The AI in Manufacturing workshop brings together researchers and practitioners from AI and the area of manufacturing, and especially also people working at the intersection. The workshop provides an opportunity for interaction and networking, and a platform for the dissemination of problem statements, early ideas and work-in-progress presentations for potentially ground-breaking research, in the broad area of AI with application in manufacturing. Our intention is to facilitate AI advances relevant for manufacturing. The scope includes theory, algorithms, systems, and applications related to this topic.
Submission deadline: June 20 July 4 (extended), 2022 (https://easychair.org/my/conference?conf=ai4manufacturing)
Author notification: July 20, 2022
Camera ready copy due: Aug 15, 2022
Workshop day: September 19, 2022
ECML-PKDD is organized as a hybrid conference. Details will need to be confirmed, but it should be possible to follow the workshops both in-person and online, but all presentations require on-site presence.
In manufacturing, the traditional process optimization, diagnostics and maintenance of systems depended largely on domain engineer's knowledge and expertise. This practice is not sufficient anymore for the contemporary manufacturing requirements. Machine Learning and AI may enhance the automation in monitoring, diagnostics, and control for High Volume Manufacturing, and provide domain engineers with the necessary tools for more effective process optimization. However, several key challenges are faced when applying ML and AI techniques in this field:
Often only small amounts of labelled data are available in contrast to very complex high-dimensional problem spaces, and this data may be biased or imbalanced to specific scenarios or contexts. To this end, techniques exploiting human experts and integration of domain knowledge in the learning or reasoning is studied and applied.
Also, most manufacturing use cases require interpretability and uncertainty quantifications because AI predictions are expected to contribute to decisions on fab processes with high financial impact. Thus, advanced visualisation and methods to enhance interpretability are necessary.
Finally, due to the physical complexity of the systems, problems like concept drift quickly arise, warping the problem space through time.
The aim of this workshop, held on September 19, 2022, in conjunction with the ECML-PKDD Conference, is to be a discussion forum for the most recent advances on these topics. We will encourage both theoretical and practical contributions to stimulate interactions between participants, by mixing longer mature contributions with short, open for discussion, ideas.
We invite contributions of the following types: short (work-in-progress or position) papers (5-9 pages) to present ongoing and preliminary works, long research papers (10-16 pages), and demos (3-6 pages). A selection of papers will be invited to give an oral presentation, all accepted papers shall be presented in the poster session.
Authors should indicate in their abstracts the kind of submission, to help reviewers better understand their contributions. Submissions must be in PDF, written in English, not longer than 16 pages (including references and supplemental material), and formatted according to the standard Springer LNCS style. All papers accepted to the workshop should be presented as a poster/demo during the poster session and several (to the extent that time allows; depending on the number of contributions) also given an oral presentation. The papers will be posted on the workshop website (unless the authors opt out), but will not be published in a formal proceeding. Given enough interest by the workshop participants, we will look to edit a special issue in a top journal.
Reviewing will be done single blind, so authors should list their names and affiliations, and relevant funding sources should be mentioned for example in an acknowledgements paragraph.
For accepted papers, at least one author must register for the conference and attend the workshop in person to present the work.
Submit via EasyChair: https://easychair.org/my/conference?conf=ai4manufacturing
Topics of interest are all aspects of AI for Manufacturing, including (but not limited to) methods and tools to address the key challenges for applications in semiconductor manufacturing:
ML when labelling is expensive: active learning, semi-supervised learning, weak supervision, transfer learning. Techniques to maximally learn from small sample sizes, together with expert input.
Advanced visualisations for human expert interactions and interpretability: Dimensionality reduction, representation learning, embeddings.
Embedding domain expertise in ML: graphical models, physics-aware ML architectures, integration of simulation models (digital twins+ML), or other combinations of reasoning and learning frameworks.
Uncertainty quantification and learning with noisy data, or learning with only positive and unlabeled examples.
Assessment of the impact of noise, bias, outliers on model performance, ML for improving data quality, data-centric AI.
Techniques for detecting and addressing concept drift.
Organizers: Dimitra Gkorou (ASML), Pieter Van Hertum (ASML), Jefrey Lijffijt (Ghent University), Mykola Pechenizkiy (TU Eindhoven), Joaquin Vanschoren (TU Eindhoven)
Website: https://sites.google.com/view/ai4manufacturing/home
Submission url: https://easychair.org/conferences/?conf=ai4manufacturing
Contact email: ai4manufacturing _at_ gmail.com