Machine Learning and the Physical Sciences

This website was used for the 2019 instance of this workshop.
Please visit ml4physicalsciences.github.io for up-to-date information.

DECEMBER 14, 2019 AT WEST MEETING ROOMS 109 & 110

ACCEPTED PAPERS ARE NOW ONLINE: SEE BELOW

TALK VIDEOS ARE NOW ONLINE: 1 2 3 4

About

Machine learning methods have had great success in learning complex representations of data that enable novel modeling and data processing approaches in many scientific disciplines. Physical sciences span problems and challenges at all scales in the universe: from finding exoplanets in trillions of sky pixels, to finding machine learning inspired solutions to the quantum many-body problem, to detecting anomalies in event streams from the Large Hadron Collider, to predicting how extreme weather events will vary with climate change. Tackling a number of associated data-intensive tasks including, but not limited to, segmentation, 3D computer vision, sequence modeling, causal reasoning, generative modeling, and efficient probabilistic inference are critical for furthering scientific discovery. In addition to using machine learning models for scientific discovery, the ability to interpret what a model has learned is receiving an increasing amount of attention.

In this targeted workshop, we aim to bring together computer scientists, mathematicians and physical scientists who are interested in applying machine learning to various outstanding physical problems, including in inverse problems; approximating physical processes; understanding what a learned model really represents; and connecting tools and insights from the physical sciences to the study of machine learning models. In particular, the workshop invites researchers to contribute short papers that demonstrate cutting-edge progress in the application of machine learning techniques to real-world problems in the physical sciences, and using physical insights to understand what the learned model represents.

By bringing together machine learning researchers and physical scientists who apply machine learning, we expect to strengthen the interdisciplinary dialogue, introduce exciting new open problems to the broader community, and stimulate the production of new approaches to solving challenging open problems in the sciences. Invited talks from leading individuals in both communities will cover the state-of-the-art techniques and set the stage for this workshop.

NeurIPS 2019

The Machine Learning and the Physical Sciences 2019 workshop will be held on December 14, 2019 as a part of the 33rd Annual Conference on Neural Information Processing Systems, at the Vancouver Convention Center, Vancouver, Canada. Please check the main conference website and FAQ for information about registration, schedule, venue, and other arrangements. More information in the Registration section below.

Schedule

Organizers

• 

  Atılım Güneş Baydin
  University of Oxford

• 

  Juan Felipe Carrasquilla
  Vector Institute / University of Waterloo

• 

  Shirley Ho
  Flatiron Institute / Princeton University

• 

  Karthik Kashinath
  NERSC, Berkeley Lab

• 

  Michela Paganini
  Facebook AI Research

• 

  Savannah Thais
  Princeton University / IRIS-HEP

Steering Committee

• 

  Anima Anandkumar
  California Institute of Technology / NVIDIA

• 

  Kyle Cranmer
  New York University

• 

  Roger Melko
  University of Waterloo

• 

  Prabhat
  NERSC, Berkeley Lab

• 

  Frank Wood
  University of British Columbia

Call for papers

We invite researchers to submit work particularly in the following and related areas:

• Application of machine and deep learning to physical sciences
• Generative models
• Likelihood-free inference
• Variational inference
• Simulation-based models
• Implicit models
• Probabilistic models
• Model interpretability
• Approximate Bayesian computation
• Strategies for incorporating prior scientific knowledge into machine learning algorithms
• Experimental design
• Any other area related to the subject of the workshop

Submissions of completed projects as well as high-quality works in progress are welcome. All accepted short papers will be made available on the workshop website. This does not constitute an archival publication or formal proceedings; authors retain full copyright of their work and are free to publish their extended work in another journal or conference. We allow submission of short papers that overlap with papers that are under review or have been recently published in a conference or a journal. However, we do not accept cross submissions of the same short paper to multiple workshops at NeurIPS. Submissions will be kept confidential until they are accepted and authors confirm that they can be included in the workshop. If a submission is not accepted, or withdrawn for any reason, it will be kept confidential and not made public.

Accepted submissions will be presented as posters during the workshop. Several accepted submissions will be selected for contributed talks.

Submission instructions

Submissions should be anonymized short papers up to 4 pages in PDF format, typeset using the NeurIPS style. References do not count towards the page limit. Appendices are discouraged, and reviewers are not expected to read beyond the first 4 pages. A workshop-specific modified NeurIPS style file will be provided for the camera-ready versions, after the author notification date.

Please note that at least one coauthor of each accepted paper will be expected to attend the workshop in person to present a poster or give a contributed talk. Submissions page is here.

