Adventures in Signal Processing and Open Science

Magni: A Python Package for Compressive Sampling and Reconstruction of Atomic Force Microscopy Images

Our new software metapaper Magni: A Python Package for Compressive Sampling and Reconstruction of Atomic Force Microscopy Images has just been published in Journal of Open Research Software. The paper describes our new software package Magni:

Magni is an open source Python package that embraces compressed sensing and Atomic Force Microscopy (AFM) imaging techniques. It provides AFM-specific functionality for undersampling and reconstructing images from AFM equipment and thereby accelerating the acquisition of AFM images. Magni also provides researchers in compressed sensing with a selection of algorithms for reconstructing undersampled general images, and offers a consistent and rigorous way to efficiently evaluate the researchers own developed reconstruction algorithms in terms of phase transitions. The package also serves as a convenient platform for researchers in compressed sensing aiming at obtaining a high degree of reproducibility of their research.

The software itself is on GitHub as well as on Aalborg University’s repository: DOI 10.5278/VBN/MISC/Magni

Go ahead and check it out if you are into compressed sensing or atomic force microscopy. Pull requests welcome if you have ideas.

Live-tweeting iTWIST 2014 workshop

As an experiment I am live-tweeting the workshop iTWIST in Namur, Belgium. Look for the tag #itwist14.
See also http://www.ploscompbiol.org/article/info%3Adoi%2F10.1371%2Fjournal.pcbi.1003789 for inspiration (by @collabchem and @eperlste)

Episciences.org progress

Lately, I have been following the Episciences project as you may have noticed in my previous post. It seems there has been some more progress recently: I have just noticed that another “epi-committee” has been added to the site (I understand these epi-committees as a sort of editorial boards responsible for a given subject area). In addition to the existing math committee, the new committee is Episciences IAM (Informatics and Applied Mathematics). This sounds a bit closer to my area. I wonder if they consider signal processing to be in their area?
The page so far says that the committee is being formed and as such does not list any members yet. It will be interesting to see what this turns into.

Episciences.org update

I mentioned the Episciences project the other day in Scientific journals as an overlay. In the meantime I have tried to contact the people behind this project and The Open Journal, apparently without any luck.

I went and checked the Episciences website yesterday and it actually seems that they are moving forward. They changed the page design completely and there is now a button in the upper right corner to create an account and log in. I took the liberty of doing so to have a look around. I was able to create an account, but is just about it so far. The site still seems quite “beta” – I was not able to save changes to my profile and I cannot yet find anywhere to submit papers. It is nice to see some progress on the platform and I will be keeping an eager eye on it to find out when they will go operational.

Compressed Sensing – and more – in Python

Compressed Sensing – and more – in Python

The availability of compressed sensing reconstruction algorithms for Python has so far been quite scarce. A new software package improves on this situation. The package PyUnLocBox from the LTS2 lab at EPFL is a convex optimisation toolbox using proximal splitting methods. It can, among other things, be used to solve the regularised version of the LASSO/BPDN optimisation problem used for reconstruction in compressed sensing:

\underset{x}{\mathrm{argmin}} \| Ax - y \|_2 + \tau \| x \|_1

See http://pyunlocbox.readthedocs.org/en/latest/tutorials/compressed_sensing_1.html

Heard through Pierre Vandergheynst.

I have yet to find out if it also solves the constrained version. Update: Pierre Vandergheynst informed me that the package does not yet solve the constrained version of the above optimisation problem, but it is coming:

\underset{x}{\mathrm{argmin}} \quad \| x \|_1 \\ \text{s.t.} \quad \| Ax - y \|_2 < \epsilon

Scientific journals as an overlay

There is an update on this post in Episciences.org update

In many of my posts since I started this blog, I have been writing about open peer review. Another topic related to open science that interests me is open access (to scientific papers). Part of open access in practice is about authors posting their papers, perhaps submitted to traditional journals, to preprint servers such as arXiv. This is used a lot, particularly in physics and mathematics.
Read the rest of this entry »

The Winnower officially launches today

I have written about The Winnower here before. I have been involved in testing the platform during the past couple of months and must say that it looks very promising.
Today they officially launch! Now it is just up to us to participate and make a change towards transparent publishing with open review.

How to attract reviewers for open / post-publication review?

In traditional journals with closed pre-publication peer review, reviewers are typically invited by the editor. Editors can for example draw on previous authors from the journal or (I guess) their professional network in general. With the recent appearance of several open peer review platforms, for example PubPeer, Publons etc. – more here, there will be a need to attract reviewers to such platforms. Sufficiently flawed papers seem to attract enough attention to trigger reviews, but it is my impression that papers that are generally OK do not get a lot of post-publication review. This is perhaps not such a big deal for papers that have been already been published in some journal with closed pre-publication review – the major downside is that the rest of us do not get to see the review comments. But, if you are going to base an entire journal on post-publication peer review you will want to ensure at least a few reviews of each published paper as a sort of stamp of approval.

The Winnower – the new journal based entirely on open post-publication peer review that I have previously written about here – is about to launch. First and foremost they will of course need to attract some papers to publish. Their in my opinion very fair (when you compare to other open access journals) price of $100 should help and the fact that they span a very broad range of scientific disciplines should also give them a lot of potential authors. They also want to attract reviewers to their papers. The platform is open to review by anyone, but in order to ensure a minimum number of reviews of each paper with at least some experts on the topic among them, this seems like a good idea. But how do you do this? As a new journal there are no previous authors to draw on. It can probably get difficult to get in touch with sufficiently many qualified reviewers across all of the journal’s disciplines and on top of that, reviewers may be more reluctant to accept since the reviews will be open with reviewers’ identities disclosed.

