WAX is a workshop on approximate computing, a research direction that asks how computer systems can be made better—faster, more efficient, and less complex—by relaxing the requirement that they be exactly correct. Approximation arises from sources as diverse as sensors, machine learning algorithms, and big data applications. Approximate systems raise questions from across the system stack, from circuits to applications. WAX is a venue for discussion, debate, and brainstorming on all of these topics.

Here’s a (tentative) timeline for the workshop:

  • May 1, 2019: position papers due via HotCRP
  • May 10, 2019: position paper notification
  • June 22, 2019: WAX!

More details about WAX and the call for position papers are available at http://approximate.computer/wax2019/

Sat 22 Jun

pldi-2019-catering
08:00 - 09:00: Catering - Breakfast at 301 Foyer
pldi-2019-catering08:00 - 09:00
Other
WAX-2019-papers
09:00 - 09:15: WAX 2019 - Opening and Introduction at 105C
Chair(s): Ulya KarpuzcuUniversity of Minnesota, USA
WAX-2019-papers
09:15 - 10:15: WAX 2019 - Keynote: Song Han, MIT at 105C
WAX-2019-papers
10:15 - 11:00: WAX 2019 - Talks 1 at 105C
WAX-2019-papers10:15 - 10:25
Talk
Jesse MichelMassachusetts Institute of Technology, Sahil VermaIIT Kanpur, Benjamin ShermanMassachusetts Institute of Technology, USA, Michael CarbinMassachusetts Institute of Technology
WAX-2019-papers10:25 - 10:35
Talk
WAX-2019-papers10:35 - 10:45
Talk
Samuel Triest, Daniel Nikolov, Jannick RollandUniversity of Rochester, Yuhao ZhuUniversity of Rochester
pldi-2019-catering
11:00 - 11:20: Catering - Coffee Break at 301 Foyer
pldi-2019-catering11:00 - 11:20
Coffee break
WAX-2019-papers
11:20 - 12:30: WAX 2019 - Talks 2 at 105C
WAX-2019-papers11:20 - 11:30
Talk
Vimuth FernandoUniversity of Illinois at Urbana-Champaign, Keyur JoshiUniversity of Illinois at Urbana-Champaign, Darko MarinovUniversity of Illinois at Urbana-Champaign, Sasa MisailovicUniversity of Illinois at Urbana-Champaign
WAX-2019-papers11:30 - 12:00
Talk
Yipeng HuangPrinceton University
Link to publication Pre-print
WAX-2019-papers12:00 - 12:30
Talk
Yongjoo ParkUniversity of Michigan
pldi-2019-catering
12:30 - 14:00: Catering - Lunch at 301A
WAX-2019-papers
13:30 - 14:30: WAX 2019 - Discussion at 105C

Call for Papers

WAX is a workshop on approximate computing, a research direction that asks how computer systems can be made better—faster, more efficient, and less complex—by relaxing the requirement that they be exactly correct. Approximation arises from sources as diverse as sensors, machine learning algorithms, and big data applications. Approximate systems raise questions from across the system stack, from circuits to applications. WAX is a venue for discussion, debate, and brainstorming on all of these topics.

Topic

With transistor scaling becoming less effective at improving computer system performance and energy efficiency, we urgently need new paths forward for expanding the capabilities of computers. Trading off accuracy for better performance and energy efficiency is an attractive option for many important and resource-hungry applications, including image and video processing, computer vision, machine learning, simulations, big data analytics, embedded systems, etc. For that reason, approximate computing has become a “hot topic,” with active research in computer architecture, programming languages, operating systems and user-facing areas such as ubiquitous computing and HCI.

Making approximate computing successful requires cooperation among all layers of the stack, from algorithms to programming languages to OSes to architecture to circuits, as well as system components like storage and networks. This workshop aims to bring together an interdisciplinary group of researchers to present and discuss thoughts and ideas on how to effectively exploit approximate computing.

Topics for WAX include:

  • Hardware support for approximate computing
  • Programming languages and compiler support for approximate computing
  • Tools for writing, debugging, and testing approximate programs
  • Modeling and understanding approximate computing opportunities and systems
  • Applications amenable to approximation and domain-specific strategies
  • Formal reasoning about programs with approximations
  • Retrospectives on past approximate-computing work, including both reflections on your own past projects and reproduction of others’ results
  • Position papers on approximate computing, potential, how it could fail, what we need to succeed, etc.

How to Participate

We invite participation in three forms: position papers, lightning talks, and discussion topics.

Peer-Reviewed Position Papers

The workshop will include a peer-reviewed program of short position papers. Papers can describe an early-stage research project, advocate an opinion about approximate computing, reflect on trends in the community, or reproduce someone else’s published result.

Position papers will go through a full peer-review process by a program committee of experts in approximate computing (see below). Papers will not be published in a proceedings, so they do not preclude future publication; instead, we will post PDFs on the workshop’s Web site. We also encourage authors of accepted papers to include artifacts—code, data, analysis scripts, etc.—which we will also host alongside the papers.

Accepted papers will be presented in short talks, around 5 to 15 minutes. We will not publish the paper in a formal proceedings. Instead, we will post PDFs of accepted papers on the workshop’s Web site. Authors are also encouraged to submit supporting material (code, data dumps, etc.) after acceptance, which we will also host. We especially encourage talks with live demos of applications and working tools.

Papers should use the formatting guidelines for SIGPLAN conferences (the acmart format with the sigplan two-column option) and not exceed 2 pages, excluding references. Review is single-blind, so please include authors’ names on the submitted PDF.

Paper submission will is via HotCRP.

Lightning Talks

WAX will feature a session for short talks in the morning that present a single opinion, a nugget of an idea, or just food for thought. Speakers will have approximately 60 seconds. Talks will consist of two slides, one of which is a title slide (including at least the title itself and the speaker’s name, affiliation, and email address).

Lightning talks will not be peer reviewed. We will vet slides ahead of time to ensure that they are on topic, but there will be no full review process.

Discussion Topics

The WAX program will feature a debate among the attendees. We need your help building a list of controversial topics to serve as grist for the discussion mill.

Please submit a sentence or two about an open problem, philosophical question, or other thought you’d like to see discussed at the workshop. You can submit as many of these as you like. We’ll use these suggestions to set up a debate during the workshop.

Add your topic suggestions by editing this wiki page on GitHub.