Call for Papers

The increasing use of machine learning (ML) combined with the software architectural challenges of the modern era has given rise to two broad research areas of investigation: i) software architecture for ML-enabled systems (SA4ML) and ii) using ML techniques to improve the architecting of any software system (ML4SA). For several years, both the research and practitioner communities have now explored topics at the intersection of SA and ML, and important contributions from both communities have emerged. However, these contributions are scattered across different venues of software engineering, self-adaptation, ML, etc., and especially the rise of large language models (LLMs) has introduced new opportunities and challenges. Moreover, it is not clear how to take advantage of existing techniques while clearly defining open, industry-relevant research problems.

The goal of SAML is therefore to bring together practitioners and researchers in one common platform to explore i) relevant SA techniques and practices for architecting ML-enabled systems, ii) how to leverage ML techniques to better architect software systems, and iii) potential synergies of these two areas.

The field of ML and AI evolved rapidly, and GenAI/LLMs are of high interest within our community, the SAML scope also includes all kinds of AI.

Thus, for this 5th-edition anniversary of SAML, we also choose a workshop theme for the first time: LLM Agents and Software Architecture. Other relevant topics include but are not limited to:

  • Architecture design & evaluation of ML-enabled systems
  • Architecture frameworks, patterns, and models for ML-enabled systems
  • Quality attributes of ML-enabled systems
  • Architecture and technical debt in ML-enabled systems
  • Architecting data or ML pipelines
  • Software architecture for LLM and agentic LLM systems
  • Software architecture and MLOps practices
  • Social and organizational aspects of architecting ML-enabled systems
  • Role of the software architect for ML-enabled systems
  • Architecting self-adaptive systems using ML
  • ML for synthesizing architecture documentation
  • ML for detecting architecture and design (anti-)patterns
  • ML for guiding or conducting architectural refactorings
  • ML for architecture evaluation
  • ML for architectural knowledge extraction
  • LLMs for software architecture design and analysis

Types of Contributions

We solicit two types of submissions for SAML:

  • Registered reports (RRs) i.e., detailed study designs that are submitted for feedback before study execution (full research papers without Results and Discussion sections). Accepted RRs have a clear path to publishing a full paper in ACM TOSEM, our journal collaboration partner. RRs must not exceed 6 pages + 2 pages for references. They will mostly be evaluated based on the significance and novelty of the hypotheses or techniques, and the soundness and reproducibility of the methodology specified to validate the claims or hypotheses. More details about RRs and the expected timeline will be described below.

  • Short papers e.g., industry experience reports or position, vision, or reflection papers, which must not exceed 5 pages + 1 page for references. They will be evaluated based on their relevance, novelty, soundness, and presentation.

All contributions should be submitted before the submission deadline using ICSA’s EasyChair online submission site (select the track “5th International Workshop on Software Architecture and Machine Learning”). Submissions must follow the IEEE conference proceedings format and abide by the single-anonymous submission process (author identities are known). All submissions must be original work and must not have been previously published, nor be under consideration for publication elsewhere. Artifacts related to the paper, if any, can be submitted by uploading the data/code to repositories such as Zenodo or Figshare, with the link to these contributions included in the paper.

Note that SAML explicitly moved away from soliciting full research papers, which we do not consider suitable for workshop publications. Instead, we invite submissions that align well with the interactive and discussion-focused format of a workshop.

Proceedings

The accepted papers will be published in ICSA 2026 Companion proceedings, and appear in IEEE Xplore Digital Library. At least one author of each accepted paper should register for the conference and participate in the workshop to present the paper in person. Papers that are not presented will be excluded from the proceedings.

Submission Deadlines

All deadlines are 11.59 pm UTC -12h (“Anywhere on Earth”).

Paper Deadline: February 13, 2026 February 20, 2026

Notification of acceptance: March 13, 2026

Camera ready due: March 27, 2026

Workshop dates: Either June 22 or June 23, 2026 (TBA)

Registered Reports & Expected Timeline

A registered report (RR) is a type of scientific publication that focuses on the importance of the significance and design of the planned study rather than the study outcomes. It makes use of preregistration to submit a study design for review before study execution (stage 1 publication). Think of it as a full research paper without the Results and Discussion sections. The goals of RRs are to receive early feedback on the study design, to prevent problems like p-hacking or reviewer bias against negative results, and to allow a structured path to publishing a full paper in a journal after study execution (stage 2 publication). Naturally, RRs are more suited to confirmatory and quantitative research, but it is possible to publish an RR for more exploratory and qualitative research. A more in-depth explanation of RRs and their use in software engineering has been compiled by Ernst and Baldassarre (2023). For more details on how our collaboration partner ACM TOSEM handles RRs, please refer to their website.

With SAML, publishing RRs will play out according to the following process:

  1. SAML solicits RR stage 1 submissions, i.e., the study design descriptions (deadline: Feb 13, 2026).
  2. SAML reviewers that agree to review RR submissions also agree to review the stage 2 submissions at TOSEM later on, i.e., the full papers written after study execution.
  3. Each RR stage 1 submission at SAML gets one of three outcomes: reject, accept with revisions, or accept as is (notification: March 13, 2026).
  4. All accepted RR stage 1 submissions are published in the ICSA companion proceedings (camera-ready deadline: March 27, 2026).
  5. At the SAML workshop, the accepted RR stage 1 submissions are presented and discussed, with potential further improvements to the study design.
  6. The authors submit these revised RR stage 1 papers to TOSEM in the RR track (deadline: July 22, 2026, but the earlier the better).
  7. The reviewer contact details for each accepted RR submission are forwarded to TOSEM.
  8. TOSEM invites the reviewers. If there were no (requested) changes, they greenlight the submission. Otherwise, they check the updates. Sometimes, there is a minor revision. The stage 1 submission is not published again at TOSEM.
  9. The RR stage 1 submission is accepted at TOSEM, which forms an agreement between authors and reviewers, i.e., if the described study design is followed and deviations are reasonably explained, the paper will be accepted, irrespective of the study results. The authors then execute the study and write the paper about the results.
  10. The authors submit the stage 2 submission, i.e., the full paper, to TOSEM, and the same reviewers are invited. It is treated internally as a major revision (deadline: ~6 months after the stage 1 RR acceptance, i.e., roughly Jan 2027).
  11. If the authors did everything according to plan or can convincingly justify deviations, the paper gets accepted, even with negative results. A minor or major revision may be used to request additional justifications for deviations or to address substantial issues regarding presentation or discussion of results. A rejection will only be used if the authors deviated strongly from their initial study design without reasonable justifications.