Seeking four new members for the international SPARS Steering Committee

Remi Gribonval just sent me the following:

Dear Igor,
Would you mind sharing this call with the readers of Nuit Blanche ?
All the best,
Remi.

 Sure Remi, here it is:

Dear colleagues, 

We are seeking up to four new members of the community to join the international SPARS Steering Committee, as part of its renewal process.
The SPARS Steering Committee was created during the SPARS 2009 workshop (Signal Processing with Adaptive/Sparse Representations) in St-Malo, France.
Its main purpose is to provide a framework to manage forthcoming SPARS workshops.
Members of the Steering Committee will be expected to: have an active interest in the field; have presented at and/or attended previous workshops; and contribute their experience and ideas for the benefit of the research community.
The next meeting of the SPARS Steering Committee, including the newly elected members, will take place during SPARS 2013 (July 8-11 2013, EPFL, Lausanne, Switzerland.http://spars2013.epfl.ch/).
If you wish to nominate yourself or a colleague, please send the following to me, Remi Gribonval (remi.gribonval@inria.fr), as attachments to an email (PDF preferred):
(1) A statement of qualification for the nominee (up to 1 page, CV style)
(2) A Letter of Support (up to 2 pages)
Deadline for nominations:
Friday 9 November 2012 (midnight GMT/UTC)
Best wishes,
Remi Gribonval
Chair, international SPARS steering committee
--
Rémi GRIBONVAL, Directeur de Recherche Inria
Projet METISS, IRISA
Campus de Beaulieu
35042 Rennes cedex
France
E-mail: Remi.Gribonval@inria.frWWW   : http://www.irisa.fr/metiss/members/remi

Join our Reddit Experiment, Join the CompressiveSensing subreddit and post there !
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.

Open Access and Post Publication Peer Review

Matthieu just sent me the following:

Dear Igor,
I found a nice video about Open Access Publishing on PhdComics (http://www.phdcomics.com/comics.php?n=1533). At the end, it provides a link to the 6th Open Access Week (http://www.openaccessweek.org/), which is described as:

I think it might be of interest for your Nuit Blanche readers following the Post Peer-Review discussion.
Best Regards, 

 A global event, now in its 6th year, promoting Open Access as a new norm in scholarship and research.

Matthieu  

The post peer-review discussion and open access certainly are dealing with similar issues. Open Access could enable enterprising minds to set up a post peer review process. However most current versions of open access journals currently still rely on pre-publication peer review which is the crux of the assumption we are questioning. In short, the value of a paper is not in the quality of the journal or the very few and possibility uninformed gate-keepers, it is about how that work stands the unrelenting scrutiny of time. Thanks Matthieu.

Join our Reddit Experiment, Join the CompressiveSensing subreddit and post there !
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.

Postdoctoral Researcher position at the Linnaeus Centre ACCESS

Cristian Rojas just sent me the following:

Dear Igor,

My name is Cristian Rojas, and I am an assistant professor at KTH, Stockholm, Sweden. At the ACCESS Linnaeus Centre at KTH, we are announcing a postdoctoral position on sparsity related techniques, and we are looking for someone with experience on compressive sensing, sparse estimation methods or similar areas. As such, the call is quite broad. I hope you could help us to distribute this call. Please find it attached, and at the website:

http://www.kth.se/en/om/work-at-kth/vacancies/postdoctoral-researcher-position-at-the-linnaeus-centre-access-1.342681

Thank you very much.

Best regards,

Cristian R. Rojas

Assistant Professor

Automatic Control Lab and ACCESS Linnaeus Centre

School of Electrical Engineering

KTH - Royal Institute of Technology

SE 100 44 Stockholm

Sweden

Join our Reddit Experiment, Join the CompressiveSensing subreddit and post there !

Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.

Eulerian Video Magnification for Revealing Subtle Changes in the World - implementation -

Remember the Eulerian Video Magnification for Revealing Subtle Changes in the World by Hao-Yu Wu, Michael Rubinstein, Eugene Shih, John Guttag, Frédo Durand, William T. Freeman. I stumbled upon the page again this week. The abstract of the paper reads:

Our goal is to reveal temporal variations in videos that are difficult or impossible to see with the naked eye and display them in an indicative manner. Our method, which we call Eulerian Video Magnification, takes a standard video sequence as input, and applies spatial decomposition, followed by temporal filtering to the frames. The resulting signal is then amplified to reveal hidden information. Using our method, we are able to visualize the flow of blood as it fills the face and also to amplify and reveal small motions. Our technique can run in real time to show phenomena occurring at temporal frequencies selected by the user.

