Over the past few years, machine learning (ML) algorithms have become increasingly prevalent as tools for data analysis and decision-making. Motivated by the rapid advances in the field of machine learning, the RSS
Discussion Meetings Committee, along with the RSS Sections on
Computational Statistics and Machine Learning,
Applied Probability and
Data Ethics and Governance, are jointly inviting submissions of discussion papers on probabilistic and statistical perspectives on machine learning. These papers can cover any probabilistic or statistical aspect of the methodological development or theoretical underpinning of machine learning algorithms. This could include methodological advances in machine learning, mathematical insights into existing ML algorithms, and data ethics for ML systems addressing the sociological impacts of ML. Examples of topics of interest include (but are not restricted to):
- Generalisation and robustness in machine learning;
- Theoretical guarantees for machine learning algorithms;
- Interpretable machine learning;
- Statistical perspectives on deep learning;
- Featurisation and dimension reduction in machine learning;
- Generative and probabilistic approaches to privacy;
- Novel mathematical machine learning models (such as neural SDEs, geometric deep learning).
To be considered, papers should have both a strong probabilistic or statistical component, and where appropriate, exploration of related ethics, policy, or governance implications.
Papers selected for publication will be presented at a multi-paper discussion meeting held at the RSS International Conference in September 2023, and subsequently published in one of the series of Journal of the Royal Statistical Society (JRSS), together with all contributions to the discussion at the meeting itself or submitted in writing shortly afterwards. All submitted papers are refereed, both for their scientific quality and their potential to generate discussion. Papers that meet the first criterion but not the second may, with the agreement of the authors, be referred to the editors of the Journal for review as a regular paper.
Submitted papers should be substantially shorter than is typical for a single‐paper discussion meeting (16 pages max excluding supplementary material following standard JRSS formatting instructions). We shall employ a two-step process to expedite the peer review process.
Full details are as follows:
- Abstract submission. Authors are invited to send a single‐page abstract (400 words max) of their proposed paper to Judith Shorten, the RSS Journals Manager (journal[at]rss.org.uk) by 20 May 2022.
- Full paper submission. Notification of accepted abstracts will be made by 28 June 2022 together with an invitation to submit a full paper. Full papers (16 pages max) should then be submitted via manuscript central to the most appropriate journal series (A, B or C) selecting the “Discussion Paper” option. The deadline for full paper submission is 4 November 2022.
- Refereeing. All papers received by 4 November 2022 will be refereed using the Society’s standard criteria for discussion meeting papers (scientific quality and potential to generate discussion).
- Final versions of accepted papers will be ready for pre‐printing by mid‐2023.
Informal enquiries about the call can be made by email to the Discussion Papers Editor, Adam Sykulski at
a.sykulski@lancaster.ac.uk.