Call for Papers

Big data analytics has been in commercial use in search engines, social networks and online shopping and advertisement only in recent decades. In contrast, the scientific community has long relied on generating or recording massive amounts of data to be analyzed through high-performance computing (HPC) tools on supercomputers. Fortunately, the underlying technologies that build the backbone for big data analytics are now converging from algorithms to programming paradigms and software/system stacks to process, communicate and store data. Moreover, much like classic areas of computer science that are witnessing transformations from conventional paradigms to deep learning, many important applications in scientific computing relying on simulation (e.g., protein folding, fluid dynamics, brain simulation, food composition) are also witnessing a transformation to paradigms that adopt deep learning with abbreviated simulation to reduce turnaround and resource utilization and/or improve results.

The EuroLab4HPC workshop on the Convergence of ML and HPC will include invited talks from industry and academic institutions on technologies at the intersection of machine learning and high-performance computing. The workshop also solicits poster presentations on the following key areas of interest:

  • High-performance data analytics from algorithms to hardware
  • Language technologies for large-scale machine learning
  • Distributed machine learning algorithms
  • Machine learning techniques for scientific applications
  • Simulation-assisted machine learning
  • Visualization for high-performance data analytics

Authors are requested to submit their extended poster abstracts for review by 5pm CET on February 7th, 2020. Authors will be notified of the outcome by February 3rd. Submissions should be in PDF, and should not exceed two printed pages. Travel grants are available for students whose poster submissions are accepted for a presentation.

Submission Guidelines

Formatting rules as well as the Latex template for the submission are adapted from ASPLOS'20. Authors are encouraged to use the provided Latex template, and follow the rules given below:

  • Paper must be submitted in printable PDF format.
  • Text must be in a minimum 10pt font.
  • Papers should contain a maximum of 2 pages of single-spaced two-column text, including any appendixes, but not including references.
  • No page limit for references.

Please submit your extended poster abstracts at EasyChair page.

Important Dates

  • Submission deadline: 5pm CET January 24th February 7th, 2020
  • Author notification: February 14th, 2020
  • Workshop: March 17th, 2020

Registration

No need for registration. The workshop will be open to Eurolab4HPC, EcoCloud and the EPFL community.

To participate in the virtual poster sessions, please first sign up here for ASPLOS2020's Slack workspace, and browse the channel: #workshop-eurolab4hpc-industry-day-on-convergence-of-ml-and-hpc.

Program

10.00 am - 11.00 am   :      Tal Ben-Nun (ETHZ)   [Slides]

11.00 am - 11.30 am    :      Virtual break & Poster discussion on Slack

11.30 am - 12.30 pm    :      Felix Schürmann (Blue Brain)    [Video]

12.30 pm - 2.00 pm    :      Virtual break

2.00 pm - 3.00 pm     :      Gunter Roeth (Nvidia)   [Slides] [Video]

3.00 pm - 4.00 pm     :      Thomas Parnell (IBM Zurich)   [Slides] [Video]

4.00 pm - 5.00 pm     :      Virtual break & Poster discussion on Slack

5.00 pm - 6.00 pm     :      Paolo Faraboschi (HPE)   [Slides] [Video]

Sponsors