The attendant code is here.

Join our Reddit Experiment, Join the CompressiveSensing subreddit and post there !
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.

Slides: Workshop on "Randomized Numerical Linear Algebra (RandNLA): Theory and Practice"

Haim Avron let me kniow that "... most of the slides of the talks in the RandNLA workshop last Saturday are now posted on the workshop's website". Thanks Haim. And also thanks to Christos Boutsidis, and Petros Drineas and Abhisek Kundu for their  work in setting up the meeting and the page. It definitely bring some visibility to the subject. Here are the presentations and the link to each researcher's page (for more):

• Petros Drineas Welcome and Introduction (slides)
• Michael Mahoney, Tutorial: Theory (and Some Practice) of Randomized Algorithms for Matrices and Data (slides, abstract)
• Nick Harvey , Matrix Concentration and Sparsification (slides, abstract)
• Nikhil Srivastava, Graph Sparsification (slides, abstract)
• Ilse Ipsen, Randomly Sampling from Orthonormal Matrices: Coherence and Leverage Scores (slides, abstract)
• Haim Avron, Randomized Preconditioning (slides, abstract)
• Ioannis Koutis, Solving Laplacian Linear Equations: Theory and Practice (slides, abstract)
• Anastasios Zouzias, Randomized Gossip Algorithms for Solving Laplacian Systems (slides, abstract)
• Christos Boutsidis, Near-Optimal Column-Based Matrix Reconstruction (slides, abstract)
• David Woodruff, Low Rank Approximation and Regression in Input Sparsity Time (slides, abstract)
• Ben Recht, An Incomplete Survey of Matrix Completion (slides, abstract)

Credit Images: NASA/ESA



Join our Reddit Experiment, Join the CompressiveSensing subreddit and post there !
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.

A million here, a million there ....

I am still wrapping my head around that number: 1.5 million pageviews since 2010. What is the real reach out ? for all we know these pageviews could be the results of the Googlebot spider process hitting this site 1.5 million times. Here is a map of the delivery process for these 2317 blog entries

• More than 2000 potential readers are reached using different Feeds
• 440 potential readers by direct e-mail
• 543 followers on Twitter (all entries are announced there)
• 1835 members in the LinkedIn Group on CS
• 447 members in the LinkedIn Group on Advanced Matrix Factorization
• 39 subscribers to the experimental CompressiveSensing subreddit.

What about the actual impact ? As in many areas, this is a difficult question, I don't think I have a good answer, but with regards to :

• Pageviews, one should expect a minimum of 100 pageviews per blog entry. For the maximum (outside of some outliers) one should expect about 500 pageviews. 
• E-mail: this one is difficult as I have no way of tracking how many people clicked on particular items. At the very least, the blog entry received some 400 or so eyeball
• Feeds: According to feedburner, at any one time, 1/4th of the readers looked at a specific entry (that's about 500)
• Videos: the recent video by David Brady has gotten 72 viewers (a 40 minutes video) but from what I have seen, outside of some outliers, from 50 to 500 people can decide to watch a video as a result of being featured here. I have noted some statistics were being featured here actually provided a stepping stone to a much larger audience. Here is a list automatically compiled by YouTube of all the Youtube videos featured on Nuit Blanche.
• Code sharing: I am making a targeted effort at putting code implementations in a single entry in order to provide focus. It works, see Brian King's recent reaction in the comment section featuring his toolbox. 

Use of the blog: besides information downloading and sharing: Some folks are providing feedback:

• The blog itself has reached 53 Google +1s so far, while the Big Picture has 22 and the Advanced Matrix Factorization Page has 12.
• One blog entry has reached 8 Google +1
• This is the second time so far recently, that somebody has used the comment section to provide some anonymous peer review of a paper featured on the blog, I like it very much.

Let's see how those stats will be different when we hit 2 million page views!

Join our Reddit Experiment, Join the CompressiveSensing subreddit and post there !
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.

Random Bits: A course, a talk and a Postdoc Position in Randomized Numerical Linear Algebra, IBM

If you think that I did not cover your stuff this month and want to be counted in the new Nuit Blanche in Review for October, please let me know. 

