Ian @yumsoonhan - Tumblr Blog

Generating test data from ReMarkable

For a while I’ve wanted to train a neural net on my handwriting, as part of a larger test project that involves a drawing interface. To minimize error effectively I need the highest recognition rate possible, so using a generic dataset like IAM for testing is not necessarily the best. Plus, I don’t know enough about machine learning currently to entertain adaptive networks, that start from some generic test set and gradually adapt to the writer’s writing style.

I used to contemplate developing an HTML5 app connecting to a backend that records writing on my iPhone, but although that allowed for on-line input, it was less straightforward for a hobby project. An alternative was to develop a template paper with a grid and scan them, but this was also time consuming. Luckily, I have a ReMarkable tablet to speed up the process!

The ReMarkable has a grid template that makes it easy to enter characters en masse:

The ReMarkable desktop app lets you easily export this writing to a folder of PNGs. Within two days I already have, for instance, 329 images of the letter “x” in my handwriting (extracted as separate images of characters using a Python script). My task does not necessarily require word input (yet), so I’m going to make do with what I have and build the model from here.

One of the things I realized as I was drawing was how similar certain letters are to others (and how different). “a” and “d”, for instance, look similar in my handwriting, the only different being the stem on the “d”. “D” and “O” also could be confused, or “U” and “V.” Using this knowledge we can restrict ourselves to a subset of the Latin alphabet which are all pretty mutually distinct, like X Y Z.

#remarkable #training data #neural networks #machine learning

...what is given to us is not just there for the taking as data for collection, but is an offering, the acceptance of which carries a responsibility of care. Anthropology shows that curiosity and care, pried apart in mainstream science policy by a spurious and ethically indefensible division between research and impact, are inseparable aspects of our relations with those to whom we owe our education in the ways of the world.

https://culanth.org/fieldsights/841-enough-about-ethnography-an-interview-with-tim-ingold / https://www.journals.uchicago.edu/doi/10.14318/hau4.1.021

#tim ingold #anthropology #ethnography #quotes

I don't believe in PowerPoint because when we use PP we project images on the screen. PP is the epitome of the logic of projection which I'm arguing against. And the reason why I like blackboards is the blackboard is the epitome of the process of creativity that I'm arguing for. When you stand at the blackboard and scrape a line, then your movement, your awareness, the trace of the materials, all are bound up in that one performance and what you see is the outcome of that performance.

Tim Ingold, https://www.youtube.com/watch?v=Ygne72-4zyo

Men who may have seeds of negativity and domination within them along with positive traits may find the negative burgeoning at times of crisis in their lives.

from bell hooks’ “The Will to Change”

#bell hooks #toxic masculinity

El sueño de la razón produce monstruos

can i put emoticons in paragraphs of my novel? :/

Life tip: When someone sends you an aggressive email, write your response twice, and pick the second.

A long-time adherent to Seymour Papert's constructionist model of learning, *which viewed children as innately able to teach themselves*, Negroponte insisted that computers could allow all children to self-educate...

From Networking Peripheries by Anita Chan. The emphasized statement is actually incorrect and puts forward a common misconception perpetuated by “technorealist” circles: in fact, constructionism is not the kind of constructivism Papert rebelled against, where students should self-educate. Here, in writing Chan betrays a technocentric assumption that equates Negroponte’s ideas to those who work(ed) in the lab he founded, flattening the “rural” as a privileged untapped source of disruptive or inconvenient knowledge and the “urban” as a spigot from which techno-solutionism flows, forgetting that disruption is not just (always) on the periphery: it can be front-and-center, operating not merely through what is spoken but what is *heard*, how things are taken, translated, and misconstrued across conglomerations of multi-national actors with varying aspirations and political motivations. The silence (or un-examination) of those who may contend with OLPC within its very womb is not actually proof of techno-solutionist homogenization.

Put another way, Chan has learned what constructionism is from a translation -- Negroponte’s -- while leaving unexamined the possibility that his translation may itself be a misrepresentation. The mistranslation, I think, may be the actual danger, whereby actors equate the success of a few educational projects to technology and not social forces at work, in the process losing the (non-technological) parts of what actually made them successful -- even when some actors, in the heart and at the edges, believe and try to argue otherwise.

