions for patient cardiac monitoring are viewed with growing interest, and improvements on current implementations are frequently reported, with wireless, and in particular, wearable devices promising to achieve ubiquity. However, due to unavoidable power consumption limitations, the amount of data acquired, processed, and transmitted needs to be diminished, which is counterproductive, regarding the quality of the information produced.
Compressed sensing implementation in wireless sensor networks (WSNs) promises to bring gains not only in power savings to the devices, but also with minor impact in signal quality. Several cardiac signals have a sparse representation in some wavelet transformations. The compressed sensing paradigm states that signals can be recovered from a few projections into another basis, incoherent with the first. This paper evaluates the compressed sensing paradigm impact in a cardiac monitoring WSN, discussing the implications in data reliability, energy management, and the improvements accomplished by in-network processing.

Nonparametric Bayesian methods are employed to constitute a mixture of low-rank Gaussians, for data x 2 RN that are of high dimension N but are constrained to reside in a low-dimensional subregion of RN. The number of mixture components and their rank are inferred automatically from the data. The resulting algorithm can be used for learning manifolds and for reconstructing signals from manifolds, based on compressive sensing (CS) projection measurements. The statistical CS inversion is performed analytically. We derive the required number of CS random measurements needed for successful reconstruction, based on easily computed quantities, drawing on block–sparsity properties. The proposed methodology is validated on several synthetic and real datasets.

Bayesian Robust Principal Component Analysis by Xinghao Ding, Lihan He, and Lawrence Carin. The abstract reads:
A hierarchical Bayesian model is considered for decomposing a matrix into low-rank and sparse
components, assuming the observed matrix is a superposition of the two. The matrix is assumed
noisy, with unknown and possibly non-stationary noise statistics. The Bayesian framework infers an approximate representation for the noise statistics while simultaneously inferring the low-rank and sparse-outlier contributions; the model is robust to a broad range of noise levels, without having to change model hyperparameter settings. In addition, the Bayesian framework allows exploitation of additional structure in the matrix. For example, in video applications each row (or column) corresponds to a video frame, and we introduce a Markov dependency between consecutive rows in the matrix (corresponding to consecutive frames in the video). The properties of this Markov process are also inferred based on the observed matrix, while simultaneously denoising and recovering the low-rank and sparse components. We compare the Bayesian model to a state-of-the-art optimization-based implementation of robust PCA; considering several examples, we demonstrate competitive performance of the proposed model.
The code implementing this Bayesian PCA is here.

I usually don't post about graduate studentships but the sheer size of this announcement makes it worthwhile for a post. Felix Herrmann just sent me this announcement for three postdocs and ten graduate studentships at UBC. The jobs are posted in the compressive sensing jobs page. They are also listed below:

  • September 27th, 2010, Three postdoctoral positions at the Seismic Laboratory for Imaging and Modeling (SLIM) of the University of British Columbia, Vancouver, BC, Canada



    Project:
    DNOISE is a 5-year NSERC and industry-funded project for research in seismic data acquisition, processing, and imaging. Our interdisciplinary approach builds on recent developments in compressive sensing, large-scale optimization, and full-waveform inversion from severely sub-sampled data. The project includes 10 graduate students, 3 postdocs, and a research associate. The postdoctoral positions, under supervision Felix J. Herrmann (Earth and Ocean Sciences), Ozgur Yilmaz (Mathematics), and Michael P. Friedlander (Computer Science), are available immediately. 

    Description:
    The aim of the DNOISE project is to design the next generation of seismic imaging technology to address fundamental issues related to the quality and cost of seismic data acquisition, the ability to invert exceedingly large data volumes, and the capacity to mitigate non-uniqueness of full-waveform inversion.

    You will be part of a dynamic interdisciplinary research group and will present your research at international conferences and to industry. You will also be involved in industry collaborations that include internships and projects on real field data. You will have extensive contact with graduate students, fellow postdocs, and faculty. We seek excellence in any of a wide variety of areas, spanning from theory, algorithm design, to concrete software implementations to be applied to field data. SLIM has state-of-the-art resources, including a 288 CPU cluster, Parallel Matlab, and seismic data-processing software.

