SCL@ECML2026 - Workshop on Streaming Continual Learning (SCL 2026) @ ECML 2026

24 views

Skip to first unread message

MARCO Piangerelli

unread,

Apr 8, 2026, 9:48:16 AM (12 days ago) Apr 8

to Machine Learning News

[Apologies for cross-posting]

-----------------------------------------------------

Call For Papers

-----------------------------------------------------

The Streaming Continual Learning (SCL) Workshop welcomes contributions on all aspects of learning in non-stationary and streaming environments, spanning both theoretical foundations and practical applications across different domains (e.g., IoT, robotics, finance, monitoring systems, basic sciences, and industrial applications), and across a wide range of data types (e.g., tabular data, time series, spatiotemporal data, graphs, logs, multimedia, and data streams).

Workshop adjunct proceedings published by Springer-Verlag:

Date: September 7th, 2025 - Naples (Italy)

Web: https://www.ecml26.scl-ai.org/home

-----------------------------------------------------

Important Dates

-----------------------------------------------------

Paper submission deadline: May 31st 2026

Notification to the authors: June 30th 2026

Camera ready submission: July 10th 2026

All Deadlines are Anywhere on Earth (AoE) at 23:59.

------------------------------------------------------

Workshop Aims and Scope

------------------------------------------------------

Real-world AI systems increasingly operate on continuous, high-velocity data streams where distributions evolve over time and temporal dependencies are present. Traditional "train-once, deploy-forever" paradigms break down, motivating learning systems that adapt continuously under strict computational and memory constraints. Streaming Continual Learning (SCL) is emerging as a unifying framework bridging Streaming Machine Learning (SML) and Continual Learning (CL), combining rapid online adaptation with selective knowledge retention. Interest in SCL is rapidly growing, building on a history of community events including the ESANN 2025 special session and the AAAI SCL bridge, which attracted strong participation across CL, SML, and related areas.

The workshop solicits contributions on learning paradigms (CL, SML, online/OCL), adaptation under drift, temporal and sequential learning, evaluation and benchmarking, and emerging directions such as continuous adaptation of foundation models, reinforcement learning, and resource-constrained/edge learning.

Through oral and poster presentations, invited talks, and open discussion sessions, the workshop fosters interaction and exchange among researchers working on dynamic environments and evolving data streams. Uniting these communities provides a shared forum to identify open challenges, promising research directions, and practical solutions for Streaming Continual Learning.

--------------------------------------------------------

Workshop Topics

SCL encourages submissions presenting novel algorithms, architectures, evaluation frameworks, benchmarks, and application-driven studies, as well as position papers and perspectives outlining open challenges and future research directions in SCL. Topics include, but are not limited to:

Learning paradigms

Continual and lifelong learning
Streaming Machine Learning
Online learning in non-stationary environments
Online Continual Learning
Streaming Continual Learning: Unified perspectives on Streaming Machine Learning and Continual Learning
Adaptation and Non-Stationarity
AutoML/Automated adaptation

Concept drift detection and adaptation

Real and virtual drift detection/handling
Transfer learning in streaming settings
Domain adaptation and test-time adaptation under distribution shift

Temporal and Sequential Learning

Time series analysis in streaming settings
Learning continuously with temporal dependence
Sequential and temporally-aware continuous learning

Evaluation and Benchmarking

Design of realistic Streaming Continual Learning benchmarks
Evaluation protocols and metrics for streaming and continual learning
Trade-offs between accuracy, adaptation speed, memory, and computation Emerging Directions in Streaming Settings

Continuous learning and adaptation of Foundation Models
Streaming Continual Learning applied to Reinforcement Learning
Efficient updates without retraining from scratch
Resource-constrained and edge learning scenarios

-------------------------------------------------------

Submission and Publication

-------------------------------------------------------

We welcome submissions of completed research, preliminary results, and innovative ideas in the form of full and short papers. Submitted papers should not have been previously published or accepted for publication in substantially similar form in any peer-reviewed venue, such as journals, conferences, or workshops. Papers must be written in English and formatted in LaTeX, following the outline of ECML-PKDD author kit.

We will consider two submission types:

Full papers (up to 12 pages) may include research contributions, reproducibility or replicability studies, as well as case studies. They should clearly situate the work within the state of the art and describe the proposed methodology, experimental setting, or real-world application in sufficient detail.
Short papers (up to 8 pages) include position papers, preliminary or ongoing work, and practice or experience reports. They may introduce new perspectives on the workshop topics or describe real-world scenarios and lessons learned.

Submissions should not exceed the indicated pages, including any diagrams and references. All submissions will go through a double-blind review process and will be reviewed by at least two reviewers based on relevance for the workshop, novelty/originality, significance, technical quality and correctness, quality and clarity of presentation, quality of references, and reproducibility. Submitted papers will be rejected without review if they are not properly anonymized, do not comply with the template, or do not follow the above guidelines.

The authors might have a (non-anonymous) pre-print published online, but it should not be cited in the submitted paper to preserve anonymity. Reviewers will be asked not to search for them. The authors are also strongly encouraged to adhere to the best practices of Reproducible Research, by making available data and software tools that would enable others to reproduce the results reported in their papers. Tools such as https://anonymous.4open.science are encouraged to be used to ensure the anonymity of the submitted code.

The accepted papers and the material generated during the meeting will be available on the workshop website. The accepted papers will be included in a joint Post-Workshop proceeding published by Springer Communications in Computer and Information Science, in 1-2 volumes, organized by focused scope. Moreover, the authors of selected papers may be invited to submit an extended version in a special issue hosted by a top-ranking journal.

Notice: According to the ECML PKDD 2026 policy, the SCL Workshop is an in-person workshop, so each accepted workshop paper must be presented in person. Each accepted workshop paper must be accompanied by at least one distinct full author registration, completed by the early registration date cut-off and must be presented at the workshop even if they opt-out of the post-proceedings. We expect the authors, the program committee, and the organizing committee to adhere to the ECML-PKDD Code of Conduct.

----------------------------------------------------------

Workshop Chairs

---------------------------------------------------------

Marco Piangerelli

University of Camerino, Camerino (Italy) and Vici & C. S.p.a.

Email: marco.pi...@unicam.it

Maroua Bahari

Sorbonne Université, Paris (France)

Email: maroua...@lip6.fr

Federico Giannini

DEIB - Politecnico di Milano

Email: federico...@polimi.it