Category Archives: epistemology

Most notable science and technology from the last 20 years, and predictions for the next 20

Here is a selection of the most notable advances in science and technology over the last twenty years.

I’ve collected these from people working around me (there may be a Berkeley Lab or Optics bias!) or by looking at what around me had made life different (a  an Academic life or California bias!) They are listed in no particular order, but the ordering tries to highlight some relationship between the topics.

Continue reading

Threads

I’ve been using Twitter (@awojdyla) more frequently over the last 3 years, finding a lot value in this tool which allows to address a worldwide audience and reach out to people in a very effective way.

Straight goals

Twitter is a very strange medium, in that it can be extremely helpful to reach out to people (the six degrees of separation collapse to one, basically), but whose rules and purpose are hard to understand.

Here’s a few remarks on my experience, and some resources if you’re interested in engaging the tweet game!

Continue reading

How to promote diversity among scientists

This is a compendium of the things I’ve learned discussing the issue with other trained scientists and running an association with many young researchers. This is not meant to be a comprehensive list, but should help with the discussion. This discussion is primarily based on gender imbalance, but intersectionality applies (sometimes in weird ways), though the case is not as thoroughly documented.

(I wrote this initially for colleagues in my organization, since I couldn’t find a good resource. Here’s a bunch of additional resources – Ideas In Action – that have sprung out since. I’d be very grateful if you could point me to other concise lists)

*   *
*

The lab touts about the number of Nobel Prizes in spurned out, but it is remarkable, if not surprising that none of them are women (Nobel Prize in physics awarded to women (3) are rarer than total solar eclipses in California.) While there are many great women scientists at the lab who will eventually receive the coveted distinction, we must, given the historical significance of the lab and its stature, move forward, lead the way set some guidelines.

 

Personal safety

The first and most important step to promote diversity among scientists is to promote and enforce personal safety, first in terms of fighting against harassment (sexual or other), not only in physical safety terms and how women are depicted (sexist and lewd jokes, prejudiced opinions; see Tim Hunt’s comment on how “Women in labs ‘fall in love with you … you criticize them, they cry’”), but also in terms of financial safety, especially when some populations can experience hardship due to delayed entry into the (paid) workforce without external support (see also “parental leave” below.) It is also generally a good idea not to eschew virtue signaling when it can make a difference (a rainbow flag on a door can go a long way.)

 

Acknowledge your bias

A common objection from researchers is that they are not biased because, you know, they only judge with the data in hands, and that they can tell a good scientist from another using objective information – a resume, a publication list, a curriculum. The trouble here is that biased are ingrained in us (see Daniel Kahneman), which are somehow necessary for us to navigate a complex world (see Gigerenzer), and that even women have biases against women (see Walton). It is therefore important to acknowledge our own biases in order to try to compensate when needed (e.g. in hiring committees.)

In my experience, the mere fact of pointing at an implicit bias does wonder. People are often willing to help, but they don’t even realized there is a problem. It’s often good to point out to organizers of conferences or panels that they are actually manels when you see one.

 

The luster of meritocracy

The most common bias in science is the idea that no matter who you are, your application should be based purely on your academic records, and not on other factors… such as gender. While it seems logical at first glance, we know better and should acknowledge that these seemingly objective metrics must be understood in a more general social context (for example have learned that even artificial intelligence, supposedly fair and objective, is far from immune to bias, see Katie O’Neil.)

 

The need to fight for diversity

Some scientist would come and say, hey, yes there’s an imbalance, but maybe we should let it be (don’t force women into physics if they don’t like it.) The trouble is that the end result is deeply shaped by the system (Fig. 1) and it is important to act early, ideally at the PhD level by making sure that the contingent is not too skewed towards one gender.)

 

Figure 1. The making of inequality (from Paul Walton)

 

While in a democracy it should be obvious that laws should be made by a legislative body with balanced gender, one might notice that the representative democracy in America is not very representative (only 20% of US representatives are women, while roughly half of US constituents are indeed women), while there is no reason other than history to explain this imbalance.

Though science is not a democratic process (there is no expressed need for equal representation), similar historical factors are at play, and diversity or lack thereof can cancel any competitive advantage in terms of science (see Marie Hicks) and inclusivity of technology (see Caroline Criado Perez.)

