By using Excel, which was never designed for scientific research, they institutionalized mouse clicks and other untraceable actions into a scientific workflow, which must be avoided since it makes explaining to others (and to oneself) how to replicate the findings next to impossible and too easily introduces inadvertent mistakes.
— Victoria Stodden, on Reinhart and Rogoff (via tofias)
Wikipedia defines Machine Learning as “a branch of artificial intelligence that deals with the construction and study of systems that can learn from data.”
Below is a compilation of APIs that have benefited from Machine Learning in one way or another, we truly are living in the future so strap into your rocketship and prepare for blastoff.
You’ve probably heard by now that Chelyabinsk, Russia was hit by a meteor strike on Friday.
On Friday? (so neutral, I feel nothing, next…)
You mean TODAY. (OMG, TODAY?)
Next week I would like that to read “on Friday,” on next Friday read that as “last Friday,” then you can just use a date, or X days ago if X < 31.
It is known how to do this in javascript since 2008. Just have it in the article as a datetime microformat and use some magical javascript to display something relative to the visitor’s local date.
Relate things to the visitor’s frame of mind, space and time are the easiest to deal with automatically.
And for the semantic problem, [systems theorist Howard Pattee] adds, “[T]he concepts of causation have completely different meanings in statistical or deterministic models,” and gives the following example: If you were to ask “What is the cause of temperature?” a determinist will assume that cause refers to a microscopic event and say it is caused by the molecules exchanging their kinetic energy by collisions. But the skeptical observer, scratching his head, will note that the measuring device averages this exchange, and does not measure the initial conditions of all the molecules and that averaging, my dear sir (or madam), is a statistical process. An average cannot be observable in a microscopic, determinist model. We have a case of apples and oranges. Pattee wags his finger at those who champion one model over the other and instead champions the idea that they are both needed and are complementary to each other. “I am using complementary here in Boltzmann’s and Bohr’s sense of logical irreducibility. That is, complementary models are formally incompatible but both necessary. One model cannot be derived from, or reduced to, the other. Chance cannot be derived from necessity, nor necessity from chance, but both concepts are necessary… . It is for this reason that our concept of a deterministic cause is different from our concept of a statistical cause. Determinism and chance arise from two formally complementary models of the world. We should also not waste time arguing whether the world itself is deterministic or stochastic since this is a metaphysical question that is not empirically decidable.
— Gazzaniga, Michael S. Who’s in Charge?: Free Will and the Science of the Brain. New York: Ecco, 2011. (via carvalhais)
I am over laptops and the posture that comes with them. I am coding a lot less, so I use my computers a lot less. I still want to simplify even further and carry just one device. So, I want to try the iPad Mini with cellular antenna as my only device and as a phone replacement, and use Skype and/or Google Voice instead.
What we believe in.
(via Shawn Blanc)
About a week ago I flew in from London to Bucharest for the Open Media Challenge (OMC), a Knight-Mozilla OpenNews funded event initiated by the Sponge Media Innovation Lab for Eastern Europe.
I landed at about 2PM in a sunny hot weather, quite unusual for this period of the year and in a stark contrast with the (lovely) cold London.
The taxi trip to the university—where the hackathon was already in full swing—was quite unusual, at couple of minutes distance from the airport there we were stopped by the police that checked the papers of the driver then surrounded by journalists asking the nigerian (I believe) driver on his right to work in Romania.
Also I was asked how I chose this taxi… “I just took the next in line” I answered. (I was unaware of the recent murder of an Japanese girl after airport pickup.)
I arrived in the end at the university, the teams were spread amongst several rooms and everyone was in that deep hack mode. Later in the afternoon I discussed with each team their project idea and where they thought they would be versus where they are now, and I reassured that all that matters is the potential of the project.
The second day was in the functional room of an absinthe bar (we had only coffee, I swear) set in the beautiful historical centre of Bucharest.
After presenting them the work the OpenNews fellows did with popcorn.js at BBC, Al Jazeera, etc., we proceeded with final projects presentation which took roughly two hours.
At lunch Stephen King partner at Omidyar Network talked on how transparency and participation can be supported by technology and how it can drive more accountability.
The winners were announced shortly after the lunch:
It was a hard choice for the jury, all the projects were very good and I’m sure they will evolve into crazy things as one of the most amazing thing I witness there was the participants discovering how their projects and datasets can interact further.
One of the projects was “Mapping the stray dogs in Bucharest”, and a member of the jury got bitten by a (non-stray but free roaming) dog the day before… I’m sure the incident was carefully planned just to stress the importance of a such project.
In two months, my Knight-Mozilla OpenNews fellowship at BBC News Specials will end and everybody wonders what I’ll do next.
In a previous century, while studying medicine (long story) I got a part-time webmaster job at one of the most important regional newspapers in Romania (Monitorul de Iași). After working there on the newspaper site and other related online publications for over a year, I left with two colleagues — a designer and a strategist (both also med students and working part-time at the same newspaper) and founded what became one of the leading branding and interactive agency in Romania: Grapefruit*
That unlikely path led me to this fellowship, which has been its own adventure. Let me summarise the past seven months:
January–February: Settling in at the BBC (fighting their proxies), doing choropleth maps experiments with D3.js and Raphaël, also various painful Internet Explorer 6 (don’t ask) tests.
March: South by Southwest (SXSW) Interactive in Austin (Texas, US), then Malofiej20 (Infographic World Summit) in Pamplona (Spain).
April–May: Getting embedded in the BBC UK Election team and moved to another building for that project. Played all their agile games while working on a flash-free election map that had custom zoom behaviour, had to work from iPad to Internet Explorer 6, had to integrate with various disparate data sources, and had to be bilingual: English and Klingon (actually Welsh). Post elections, I spent some time prototyping a flash-free alternative for BBC’s audio slides while fantasising about a complete replacement of the SoundSlides.com tool.
June: Eyeo Festival in Minneapolis (Minnesota, US), MIT-Knight Civic Media conference at MIT Media Lab (Heaven) where at the pre-conference hack day we teamed up with a Dalek and prototyped newsquest.me.
July: Tor Project hackfest in Florence (Italy), Data Live (Knight Mozilla OpenNews Fellowship Hack Jam) in Dundee (Scotland), Guardian Discovery week in London while eavesdropping on BBC News Labs hackathon that was happening at the same time with the Guardian’s.
In the meantime I also worked on “Tomlinson’s last movements” map/slideshow with video transitions (popcorn.js magic).
Working with the other fellows was the most rewarding experience of this fellowship. While day-to-day work at BBC was challenging, it was limited in the aspect that it was serious work and not experimental bat-shit crazy stuff that may work only on one browser. With my fellow fellows, we played with arduino, scrapers, speech recognition and video transcripts, natural language processing, seriously.js, processing.js and all the other *.js cool toys at the hack days we attended together.
BTW, if you think you’ll survive that much fun you can apply too, the deadline is August 11th.
Before this fellowship the options were quite simple: go back to academia or be a front-end developer in a cool place. Now, all the conferences I attended, all the hack days and experiments, all conversation with all the crazy people I met, have made me even more of a generalist than I was before. Not so long ago being one made you unemployable; fortunately this is not longer the case. At least not in the cool places.
There are still two months left and a lot of experimental code to be polished, documented, published, and blogged about. I believe that what’s next for me will be shaped by those, by how I package and communicate them: the range is from RTFS to extensive tutorials as I can play anything from the over-caffeinated developer to the overly pedantic teacher.
Actually I will tell you how I’ll determine what’s next: I will use the atemporal Feynman method, and yes, this was STEP 2.