What can be done to attract sufficiently many reviewers? Should the journal gamble on being the cool new kid in class that everyone wants to be friends with or simply try to buy friends? I have been discussing this with their founder Josh Nicholson. One possibility is to pay reviewers a small amount for each review they complete. If we were talking one of the traditional publishers, whom I think in many cases are exploiting authors and reviewers shamelessly to stuff their own pockets, I think it would only be reasonable to actually start paying the reviewers. In the case of The Winnower, it may be different. The Winnower is a new journal trying to get authors and reviewers on board. With a very idealistic approach and pricing, I do not think people are likely to think that they are just trying to make money – being a “predatory publisher”. But on the other hand, paying reviewers might somehow make it look like they are trying to “buy friends”. With the journal’s profile, potential reviewers might mainly be ones that like to think of themselves as a bit idealistic and revolutionary too and that might just not go well with being paid for reviews? Josh told me an anecdote he had been told recently:

To illustrate this point, imagine walking down the street and an able bodied young man asks for your help loading a large box into a truck.  If he were to politely ask for help, most people would be highly likely to assist him in his request.  However, if he were to politely ask and also mention that for your time, he will pay you $0.25 most people will actually turn him down.  Despite being totally irrational, given that under the same circumstance they would do it without the promise of a quarter, the mention of money evokes the passerby to calculate what their time is worth to them.  To many, the $0.25 isn’t going to be worth the effort.  I think this may pose a similar issue.

Then again, paying reviewers could also send the message that they are taking their reviewers and the work they do very seriously? Ideally, I think it might work better with an incentive structure and “review of reviews” / scoring of reviews like the Stackexchange network for example. I am just afraid that something like that will take considerable “critical mass” to be effective. Another option Josh mentioned could be to let reviewers earn free publications with the journal by completing a number of reviews. This sounds better to me: you do not risk offending potential reviewers with a “price on their head”, but there is still something to gain for reviewers.

I think this is a very interesting question and probably one that a lot of people have much more qualified answers for than I. Let me know what you think?!

Standalone peer review platforms

Standalone peer review platforms

I have previously mentioned some platforms for open / post-publication peer review in Open Review of Scientific Literature and discussed the roles of such platforms in Third-party review platforms. I just wanted to mention the above document in Google Docs which seems to have been started by Jason Priem(?). The document contains a list of peer review platforms; both standalone and including manuscript publishing as well. Go have a look – there are probably some that you don’t know yet. Anyone can edit the document, so please add platforms if you now any additional ones.

Compressed sensing with linear correlation between signal and measurement noise

Torben Larsen and I have recently published a paper, “Compressed sensing with linear correlation between signal and measurement noise” in EURASIP Signal Processing. This post is an attempt and a sort of experiment to provide a front page summarizing the paper’s contributions and providing an overview of available versions of the paper and its accompanying code.

We considered compressed sensing with measurement noise in the case where the measurement noise is linearly correlated with the signal of interest. So we have the typical compressed sensing model with measurement noise:

\mathbf y = \mathbf{Ax} + \mathbf n

where the noise \mathbf n is now correlated with \mathbf x. This can be modelled as a scaling by some factor \alpha of the measured signal in addition to additive random noise:

\mathbf y = \alpha \mathbf{Ax} + \mathbf w

The difference in the measurement between the original and scaled signals constitutes the part of the resulting measurement noise that is correlated with the input signal:

\mathbf n = \alpha \mathbf{Ax} + \mathbf w - \mathbf{Ax} = (\alpha - 1) \mathbf{Ax} + \mathbf w

We show that in the case of reconstruction of the measured signal by basis pursuit de-noising (BPDN), the correlation between the measurement noise and the measured signal can be compensated simply by scaling the BPDN solution by 1/\alpha.

It turns out that this simple correlated noise model models the error introduced by low-resolution quantisation quite well. We have tested the proposed reconstruction approach on compressed measurements quantised to 1, 3, and 5 bits, respectively. Especially in the extreme case of 1 bit quantisation we see substantial improvements in reconstruction error, reducing the error by up to around 7dB. This simple modification of BPDN performs better than BIHT (which is specifically designed for 1 bit quantisation) in a large portion of the undersampling/sparsity phase space.

Reconstruction MSE

Relative reconstruction MSE of the proposed approach. The fat contour line marks the region (above and left of it) where the error is below that of BIHT reconstruction.

Below, you can find links to both the official published version of the paper, all versions from the review process on arXiv, and the code for running the numerical simulations.

Paper versions and simulation code

 

Pandelis Perakakis, PhD

Academic Website

chorasimilarity

computing with space | open notebook

PEER REVIEW WATCH

Peer-review is the gold standard of science. But an increasing number of retractions has made academics and journalists alike start questioning the peer-review process. This blog gets underneath the skin of peer-review and takes a look at the issues the process is facing today.

Short, Fat Matrices

a research blog by Dustin G. Mixon

www.rockyourpaper.org

Discover and manage research articles...

Science Publishing Laboratory

Experiments in scientific publishing

Open Access Button

Push Button. Get Research. Make Progress.

Le Petit Chercheur Illustré

Yet Another Signal Processing (and Applied Math) blog

Follow

Get every new post delivered to your Inbox.

Join 158 other followers