 

Alleged differences

Some people will go as far as to say that women are actually undesirable in science, based on alleged differences in mental capacity (see Saini), or more subtly in the “variance” of the population (men supposedly show more variance, therefore more chances of fringe cases; see Strumia, Fig. 2)

 

Figure 2. Excerpt of the infamous Strumia’s talk at the workshop on gender diversity at CERN

 

Similar arguments from the Charles Murray’s racist book Bell Curve are used to promote borderline anti-semitic ideas – see also Jordan Peterson.

Ahh! intersectionality….

 

Role models

It is important in order to bring more balance to have role models to whom young scientist can identify, ideally more recent (and more diverse) than Albert Einstein or Marie Curie. The role models should be invited to give talks on site, but ideally *not in the context diversity* — it is important not to fall in the Bechdel test trap, since (i) you will lose a lot of speaker who are tired of having to repeat again and again the many hurdles they faced (ii) these interventions are usually not very interesting outside the TED talk format (iii) the point of role models is to inspire to do science because of it, not in spite of it (that’s what we have experimented with Series X).

 

Mentors

Similar to role models but closer to the person, it is important to have mentors, that can come naturally or through some kind of pairing. These mentors can provide help and support, through sharing their personal experience, advices, and promoting their mentees through invitations to talks and workshops (that’s what we experimented with forum@MSD and  forum@ESA). Be careful that mentors should promote their mentee, not undermine them – the mentor should acknowledge the accomplishments of the mentee, not their own.

 

Provide a platform

Given the current imbalance in gender and diversity as whole, we must make sure that people from underrepresented groups get invited to the lab and are being able to leverage this position, through announcements and support. Success begets success, and the lab is a good reference when someone wants to get booked in other places. Given that diverse speaker are paradoxically more rare, a budget must be set aside to fly them here and/or for an honorarium (they should not work for free, especially when they are themselves not in a position of power.)

Scientists at the lab usually yield considerable power in their own field, and they should be encouraged to seek and promote diversity when they look for invited speakers (best practices and guidelines can be useful here, e.g. never have an all-male panel or seminar series.)

 

Promote scientists to leadership position

Academia is a very competitive environment, and any differentiating factor is useful when it comes to apply to position, especially when women face external factors that gradually push them outside academia (Fig. 1) Encouraging women to pursue ancillary activities (association, EAA or ERGs), where they can learn leadership skills and strengthen their network, is seen as important, and the lab should further its support to extra-curricular activities.

 

Work-life balance

Work-life balance is a vague concept that still has very real implications: while men do not face the discrimination related to their potential being pregnant, women do. The consequence of childbearing should be shared between the two parents, and initiatives such as (non-gendered) parental leaves are very useful to bridge the gap.

Some policies to make it more convenient to raise children can be implemented, such as policies against emails past 5pm during week-ends, or enforcing one day per week without meeting, to allow for telecommuting (parking at the lab is nearly impossible when an appointment to a doctor pushes your commute later in the day.)

 

Smash the patriarchy

Oftentimes I hear people (old white male) arguing about the current push for equality, dismissed as a PC coup and a threat to the freedom of expression, where they feel that *they* might become victims. This is baloney, and they should learn about the distinction between men (them) and patriarchy (the system), and not feel threatened – this is not about them, it is about all of us. They might have to change their habits coming from a position of power they rarely acknowledge, and learn to speak up when they see something wrong.

This applies to men… but also to women. I seen many times over

 

References

Paul Walton: Gender equal­ity in Aca­demia – what we have learnt

Angela Saini, Inferior: How Science Got Women Wrong

Daniel Kahneman, Thinking, fast and slow

Katie O’Neal, Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy

Marie Hicks, Programmed Inequality: How Britain Discarded Women Technologists and Lost Its Edge in Computing

Nearly half of US female scientists leave full-time science after first child, Nature, 19 February 2019

https://www.nature.com/articles/d41586-019-00611-1?fbclid=IwAR2unznTjBTUUTlkEQJTiFmV6ZOxQUxTmhFQw_5j48r6YFXnHjoDoUEH2wM