The Unexpected Union between NLP and Ethnography

Lately I have been learning the (extremely lengthy and disciplining) process of turning “jottings” into full field-notes. After spending over a day converting only one four-hour session, I’ve realized that much of the process’ tedium is marred by manual expansion of shorthand.

Shorthand has been around for centuries, consisting of (sometimes incomprehensible) ticks and slashes. Typing on computer makes it slightly different: I’ve adopted my own shorthand based on disemvoweling, IM abbreviations, and making abbreviations for local common words like “Swahili” (it’s Sw). I’ve also adopted a convention whereby gestures are placed in-line with words using (when I’m good) double-brackets e.g. [[points to board]]; and a naming convention where commonly-used names are abbreviated as capital first letters.

Facing the daunting task of converting more jottings, I wrote a simple Python program that reads a text file, finds the words and then converts any shorthand it finds into longhand with a dictionary lookup. This alone has already saved me tons of time, for those shorthand abbreviations I can be sure won’t be confused with other things.

However, one can go further... For instance, the abbrev. “ur” for “you’re” and “your” makes it impossible to use simple conversion; a full NLP parser would infer which is most accurate. (This is probably an easy task for any expert NLP practitioner). In another instance, I spend time isolating the gestures and names in dialogue to full, properly-written lines, like ‘“Hello X,” said M, pointing to X.’ The ethnographer can then work with these expanded jottings when making field-notes, reducing a lot of tedious effort in the process.

I hope to work on this more and release it when I’m back, if other field researchers find it useful.

#fieldnotes #ethnography #shorthand #jotting #notes #taking notes

In an age of mendacity and criminality which is our own, just telling the truth is revolutionary.

Cornel West (https://www.youtube.com/watch?v=X2kH6kSY6ps)

In times of extremes, extremists win. Their ideology becomes a religion, anyone who doesn't puppet their views is seen as an apostate, a heretic or a traitor, and moderates in the middle are annihilated. Fiction writers are particularly suspect because they write about human beings, and people are morally ambiguous. The aim of ideology is to eliminate ambiguity.

Margaret Atwood (https://www.theglobeandmail.com/opinion/am-i-a-bad-feminist/article37591823/)

The first things we did were change, giving a name to the change came after. We began to 'know' a long time before saying that we know. We learn before teaching.

Paulo Freire (https://www.youtube.com/watch?v=FnVCyL9BwS8 21:26) (http://www.papert.org/articles/freire/freirePart2.html)

#paulo freire #quotes

When you start to humanize your enemy, you, in turn, may be dehumanized by your community.

Cassie Jaye, TEDxMarin https://www.youtube.com/watch?v=3WMuzhQXJoY

Chimamanda Ngozi Adichie speaks at the NYTimes. Normally I don’t especially enjoy author-talks (can be too light on substance), but this talk is an exception. Adichie has very enlightening and atypical responses.

According to Greek tradition, Archimedes could have moved the world had he had a firm ground on which to stand. In twentieth-century Africa, he would have been advised to stand on the firm ground of the English language in order to move the world.

Ngugi wa Thiong’o, Moving the Centre: The Struggle for Cultural Freedoms

And then, there is something else which is important: the whole illusion that, if you have a prosperous middle class, then there's development. But to me, development should be measured not by the perspectives of those who are at the top of the mountain, but be measured from the needs and perspective of those at the bottom of the mountain.

Ngugi wa Thiong’o https://www.youtube.com/watch?v=TUxkPINiimg

“[Observability] doesn’t necessarily make people better off. It does potentially increase contributions but [sometimes] people don’t like it, they feel ashamed... [...] The person may be worse off not just by the lost [cost] but they may actually enjoy the experience less. If your organization’s entire goal is to increase contributions, [making things observable] is fine, but you might also want to think about your relationship with your customers more broadly.”

Trending Blogs

Recently Viewed Blogs

Ian