First, It looks like I am getting the hang of writing these Sunday Morning Insights, so much so that now they have a tag. It's a little bit like in Aggieland, every time you do something twice it becomes a tradition. I don't know how much time I will be able to keep up on those or if they are useful but this is an interesting area to explore. There were three entries for this month. It will not have escaped the ever sharp and attentive reader that there were more than three Sundays this month. I agree but the these entries are called Sunday Morning Insight, the last word being as important as the first two:
At some point in time, people will have to notice that our community is at the forefront of open scientific inquiry not because of the substantial contribution it brings to the world but rather by the natural way of doing things , i.e. share their codes with the rest of the world. I have also noted a burgeoning trend in setting up some of these efforts on GitHub starting with SPGL1. When you talk to your colleagues let them know you are part of that trend, you are sending a very strong signal. Here are the implementation that were listed this month:
- From Bits to Images: Inversion of Local Binary Descriptors [8 G+]
- Fast Marginalized Block SBL Algorithm
- COSMOS: sparse modeling software
- Belief Propagation Reconstruction for Discrete Tomography [4 G+]
- FLAGLET: Exact wavelets on the ball
- S2LET: A code to perform fast wavelet analysis on the sphere
- Bayesian Blind Deconvolution With General Sparse Image Priors
- ReProCS: Recursive Robust PCA or Recursive Sparse Recovery in Large but Structured Noise
- Low Rank Approximation and Regression in Input Sparsity Time
- Fast Earth Mover's Distance (EMD)
- Compressive Phase Retrieval via Generalized Approximate Message Passing
- CPRL – An Extension of Compressive Sensing to the Phase Retrieval Problem
I, for one, am stunned by the quality of the diversity of the algorithms that are proposed. This month we had a few entries related to hardware hacking. This is a trend that will not die. When you go to some of these meetings, you are generally asked to provide three tags that somehow define your interests. The latest tags I used were: Dumb Sensors, Really Dumb Sensors, Really Really Dumb Sensors. It definitely is an attention getter but I suppose it can turn people off too. In light of these trends I wonder if I should not add "the Internet of Things" to the Wondering Star blog theme. The reason I started that blog was to talk about sensors that were the size of a planet. Most were government owned projects but I see how the arrival of Arduino, OpenPicus could change this altogether, stay tuned: Anyway, here are the entries on this theme this month:
Other entries that are difficult to sum in one or two themes are listed below. Several received some unusual large traffic:
- Accelerating Particle Filter using Randomized Multiscale and Fast Multipole Type Methods
- Compressive Sensing Solvers in Python and the ones that ought be in Python, R,.... [5 G+]
- Approximate Message Passing and Compressive Sensing
- Small-sample brain mapping: sparse recovery on spatially correlated designs with randomization and clustering
- Variational Properties of Value Functions
- Around the blogs in 80 summer hours
- Astounding Compression and Beyond Belief Propagation [ I particularly liked the comments of this one]
- Mind Maps of Compressive Sensing / A compressive sensing framework for seismic source parameter estimation
- Compressed Sensing with Electron Microscopy and Optical-resolution photoacoustic computed tomography
- Learning Manifolds in the Wild [4 G+]
- Fast Functions via Randomized Algorithms: Fastfood versus Random Kitchen Sinks
- OSNAP: Faster numerical linear algebra algorithms via sparser subspace embeddings
- A note from Jelani Nelson on OSNAP
- Eigenvalues of a matrix in the streaming model
- Sketched SVD: Recovering Spectral Features from Compressive Measurements - implementation - and Randomized Matrix Computations
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.
No comments:
Post a Comment