DSML 2020
Dependable and Secure Machine Learning

Topics of Interest

  • Testing, certification, and verification of ML models and algorithms
  • Metrics for benchmarking the dependability and security of ML systems
  • Adversarial machine learning (an emphasis will be placed on defenses)
  • Resilient and repairable ML models and algorithms, including mechanisms for failsafe defaults and smooth degradation of performance
  • Reliability and security of ML architectures, computing platforms, and distributed systems
  • Faults in implementation of ML algorithms and their consequences
  • Dependability of ML accelerators and hardware platforms
  • Safety and societal impact of machine learning
  • Testing, certification, and verification of ML models and algorithms

Important Dates

Website open for Submission: 3 March, 2020 (AOE)
Submission Deadline: 7 March, 2020 (AOE)7 April, 2020 (AOE)
Notification of Acceptance:21 April, 2020
Camera Ready:28 April, 2020
Workshop:29 June, 2020

Submissions

DSML welcomes both research papers reporting results from mature work, and more speculative papers describing new ideas with preliminary exploratory work. Papers reporting industry experiences, case studies, and datasets will also be encouraged. This year, we are also soliciting proposals for research talks based on work previously published elsewhere (reference to previous work is required). We strongly encourage these research talks to also include new ideas and provocative opinions and not just summarize previous work that is already published. Specifically, we accept submissions in the following formats:
  • Regular research papers (up to 6 pages + 3 pages for references and supplementary material)
  • Proposals for research talks (1 page + 3 pages for references and supplementary material)
All submissions should be in PDF format and must adhere to the IEEE Computer Society 8.5x11 two-column camera-ready format (using a 10-point font on 12-point single-spaced leading). Both LaTeX and MS Word templates are available here: https://www.ieee.org/conferences_events/conferences/publishing/templates.html
We will use a double-blind review process only for the regular research papers, so the authors must anonymize their submissions. The first page must include the title of the paper, but no information on authors names and affiliations. Research talks need not be anonymized.

Submission site: https://dsn-dsml20.hotcrp.com/

All submitted manuscripts will be peer-reviewed by the program committee. Papers will be accepted and included in the workshop program according to the following criteria: relevance of the addressed topic to the scope of the workshop; novelty and value of the proposed contribution; scientific merit; quality of the writing, presentation accuracy and style.

Proceedings

Authors of regular papers can select either of the following options for the publication of their accepted papers:
(1) Paper will appear in the supplementary DSN proceedings (archived in the IEEE Digital library), with the same page limit constraints as specified above,
(2) Only an extended abstract (up to 2 pages + 3 pages for references and supplementary material) of the paper will be included in the supplementary DSN proceedings, but the authors are required to post a full version of the paper on arxiv that will be linked from the workshop website.

Important Announcement

Given the uncertainty surrounding travel because of COVID-19, we understand that authors may be unclear about whether or not they will be able to attend in person. We will be making accommodations for virtual presentations and other alternatives.