Caroline Criado Perez, Invisible Women: Exposing Data Bias in a World Designed for Men
https://www.theguardian.com/lifeandstyle/2019/feb/23/truth-world-built-for-men-car-crashes

https://www.insidehighered.com/news/2019/03/06/new-study-nih-funding-says-women-get-smaller-grants-men

http://berkeleysciencereview.com/inclusive-mcb

http://antoine.wojdyla.fr/blog/2017/10/20/sexism-in-academia/

OSA and SPIE Professional Conduct Research Assessing And Addressing The Level Of Harassment At Scientific Meetings

(edit March 9th, 2020)

The researcher journey through a gender lens – Elsevier (March 2020)

 

 

Wokipedia

Wikipedia is probably the best thing on Earth after sunsets, but it’s still far from perfect. Some articles are quite amazing, but oftentimes article about science topics or science personalities are nowhere near where they should be, and it seems that researchers should spend more time trying spread knowledge. Unfortunately, two things are in the way: the writers never get credit for it, and it’s bad optics in science to be the judge of notoriety for others.

It is amazing what you can accomplish if you do not care who gets the credit.
– Harry S Truman

Recently I became aware of an effort to improve the representation of scientists on Wikipedia, which is the go-to place to look up someone and evaluate their authority – in a world when men seems to preternaturally commend more than women. Let’s fix this!

Here’s a few people for who I have started a page (I’ll keep this list updated as I go – yes, I do take credit, on a page no one will ever read in hopes this may inspire some wandering soul.)

Scientific Topics:

People than need to be put on Wikipedia:
  • Daniele Ushizima – https://crd.lbl.gov/departments/data-science-and-technology/data-analytics-and-visualization/staff/daniela-ushizima/
  • Haimei Zheng – https://haimeizheng.lbl.gov/
  • Pascal Elleaume – synchroton radiation pioneer; https://www.esrf.eu/news/general/elleaume-obituary/index_html https://docplayer.fr/62415068-L-archicube-numero-special.html
  • Bianca Jackson https://orcid.org/0000-0002-1515-9650
  • Ashley White (AAAS Fellow, scientific communication)
  • Lady Idos (DEI Officer at Berkeley Lab)
  • Tara de Boer (CEO) –  BioAmp diagnostics
  • Chrysanthe Preza – Computational Imaging, University of Memphis https://umwa.memphis.edu/fcv/viewprofile.php?uuid=cpreza
  • Teresa Williams (TechWomen/AAAS fellow) – https://today.lbl.gov/teresa-williams-helps-to-inspire-a-culture-of-mentorship-and-networking-in-egypt/
  • Tokiwa Smith  – https://www.blackengineer.com/news/tokiwa-smith-changing-world/
  • Susan Celniker – https://sites.google.com/a/lbl.gov/women-at-the-lab/p/susan-celniker-ph-d
update August 2019
I went to a workshop organized by SPIE and led by the very Jess Wade; it was quite useful.
Here’s what I learned:
  • Do not paraphrase bios found on other website –– but you somehow can. Better than nothing!
  • You can use pictures from governmental sources for illustration, it’s always ok to use them (copyrights)
  • You can help with translating pages to other languages.

Kolmogorov Access

Back in undergrad, I remember being fascinated by the notion of Kolmogorov complexity in computer science.

Put simply, the Kolmogorov complexity is the minimal length (number of lines) of the code needed to generate a signal, would it be a mathematical sequence (such as one listed in the OEIS) or an image, irrespective to the size needed to store it. It bears deep relations with the notion of entropy (a great book on the topic is Information Theory, Inference, and Learning Algorithms by the late David MacKay.)

For example, a series of eight billion ones in a row would require 1GB of memory, but can be written in a few lines of code:

for i in 1:1e9; print 1; end

(To some extent, this is why computer science is often problematic, since one of the goal of a good code is sometimes to reduce its Kolmogorov complexity, but the final code does not show all the lines that have been erased to get there…)

In the field of arts, culture and science, this description seems naive: can you really generate a book based on a script, or has it infinite entropy?

Science is organized knowledge. Wisdom is organized life.
– Immanuel Kant

In the age of the Internet, can we do better?

update 6/10/2019: I’ve seen recently on Twitter the embodiment of these ideas, see Nicole R.‘s thread. Way to go!

Continue reading

On truth and its perils

What is true, what is false, what is wrong?

With the rise of large scale misinformation, this question has become more and more important, as it seems political parties around the world have reached the escape velocity of facts.

polarization in US politics really goes only one way…

Continue reading

Hyperloonies

Mr Musk is having a hard time, and even though I have great appreciation for his making engineering look cool again, I won’t relent  as I believe his efforts are misguided.