This fall at Rice there is a new course CAAM 654: Sparse Optimization by Wotao Yin and: Ming Yan. I note their judicious use of Nuit Blanche as one of the links to use for students :-) It might also have been interesting to add both the Big Picture in Compressive Sensing and the Matrix Factorization Jungle Page.

There are some interesting discussions and questions on the LinkedIn Compressive Sesning Group right now. The group now boast more than 1830 members. The LinkedIn Matrix Factorization group has more than 444 members. 

There was an interesting talk this past week at Columbia by the ever interesting Bill Freeman on Leaning Matrix Decomposition Structures. The talk given a year earlier has this abstract:

Many widely used models in unsupervised learning can be viewed as matrix decompositions, where the input matrix is expressed as sums and products of matrices drawn from a few simple priors. We present a unifying framework for matrix decompositions in terms of a context-free grammar which generates a wide variety of structures through the compositional application of a few simple rules. We use our grammar to generically and efficiently infer latent components and estimate predictive likelihood for nearly 1000 structures using a small toolbox of reusable algorithms. Using best-first search over our grammar, we can automatically choose the decomposition structure from raw data by evaluating only a tiny fraction of all models. This gives a recipe for selecting model structure in unsupervised learning situations. The proposed method almost always finds the right structure for synthetic data and backs off gracefully to simpler models under heavy noise. It learns plausible structures for datasets as diverse as image patches, motion capture, 20 Questions, and U.S. Senate votes, all using exactly the same code.

I am impatient to see when it comes out. 
Finally, I found this on the NA-Digest list:

From: Haim Avron <haimav@us.ibm.com> Date: Fri, 19 Oct 2012 11:32:59 -0400 Subject: Postdoc Position, Randomized Numerical Linear Algebra, IBM The High Performance Computing for Analytics group within the Business Analytics and Mathematical Sciences Department at IBM's T.J. Watson Research Center is seeking a Post Doctoral Researcher to work on investigating and implementing randomized numerical linear algebra kernels for distributed computing platforms, with applications to machine learning problems. The candidate is expected to contribute to the development of new ideas and implementations, publish in top-tier journals, and file patent disclosures when appropriate. We are especially interested in candidates who have experience and strong interest in large-scale distributed data analysis with emphasis on linear algebra techniques. The successful candidate will work with an interdisciplinary team of researchers. The candidate must have strong programming capabilities, and have excellent verbal and written skills. Preference may be given to candidates with extensive knowledge of C/C++ along with MPI and multi-threaded programming.  Knowledge of Python is a plus. PhD candidates in Computer Science or Mathematics are preferred. For more information on the requirements, and to apply, see: https://jobs3.netmedia1.com/cp/job_summary.jsp?job_id=RES-0526905 IBM is committed to creating a diverse environment and is proud to be an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to race, color, religion, gender, gender identity or expression, sexual orientation, national origin, genetics, disability, age, or veteran status.

This image was taken by Front Hazcam: Left A (FHAZ_LEFT_A) onboard NASA's Mars rover Curiosity on Sol 77 (2012-10-24 02:36:48 UTC) . 

Image Credit: NASA/JPL-Caltech 
Join our Reddit Experiment, Join the CompressiveSensing subreddit and post there !
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.

Unbiased Risk Estimates for Singular Value Thresholding and Spectral Estimators - implementation -

Just received this from Josh Trzasko

Hi Igor- 

I wanted to highlight some recent work by Emmanuel Candes, Carlos Sing-Long, and myself on unbiased risk estimates for singular value thresholding (SVT). Amongst other things, the models developed in this work can be used to automatically tune SVT-based denoising models for image series (e.g., cardiac MRI). The manuscript, code, and data needed to replicate all experiments can be found at: http://www-stat.stanford.edu/~candes/SURE/index.html. If you think this work would be of interest to the Nuit Blanche community, we'd appreciate it being featured.  

All the best,
Josh Trzasko
Mayo Clinic

Thanks Josh for the heads-up:

Unbiased Risk Estimates for Singular Value Thresholding and Spectral Estimators by Emmanuel J. Candes, Carlos A. Sing-Long, Joshua D. Trzasko. The abstract reads:

In an increasing number of applications, it is of interest to recover an approximately low-rank data matrix from noisy observations. This paper develops an unbiased risk estimate—holding in a Gaussian model—for any spectral estimator obeying some mild regularity assumptions. In particular, we give an unbiased risk estimate formula for singular value thresholding (SVT), a popular estimation strategy which applies a soft-thresholding rule to the singular values of the noisy observations. Among other things, our formulas offer a principled and automated way of selecting regularization parameters in a variety of problems. In particular, we demonstrate the utility of the unbiased risk estimation for SVT-based denoising of real clinical cardiac MRI series data. We also give new results concerning the differentiability of certain matrix-valued functions.

