Image source: https://www.freevector.com/psychedelic-pattern-28724

Call for Papers

Learning with Small Data (LSD) will bring together researchers from various areas to discuss  the sustainability of current state of the art methods in computational linguistics which rely on very large models, such as GPT2-3, BERT, and XLNet. The conference encourages contributions from machine learning, computational linguistics, theoretical linguistics, philosophy, cognitive science, and psycholinguistics, as well as from artificial intelligence ethics and social policy. We hope to see innovative technical proposals,  and we will cultivate a wide spectrum of views within a lively dialog on the issues that the conference addresses.

Important Dates:

Submission deadline: 2023 May 5, anywhere on Earth

New submission deadline: 2023 May 12, anywhere on Earth

New submission deadline NON-ARCHIVAL EXTENDED ABSTRACTS: 2023 July 31, anywhere on Earth

Notification of acceptance: 2023 June 12, anywhere on Earth

Notification of acceptance: 2023 June 14, anywhere on Earth

Camera ready: 2023 August 14, anywhere on Earth

Conference: 2023 September 11-12, not anywhere on Earth, but in Gothenburg

Topics of interest

We welcome all relevant approaches to text-based and multimodal computational neural language modeling as well as psycholinguistic perspectives, neurolinguistic perspectives, ethical, and policy issues. Papers are invited on topics in these and closely related areas, including (but not limited to) the following:

•     small-scale neural language modeling, both text-only and multimodal

•     training corpus and test task development

•     visual, dialog and multi-modal inference systems

•     neurolinguistic and psycholinguistic experimental approaches to human language processing

•     semantics and pragmatics in neural models

•     dialog modeling and linguistic interaction

•     formal and theoretical approaches to language production and comprehension

•     language acquisition in the context of computational linguistics

•     statistical, machine learning, reinforcement learning and information theoretic approaches that embrace small data

•     methodologies and practices for annotating datasets

•     visual, dialog and multi-modal generation

•     text generation in both the dialog and monologue settings

•     semantics-pragmatics interface

•     social and ethical implications of the development and application of large or small neural language models, as well as relevant policy implications and debates.

 Submission Requirements

LSD 2023 will feature three types of submissions: long papers, student papers, and short papers. All types of papers should be submitted not later than 5 May 2023  12 May 2023. Long papers must describe original research, and they must not exceed 8 pages excluding references. They will be presented at the conference either orally or as posters. Student papers describe original research, and the first author must be a student, or at least 2/3 of the work on a paper should be done by students. Student papers must not exceed 6 pages excluding references. Reviewers will give special support to student authors through mentoring. The papers will be presented orally or as posters at the conference. Short papers present work in progress, or they describe systems and/or projects. They must not exceed 4 pages excluding references. They will be presented as posters at the conference and summarized in lightning talks. Position papers are also accepted. These should be formatted in the same way as long papers. All types of papers will be published in the 2023 ACL Anthology as a CLASP Conference Proceedings. 

Submissions should be pdf files and use the Latex or Word templates provided for ACL 2023 submissions. Submissions have to be anonymous.

Papers should be electronically submitted in PDF format via the softconf system at: https://softconf.com/n/lsd2023/. Please make sure that you select the right track when submitting your paper. Contact the organisers if you have problems using softconf.

Types of submissions

Short archival paper (4 pages + references)

Long archival paper (8 pages + references)

Student archival paper (6 pages + references)

New !!! Extended abstract non-archival (2 pages + references)

Non-archival submissions will be presented as posters. This is a great opportunity to get feedback on new work that is in progress or to present previously published work to a new audience. This means that the general rule that submissions need to be original unpublished work does not apply to this category. 

Concurrent Submissions

Papers that have been or will be submitted to other meetings or publications must indicate this at submission time using a footnote on the title page of the submissions. Authors of papers accepted for presentation at Learning with Small Data 2023 must notify the program chairs by the camera-ready deadline as to whether the paper will be presented. All accepted papers must be presented at the conference to appear in the proceedings. We will not accept for publication or presentation papers that overlap significantly in content or results with papers that will be (or have been) published elsewhere.

Camera Ready Versions

Camera ready versions should follow the same guidelines with respect to style and page numbers as the initial submission, i.e. there are no additional pages allowed in the final submission. Camera-ready versions of long, short and student papers will be given one additional page of content plus unlimited pages for acknowledgements and references. This means that long papers may have up to 9 pages of content, short papers up to 5 pages of content, and student papers up to 7 pages of content. Please submit the camera-ready version by 2023 August 14.