Bulletin #7, 2016-07-05: Montréal, Innocité, quotable links/liens citables!

Montréal gagne/wins! | Innocité | New: quotable links! Nouveau: liens citables!
No. 7, 2016-07-05                                      wrangler/responsable: @prooffreader

Welcome to the seventh edition (July 5, 2016) of the MTLDATA Bulletin. Don’t forget, you can also find us on the Web, Facebook, Twitter, LinkedIn and Meetup. This issue: ..., links, calendar, jobs and more! English content is in blue, French in green: a design choice suggested by one of our readers and implemented last issue. Keep the great ideas coming! Even the not-so-great ones, we’d love to hear any feedback you have! The eighth issue should be released the last week of July.

Bienvenue à la septième édition (5 juillet 2016) du Bulletin MTLDATA; n’oubliez pas, vous pouvez nous trouver aussi sur le Web,  Facebook, Twitter, LinkedIn et Meetup. Cette édition: ..., des liens, un calendrier, des jobs et plus! Le contenu en anglais est en bleu, le français en vert -- c’est un suggestion de design suggéré par un de nos lecteurs, et nous l’avons implémenté l’édition dernier. Merci, et SVP continuez de nous envoyer vos bonnes idées! Même si elles ne sont pas si bonnes, nous vous encourageons de nous les faire parvenir, tout feedback est super apprécié! La huitème édition devra être diffusée durant la dernière semaine de juillet.

Bulletin also published online -- > http://mtldata.blogspot.ca ← site Web du Bulletin


Image 20160702_151135.png

Tonight, July 5, at La Gare, the Innocité MTL smart city entrepreneurship accelerator is holding a 5 à 7 to explain how how your startup can be chosen in the next cohort! Ce soir, le 5 juillet à La Gare, l’accélérateur des entrepreneurs dans la domaine de la ville intelligente, Innocité MTL, présente un 5 à 7 afin d’expliquer comment votre startup peut se faire choisir dans la prochaine cohorte!


  • “A Google engineer in the Android division, Holgate is one of 18 programmers in this year’s Machine Learning Ninja Program, which pulls talented coders from their teams to participate, Ender’s Game-style, in a regimen that teaches them the artificial intelligence techniques that will make their products smarter. Even if it makes the software they create harder to understand.” How Google is remaking itself as a machine learning-first company
  • Capital Intelligent Mtl, c’est un nouveau fonds de $100M qui donnera “accès à du financement aux meilleurs entrepreneurs proposant des solutions innovantes pour la ville intelligente, que leur entreprise soit au stade de démarrage ou en croissance, d’ici ou d’ailleurs.”  •  Capital Intelligent Mtl is a new $100-million fund providing “dynamic entrepreneurs with the funding necessary to accelerate the development of innovative Smart City solutions, be they start-ups or established local and international businesses.”
  • “A central and common task for us as research investigators is to decipher what our data are able to say about the problems we are trying to solve. Statistics is a language constructed to assist this process, with probability as its grammar.” Ten Simple Rules for Effective Statistical Practice (PLOS)
  • “The bot was created by the self-taught coder after receiving 30 parking tickets at the age of 18 in and around London.” Chatbot lawyer overturns 160,000 parking tickets in London and New York (The Guardian)
  • “We argue that while this law will pose large challenges for industry, it highlights opportunities for machine learning researchers to take the lead in designing algorithms and evaluation frameworks which avoid discrimination.” EU regulations on algorithmic decision-making and a “right to explanation” (ArXiv)
  • “For almost 50 years, CSV has been the format of choice for tabular data. Given the ubiquity of CSV and the pervasive need to deal with CSV in real workflows — where speed, accuracy, and fault tolerance is a must — we decided to build a CSV reader that runs in parallel.” ParaText (wise.io)
  • “Algorithms learn by being fed certain images, often chosen by engineers, and the system builds a model of the world based on those images. If a system is trained on photos of people who are overwhelmingly white, it will have a harder time recognizing nonwhite faces.” Artificial intelligence’s white guy problem (New York Times)
  • “Convolutional neural networks require large datasets and a lot of computional time to train. Some networks could take up to 2-3 weeks across multiple GPUs to train. Transfer learning is a very useful technique that tries to address both problems. Instead of training the network from scratch, transfer learning utilizes a trained model on a different dataset, and adapts it to the problem that we're trying to solve.” A Practical Introduction to Deep Learning with Caffe and Python
  • “Depuis que j’ai créé la société ForCity à Lyon avec mon associé Thomas Lagier, j’ai su éluder, pirouetter, slalomer entre les gouttes, bref éviter la bouteille à l’encre du numérique dans les territoires.” Point de vue: C’est quoi une ville intelligente?
  • “Recommender systems are interesting because they're bold. They're bold because they attempt to model a human trait - taste - that we don't fully understand ourselves. Taste is humbling as an area of research because, like so many complex phenomena, we have to shift from asking why to describing what and how.” What's New in Recommender Systems (Fast Forward Labs)
  • “Scott Klein, who directs ProPublica’s news apps team, is unequivocal about this: it’s far better to teach a journalist to program than instil a programmer with editorial judgment and storytelling chops. He would know: most of the news apps mavens in his team, including himself, came from liberal arts backgrounds (Scott himself studied religious English poetry).” What I learned in 3 hours about doing great data journalism at the New York Times and ProPublica (Roberto Rocha, CBC Montreal)
  • “This 3TB+ dataset comprises the largest released source of GitHub activity to date. It contains activity data for more than 2.8 million open source GitHub repositories including more than 145 million unique commits, over 2 billion different file paths and the contents of the latest revision for 163 million files, all of which are searchable with regular expressions.” Making open source data more available (Github)






Labels: , , , , , , , , ,

1 comment:

  1. I laughed, I cried, I learned. A+ curator. Would consume again!