By: Julia Developers
The first ever JuliaCon was held in Chicago last year and was a great success. JuliaCon is back for 2015, this time in Cambridge, Massachusetts at MIT’s architecturally-delightful Stata Center, the home of computer science at MIT. Last year we had a single-track format, but this year we’ve expanded into a four-day extravaganza:
* On Wednesday 24th there will an introduction to Julia workshop run by David P. Sanders (@dpsanders) as well as a Julia hackathon – a great chance to get some help for your new Julia projects, or to begin contributing to Julia or its many packages.
* On Thursday 25th and Friday 26th we will be having speakers talking about a range of topics – we were fortunate to have so many fantastic submissions that we had to open up a second track of talks. The near-final schedule is on the main page. We’ll be alternating between ~40 minute long “regular” talks, and ~10 minute long “lightning” talks across all the sessions.
* On Saturday 27th we will finish with a series of workshops on a range of topics: data wrangling and visualization, optimization, high-performance computing and more. These workshops run from 1.5 to 3 hours and will be a great way to rapidly boost your Julia skills.
After getting everyone settled in, we’ll start the conference proper with a session about the use of Julia in a wide variety of scientific applications. Many of the talks at the conference focus on Julia package organizations: groupings of similar packages that promote interoperability and focussing of efforts. In the session Daniel C. Jones (@dcjones), the creator of the visualization package Gadfly, will discuss the advances being made in the BioJulia bioinformatics organization, and Kyle Barbary (@kbarbary) will present JuliaAstro, a home for astronomy and astrophysics packages. Theres something for everyone: quantitative economic modeling (QuantEcon.jl), quantum statistical simulations, and how to fit Julia into a pre-existing body of code in other languages.
After lunch we’ll be splitting into two tracks: visualization and interactivity and statistics. The visualization track will be demonstrating some of the exciting advances being made that enable Julia to both produce high-quality visualizations, but also share them. Mike Innes (@one-more-minute), creator of the Juno IDE for Julia, will be sharing his working on building web-powered apps in Julia, while Viral B. Shah (@ViralBShah), one of the Julia founders, will be discussing more about the inner workings of and plans for JuliaBox. For a different take on “visualization”, Jack Minardi of Voxel8 will be sharing how Julia is powering their 3D printing work.
The statistics session covers some hot topics in the field, including two talks from researchers at MIT about how Julia is playing a big part: probabilistic programming (Sigma.jl) and deep learning (Mocha.jl). Facebooker John Myles White, author of “Machine Learning for Hackers” and a variety of packages in R and Julia, will share his thoughts on how statistics in Julia can be taken to the next stage in development, and Pontus Stenetop (@ninjin) will educate and entertain in his talk “Suitably Naming a Child with Multiple Nationalities using Julia”.
We’ll come together at the end of Thursday to learn more about how to write good Julia code, how to write packages that Just Work on Windows, and how wrappers around C libraries can be made easier than you might think through the magic of Clang.jl. Iain Dunning (@IainNZ), maintainer of Julia’s package listing and test infrastructure will follow up on last years talk by giving a brief history and updated status report on Julia’s package ecosystem. Finally current Googler Lean Hanson (@astrieanna) will share some of her tips for people looking to get started with contributing to Julia and to open-source projects.
Whatever you get up to after the talks end on Thursday, make sure you are up in time for…
If you are interested in learning how Julia works from the people who work on it every day, then Friday morning’s session is for you. The morning will kick off with newly-minted-PhD and Julia co-founder Jeff Bezanson (@JeffBezanson), who is still recovering from his defense and will be updating us on the title of his talk soon. We’ll be learning more about different stages of the compilation process from contributors Jake Bolewski (@jakebolewski) and Jacob Quinn (@quinnj), and we’ll be covering a miscellany of other cutting-edge topics for Julia like tuning LLVM, debugging, and interfaces.
In the afternoon we’ll have four sessions split across two rooms. In the second scientific applications session we’ll be learning more about how Julia is being used to prevent airborne collisions from Lincoln Lab’s Robert Moss, and Iain Dunning (@IainNZ) will give a sequel to last years JuliaOpt talk to update us on how Julia is becoming the language of choice for many for optimization. We’ll also hear how Julia is enabling rapid development of advanced algorithms for simulating quantum systems, evolving graphs, and analyzing seismic waves.
The numerical computing track kicks of with Stanford’s Prof. Jack Poulson (@poulson), creator of the Elemental library for distributed-memory linear algebra. Right after, the linear algebra wizard Zhang Xianyi (@xianyi) will give a talk about OpenBLAS, the high-performance linear algebra library Julia ships with. After a break, we’ll hear Viral’s thoughts on how sparse matrices currently and should work in Julia, before finishing off with lightning talks about validated numerics and Taylor series.
We’ll see out the day with two sessions that hit some topics of interest to people deploying Julia into larger systems: data and parallel computing. In the data session we’ll learn how about the nuts and bolts of sharing and storing data in Julia and hear more about plans for the future by the contributors working in these areas. Make sure to check out the talk by Avik Sengupta (@aviks) about his real-world industry experiences about putting Julia code behind a web-accessible API.
The parallel computing session will tackle parallelism at all levels. Contributor Amit Murthy (@amitmurthy) will open the session with a discussion of his recent work and plans for managing Julia in a cluster. We’ll also hear about work being done to make Julia multithreaded at Intel, and about running Julia on a Cray supercomputer.
After all that you will surely be inspired to hack on Julia projects all night, but make sure to wake up for a full day of workshops on Saturday!
Remember to get your tickets and book your hotel before June 4th to take advantage of early bird pricing. We’d also like to thank our platinum sponsors: the Gordon and Betty Moore Foundation, BlackRock, and Julia Computing. We can’t forget out silver sponsors either: Intel and Invenia. We’re looking forward to seeing you there!