Submit paper

Instructions for Accepted Papers, Posters, and Digital Acceptances

Accepted papers (either as talks, posters, or digital acceptances)

Please produce a "camera-ready" (final) version of your accepted paper by replacing the "neurips_2019.sty" style file with the "neurips_2019_ml4ps.sty" file available here and using the "final" package option (that is, "\usepackage[final]{neurips_2019_ml4ps}") to include author and affiliation information. The modified style file replaces the first page footer to refer to the workshop instead of the main conference. It is acceptable if your paper goes up to five pages (excluding the references or any appendices) due to author and affiliation information taking extra space on the first page.

Please revise your paper as much as you can to reasonably address reviewer comments. The revision would include minor corrections and/or changes to directly address reviewer comments. Beyond these points, it is not acceptable to include any significant new material that was not present in the reviewed version of your paper.

Please upload the final PDF of your paper as an updated version in your existing submission on EasyChair by the camera-ready deadline.

Posters at the Workshop

If your paper was accepted as a poster presentation, we expect that you produce a poster and present it at one of the two poster sessions (morning and afternoon) at the workshop. The workshop schedule will show which session your poster is assigned to. NeurIPS suggests that you use A0 size in portrait mode and no larger than 36W x 48H inches or 90 x 122 cm. Posters are taped to the wall. Posters should be on lightweight paper, not laminated. NeurIPS will provide the tape. All posters will be presented inside the room the workshop takes place in and there are no poster boards at workshops.

Accepted Talks

If your paper was accepted as a contributed talk, presenting a poster at the workshop is optional, meaning that you can prepare and present a poster at the workshop if you choose to do so.

Digital Acceptances

If your paper was accepted as a "digital acceptance", this means that you won't have a possibility to present the work at the workshop in person. However, your paper will be listed in the workshop website along with the other accepted papers. Please produce your camera-ready paper as instructed above. Optionally you can send us, via this online form a single slide in PDF format that will serve as a spotlight of your work that we will show as a cycling presentation during the workshop poster sessions. These single slides should include the names of co-authors and paper title at the bottom of the slide and be preferably in 16:9 aspect ratio.

Travel support

Thanks to our sponsors Vector Institute, Flatiron Institute, IRIS-HEP, and DeepMind, we will be able to offer travel support for a number of presenters with posters and/or talks at the workshop. Presenters indicate their interest using a web form shared with co-authors of accepted papers. This excludes papers in the digital acceptance category because these are not going to be presented in person. Funding will be distributed based on need with preference given to junior researchers and those from under-represented backgrounds.

Important dates

• Submission deadline: September 9, 2019 September 16, 2019, 23:59 PDT
• Author notification: October 1, 2019
• Camera-ready (final) paper deadline: November 1, 2019, 23:59 PDT
• Workshop: December 14, 2019

Registration

NeurIPS conference has three main sessions (Tutorials, Conference, Workshops) to which you can register. You need to be registered to at least the Workshop session in order to be able to attend this workshop. For pricing information see NeurIPS 2019 Pricing. Registrations this year are handled using a randomized lottery, due to the demand for registrations surpassing the available capacity (this led to registrations selling out under 12 minutes for NeurIPS 2018). The official information regarding this can be accessed here.

Between September 6 and September 20, 2019, those wishing to attend NeurIPS will indicate their intention to register (intention to purchase a ticket) by signing up for the lottery at the main conference home page. On September 20, 2019, registration invitations will be sent to people randomly picked from the lottery pool, which will allow them to register (purchase a ticket) within two weeks. The invitation will expire after two weeks.

In addition to—and independent from—the regular lottery registration, this workshop will have access to a number of registration slots reserved for workshop presenters (oral or poster), which we will assign to the main presenters of accepted work after the author notification date of October 1, 2019. We will do our best to ensure that the main presenters of all accepted work will get a registration invitation through this channel (they will still need to purchase a ticket), if they were not able to register through the lottery. However, this is not guaranteed and the number of reserved registrations this workshop will get access to is beyond our control. THEREFORE WE STRONGLY ENCOURAGE THE MAIN PRESENTERS OF ALL SUBMITTED WORK TO ENTER THE REGISTRATION LOTTERY IN ORDER TO ENSURE THAT THEY GET A REGISTRATION INVITATION THROUGH AT LEAST ONE CHANNEL.

Papers

Sponsors

• 

• 

• 

• 

• 

Location

Vancouver Convention Centre, 1055 Canada Pl, Vancouver, BC, V6C 0C3, Canada