    Successful candidates will have a PhD degree obtained in 2008 or later in geophysics, mathematics, computer science, electrical engineering, or a related field, with a strong achievement record in at least one of the following areas: seismic imaging, inverse problems, PDE-constrained optimization, signal processing, sparse approximation and compressed sensing, convex optimization, and stochastic optimization. Earlier PhDs will be considered where the research career has been interrupted by circumstances such as parental responsibilities or illness. UBC hires on the basis of merit, and is committed to employment equity. Positions are open to individuals of any nationality.

    Compressive seismic-data acquisition. Development of practical seismic acquisition scenarios, sigma-delta quantization, and experimental design for seismic inversion.
    Full-waveform inversion. Development of PDE-constrained optimization algorithms that deal with large data volumes and that remedy the non-uniqueness problem.
    Large-scale optimization. Development of optimization algorithms and software for sparse approximation and problems with PDE-constraints.

    About UBC and Vancouver:
    The University of British Columbia, established in 1908, educates a student population of 50,000 and holds an international reputation for excellence in advanced research and learning. Our campus is 30 minutes from the heart of downtown Vancouver, a spectacular campus that is a 'must-see' for any visitor to the city -- where snow-capped mountains meet ocean, and breathtaking vistas greet you around every corner. 

    How to apply: Applicants should submit a CV, a list of all publications, and a statement of research, and arrange for three or more letters of recommendation to be sent to Manjit Dosanjh (MDOSANJH@eos.ubc.ca). All qualified candidates are encouraged to apply; however, Canadians and Permanent Residents will be given priority. For more information, see: http://slim.eos.ubc.ca
  • September 27th, 2010, Ten graduate students at the Seismic Laboratory for Imaging and Modeling (SLIM) of the University of British Columbia



    Project:
    DNOISE is a 5-year NSERC and industry-funded project for research in seismic data acquisition, processing, and imaging. Our interdisciplinary approach builds on recent developments in compressive sensing, large-scale optimization, and full-waveform inversion from severely sub-sampled data. The project includes 10 graduate students, 3 postdocs, and a research associate. Subject to admission by the faculty of graduate studies, prospective students can start as early January 1st, 2011 under supervision of Felix J. Herrmann (Earth and Ocean Sciences), Ozgur Yilmaz (Mathematics), or Michael P. Friedlander (Computer Science).

    Description:

    The aim of the DNOISE project is to design the next generation of seismic imaging technology to address fundamental issues related to the quality and cost of seismic data acquisition, the ability to invert exceedingly large data volumes, and the capacity to mitigate non-uniqueness of full-waveform inversion.

    You will be part of a dynamic interdisciplinary research group and will present your research at international conferences and to industry. You will also be involved in industry collaborations that include internships and projects on real field data. 
    You will have extensive contact with fellow graduate students, postdocs, and faculty. We seek excellence in any of a wide variety of areas, spanning theoretical, algorithmic, and software development. A good background in mathematics is important.

    The graduate funding is for two years for MSc students and four years for PhD students and includes generous funding for travel. PhD student funding includes a tuition waiver. 

    Successful candidates will join the PhD or MSc programs in one of the following departments at UBC: Earth and Ocean Sciences, Mathematics, or Computer Science. At SLIM you will have state-of-the art resources available, including a 288 CPU cluster, Parallel Matlab, and seismic data-processing software. 

    About UBC and Vancouver:
    The University of British Columbia, established in 1908, educates a student population of 50,000 and holds an international reputation for excellence in advanced research and learning. Our campus is 30 minutes from the heart of downtown Vancouver, a spectacular campus that is a 'must-see' for any visitor to the city -- where snow-capped mountains meet ocean, and breathtaking vistas greet you around every corner. 

    How to apply: Please send a CV, cover letter, and  academic transcripts to Manjit Dosanjh (MDOSANJH@eos.ubc.ca). For more information, see: http://slim.eos.ubc.ca


Liked this entry ? subscribe to the Nuit Blanche feed, there's more where that came from

No comments:

Printfriendly