cc @thirdreviewer@twitter.com @YourPaperSucks@twitter.com, I cited both of you to appease the wrath of the reviewers

After being awfully inefficient for some time, I figured out some tips for peer-reviewing papers: matbesancon.github.io/post/201

A talk on graphs for the study of trade in the 18th century French trade by @paulanomalie@twitter.com

Woops train delay, wait for us 😨

Math. optimization can be tricky, reasoning in high dimensions often results in me messing up. Appreciation to the @JuMPjl@twitter.com ecosystem for making the process interactive and easy to iterate on and high-performance to write libraries in

2019 @JuliaConOrg@twitter.com call for proposal is open. See you in July, Baltimore pretalx.com/juliacon2019/

Starting out with stochastic optimization for network & transportation

Lisbon clearly wins the is-it-summer-yet game 🇵🇹

Olá Lisbon, nice to meet you 🇵🇹

The internet used to be a city: it had its weird bohemian districts, its flea markets, its high-rises and back alleys. Now it's just a shopping mall, and I miss my old city so much

New year, new readings. Game theory is nice, but it's essential to keep up with the larger scale: just ordered Hedge, @Nicolas_Colin@twitter.com & The Entrepreneurial State @MazzucatoM@twitter.com

Also, many elements presented are US-specific. As a European, it seems obvious that the public authority has *some* role to play for... well the public good 🙄

Sharp, brutal, brillant essay by @annehelen@twitter.com. It made me look in a mirror couple times along the way, but totally worth the read, even if you miss your podcast, meditation or todo-list prep for tomorrow
buzzfeednews.com/article/anneh

Feel free to read the whole related article:
web.stanford.edu/~vcs/papers/C

"An article about computational science [...] is not the scholarship itself [...]. The actual scholarship is the complete software development environment and the complete set of instructions [...]" Jonathan Buckheit and David Donoho

Having to trust people claiming their method is the fastest in a context feels very unscientific. Even if that's the case, the efficiency usually relies on implementation tricks, often considers "unworthy" of mentioning in the paper

The discarded one claims high perf on the class of problem of interest, but does not leave software available. The other one lets me see for myself by cloning their project, reading through the code and trying it on instances

Dear community, today I ready two papers on the same subject, oriented on performance of models. I am discarding one of them, although excellent. Why? 👇

A Mastodon instance for maths people. The kind of people who make $\pi z^2 \times a$ jokes.
Use $ and $ for inline LaTeX, and $ and $ for display mode.