Nothing can be so amusingly arrogant as a young man who has just discovered an old idea and thinks it is his own. – Sidney J. Harris

Today, I’ll talk about Hyperloop.

edit 12/18/2018: Hyperloop startup Arrivo is shutting down -The Verge — lol

edit 10/10/2019: Was the Ocean Cleanup Just a Pipe Dream? – Outside online
Jenny Allen: “Male privilege in science is a 24 year old guy with no formal training being called a ‘boy genius’, receiving millions of dollars in funding, and referring to qualified female oceanographers as ‘Ms’ instead of ‘Dr.’ when they critique his project.”

Continue reading

Quick’n’dirty

Over the years I’ve collected quotes from people who are.

I always like quotes, because they are atoms of knowledge, quick and dirty ways to understand the world we only have one life to explore. To some extent, they axioms of life in that they are true and never require an explanation (otherwise they wouldn’t be quotations.)

Here’s a bunch of quotes that I found particularly interesting, starting with my absolute favorite quote comes from the great Paul Valery:

The folly of mistaking a paradox for a discovery, a metaphor for a proof, a torrent of verbiage for a spring of capital truths, and oneself for an oracle, is inborn in us. – Paul Valery

On research — trial and error

Basic research is like shooting an arrow into the air and, where it lands, painting a target.
-Homer Burton Adkins

A thinker sees his own actions as experiments and questions–as attempts to find out something. Success and failure are for him answers above all.
– Friedrich Nietzsche
Continue reading

Learning Deep

In the past four years, there’s been a lot of progress in the field of machine learning, and here’s a story seen from the outskirts.

Eight years ago, for a mock start-up project, we tried to do some basic headtracking. At that time, my professor Stéphane Mallat told us that the most efficient way to do this was the Viola-Jones algorithm, which was still based on hard-coded features (integral images and Haar features) and a hard classifier (adaboost.)

(I was thrilled when a few years later Amazon Firephone was embedding similar features; unfortunately, this was a complete bomb — better technologies now exist and will make a splash pretty soon.)

By then, the most advanced book on machine learning was “Information Theory, Inference, and Learning” by David McKay, a terrific book to read, and also “Pattern Recognition and Machine Learning” by Chris Bishop (which I never read past chapter 3, lack of time.)

Oh boy, how things have changed!

Continue reading

Julia Language

A few years ago, I got interested in the then-nascent Julia language (julialang.org), a new open source language based on Matlab syntax with C-like performances, thanks to its just-in-time compiler.

Large Synoptic Survey Telescope (LSST, Chile) data being processed with Julia on super computers with 225x speedup

Large Synoptic Survey Telescope (LSST, Chile) data being processed with Julia on super computers with 200x speedup (from https://arxiv.org/pdf/1611.03404.pdf)

It now seems that the language is gaining traction, with many available packages, lots of REPL integration (it works with Atom+Hydrogen, and I suspect Jupyter gets its first initial from Julia and Python) and delivering on performances.

Julia is now used on supercomputers, such as Berkeley Lab’s NERSC, taught at MIT (by no less than Steven G Johnson, the guy who brought us FFTW and MEEP!), and I’ve noticed that some of the researchers from Harvard’s RoLi Lab I’ve invited to SPIE DCS 2018 are sharing their Julia code from their paper “Pan-neuronal calcium imaging with cellular resolution in freely swimming zebrafish“. Pretty cool!

Julia used for code-sharing in a Nature publication. I wish I could see that every day!

Julia used for code-sharing in a Nature publication. I wish I could see that every day!

I got a chance to attend parts of Julia Con 2017 in Berkeley. I was amazed by how dynamic was the the community, in part supported by Moore’s foundation (Carly Strasser, now head of Coko Foundation), and happy to see Chris Holdgraf (my former editor at the Science Review) thriving at the Berkeley Institute for Data Science (BIDS).

Julia picking up speed at Intel (picture taken dusing JuliaCon 2017)

Julia picking up speed at Intel (picture taken dusing JuliaCon 2017)

I started sharing some code for basic image processing (JLo) on Github. Tell me what you think!

(by the way, I finally shared my meep scripts on github, and it’s here!)