Building Up A Bridge

R&D funding institutions sometimes insist on the importance of multi-disciplinary work in Science. Other times, they just simply seem to forget about that.

I think that a part of the recent work made by Prof. Jennifer Listgarten on using machine learning for CRISPR/Cas9 is an evident example of sum of forces.

You might want to have a look at the interview she gave HERE and in particular to her answer to the question:

"How did you get interested in applying machine learning to CRISPR technology?"

I Hadn't Realized Machine Learning Reached Maturity Already

My (BSc + MSc) background is Physics and I have always found Philosophy of Science very interesting to read and reflect on. However, I was not aware of the following until a few days ago: a knowledge corpus is being created that is called  Philosophy of Machine Learning.

Two examples:

1) Philosophy of machine learning: knowledge and causality

2) Machine Learning Meets Philosophy: From Epistemology to Ethics

A very good symptom in many senses...

Dark Matter: Machine Learning Will Shed Light On You

Very intriguing, nice and estimulating new workshops, conferences and meetings are currently appearing on applications of Machine Learning.

Have a look, for instance, to:

"Accelerating the search for dark matter with machine learning"

Description and aims: HERE

I can only partially feel you: Turning Corners Into Cameras

Katherine Bouman and co-workers just developed a clever way to infer movement of objects in a scene using subtle changes in radiance arriving at a camera, with the peculiarity that the camera does NOT see anything of the scene at all!

Good piece of work!

Paper, code and datasets can be found HERE

Youtube video below:

I probably did not say that...

For my opinion, see my previous post on "and then one president became another president" :-(

Food System Panorama

The following paper, co-authored by Prof. Molly Brown recently called my attention:

Do markets and trade help or hurt the global food system adapt to climate change?

ME Brown, et al

Food Policy 68, 154-159

Link to paper: HERE (ScienceDirect)

I really feel (I do not know her opinion. It is mine) we are assuming climate change as something to be accepted (without an alternative) and feel like saying Quo Vadis? 

On the other hand, I would have wanted to see an empirical/numerical analysis of the topic in the paper. I guess it has to be very hard to do, but worth trying it (?).

Heterogeneous dialogue

Facebook research seems to be taking research into dialogue understanding seriously.

You can read an interesting post HERE that I found looking for research papers authored by Dr. Antoine Bordes.

I see their point in trying to optimize a conversation with a machine and also trying that it might be more friendly, or more human (so to speak) but I don't feel comfortable when technology might be able to understand, some day, in an automated way, a written conversation that I may have with somebody.

May be I see ghosts everywhere (?!)

Light Fields and Mixed Reality

A world full of possibilities just in front of and close to us.

Link to Vimeo video HERE

All these ideas come in part from Prof. Gordon Wetzstein (Stanford university) and collaborators' minds.

I thik it is worth having a look at his group's web page.

Shannon will take care of me

Impact of Prof. Claude Shannon (link to Wikipedia HERE) work is so deep and diversified... It reaches almost any knowledge branch where "information" can be somehow measured.

I recently found a particularly useful and interesting application in the following paper:

R. Quax; B.D. Kandhai and P.M.A. Sloot: Information dissipation as an early-warning signal for the Lehman Brothers collapse in financial time series, Scientific Reports, vol. 3, 2013.

We can measure these (complex) dynamics and even predict their impact. The key issue for me is also a measure of immediacy.

Am I your cause or your effect?

Prof. Bernhard Schölkopf's work never disappoints. Herculean work but worth the effort.

Mooij, J. M.; Peters, J.; Janzing, D.; Zscheischler, J. & Schölkopf, B.,
"Distinguishing Cause from Effect Using Observational Data: Methods and Benchmarks",
Journal of Machine Learning Research, 2016, 17, 1-102

Link to the paper details HERE

Hammers and nails

Simple and very useful:

R. E. Kass, B. S. Caffo, M. Davidian, X. Meng, B. Yu, Nancy Reid* (2016). Ten simple rules for effective statistical practice. PLoS Comput. Biol., 12(6): e1004961. doi:10.1371/journal.pcbi.1004961

Old times' sake: "Accuracy, error, precision and uncertainty"

First principles...HERE