Data and code are here.




Join our Reddit Experiment, Join the CompressiveSensing subreddit and post there !
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.

SCoBeP: Dense Image Registration using Sparse Coding and Belief Propagation - implementation -

Found through a discussion on the LinkedIn Compressive Sensing group

SCoBeP: Dense Image Registration using Sparse Coding and Belief Propagation by Nafise Barzigar, Amin mohammad Roozgard, Samuel Cheng, Pramode Verma. The abstract reads:

Image registration as a basic task in image processing has been studied widely in the literature. It is an important preprocessing step in various applications such as medical imaging, super resolution, and remote sensing. In this paper, we proposed a novel dense registration method based on sparse coding and belief propagation. We used image blocks as features, and then we employed sparse coding to find a set of candidate points. To select optimum matches, belief propagation was subsequently applied on these candidate points. Experimental results show that the proposed approach is able to robustly register scenes and is competitive as compared to high accuracy optical flow [1], and SIFT  flow [2].

The implementation of SCoBeP is here.





Join our Reddit Experiment, Join the CompressiveSensing subreddit and post there !
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.

Two jobs for students and two talks.

I just came across these two jobs for students and these two interesting talks.

• Internship (France) STAGE Compressive Sensing (H/F) (1 poste), Employeur : Thales Systèmes Aéroportés S.A. / Elancourt, Yvelines France | 22/10/2012 | Réf. 56204214
• PhD position in MR image reconstruction - compressed sensing, Image Sciences Institute, University Medical Center Utrecht

The talks:
@ Supelec (France) but also broadcast here. Structure Based Bayesian Sparse Reconstruction
Speakers: Tareq Y. Al-Naffouri, KAUST, Saudi Arabia
Date: Wednesday, October 31, 2012 - 14:00 to 15:00
Location: Council room of L2S (room B4.40), Supélec, campus of Gif-sur-Yvette, Supélec.

Abstract: There has been increased interest in sparse signal reconstruction algorithms (commonly known as compressed sensing) due to their wide applicability in various fields. In this talk, we present a novel low complexity Bayesian approach to the estimation of sparse signals. The approach jointly utilizes 1) the sparsity information of the desired signal 2) the a priori statistical information about the signal and noise and 3) the inherent structure in the sensing matrix to obtain near optimal Bayesian estimates. The proposed approach is able to deal with both Gaussian and non-Gaussian priors. The approach also exhibits relatively low complexity compared to the widely used convex relaxation methods as well as greedy matching pursuit techniques. The discussion will be illuminated with several signal processing applications including channel estimation in UWB, seismic deconvolution, and estimation and cancellation of noise/distortion.

@ UBC, Stephan Wenger Talk - Visualization Of Astronomical Nebulae Via Distributed Multi-GPU Compressed Sensing Tomography, DATE: TUESDAY, OCTOBER 23, 2012 | 1:00PM - 2:30PM

Title: Visualization of Astronomical Nebulae via Distributed Multi-GPU Compressed Sensing Tomography

Abstract:

"The 3D visualization of astronomical nebulae is a challenging problem since only a single 2D projection is observable from our fixed vantage point on Earth. We attempt to generate plausible and realistic looking volumetric visualizations via a tomographic approach that exploits the spherical or axial symmetry prevalent in some relevant types of nebulae.Different types of symmetry can be implemented by using different randomized distributions of virtual cameras. Our approach is based on an iterative compressed sensing reconstruction algorithm that we extend with support for position-dependent volumetric regularization and linear equality constraints. We present a distributed multi-GPU implementation that is capable of reconstructing high-resolution datasets from arbitrary projections. Its robustness and scalability are demonstrated for astronomical imagery from the Hubble Space Telescope. The resulting volumetric data is visualized using direct volume rendering. Compared to previous approaches, our method preserves a much higher amount of detail and visual variety in the 3D visualization, especially for objects with only approximate symmetry."

Join our Reddit Experiment, Join the CompressiveSensing subreddit and post there !
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.