Workshop on Evolutionary Computing and Explainable AI
The workshop will span two sessions (10:50–12:40, 13:40–15:30) on the first day of the GECCO conference, Saturday, 9 July 2022.
Slot 1 (Intro, invited talk & XAI4EC papers)
10:50. Paper 1 and introduction from organizers: Jaume Bacardit, Alexander E.I. Brownlee, Giovanni Iacca, John McCall, Stefano Cagnoni, David Walker. “The intersection of Evolutionary Computation and Explainable AI” [15min + 5min Q&A] [onsite]
11:10. Invited talk: Will N. Browne. “Inherent explainability in AI via EC”. [30m+10min Q&A] [online]
11.50. Paper 2: Manjinder Singh, Alexander Brownlee, David Cairns. “Towards Explainable Metaheuristic: Mining Surrogate Fitness Models for Importance of Variables” [15min + 5min Q&A] [online]
12.12. Paper 3: Mathew Walter, David Walker, Matthew Craven. “An Explainable Visualisation of the Evolutionary Search Process” [15min + 5min Q&A] [onsite]
12.30. Open Discussion [10 min] [onsite/online]
Slot 2 (EC4XAI papers & discussion)
13.40. Paper 4: Martina Saletta, Claudio Ferretti. “Towards the Evolutionary Assessment of Neural Transformers Trained on Source Code” [15min + 5min Q&A] [onsite]
14.00. Paper 5: Leonardo Lucio Custode, Giovanni Iacca. “Interpretable AI for policy-making in pandemics” [15min + 5min Q&A] [onsite]
14.20. Paper 6: Hormoz Shahrzad, Babak Hodjat, Risto Miikkulainen. “Evolving Explainable Rule Sets” [15min + 5min Q&A]
14.40. Paper 7: Hayden Andersen, Andrew Lensen, Will N. Browne. “Improving the Search of Learning Classifier Systems Through Interpretable Feature Clustering” [15min + 5min Q&A] [online]
15.00. Open Discussion [30 min] [onsite/online]
Call for papers
Explainable artificial intelligence has gained significant traction in the machine learning community in recent years because of the need to generate “explanations” of how these typically black-box tools operate that are accessible to a wide range of users. Nature-inspired optimisation techniques are also often black box in nature, and the attention of the explainability community has begun to consider explaining their operation too. Many of the processes that drive nature-inspired optimisers are stochastic and complex, presenting a barrier to understanding how solutions to a given optimisation problem have been generated.
Explainable optimisation can address some of the questions that arise during the use of an optimiser: Is the system biased? Has the problem been formulated correctly? Is the solution trustworthy and fair? By providing mechanisms that enable a decision maker to interrogate an optimiser and answer these questions trust is built with the system. On the other hand, many approaches to XAI in machine learning are based on search algorithms that interrogate or refine the model to be explained, and have the potential to draw on the expertise of the EC community. Furthermore, many of the broader questions (such as what kinds of explanation are most appealing or useful to end users) are faced by XAI researchers in general.
From an application perspective, important questions have arisen, for which XAI may be crucial: Is the system biased? Has the problem been formulated correctly? Is the solution trustworthy and fair? The goal of XAI and related research is to develop methods to interrogate AI processes with the aim of answering these questions. This can support decision makers while also building trust in AI decision-support through more readily understandable explanations.
We seek contributions on a range of topics related to this theme, including but not limited to:
- Interpretability vs explainability in EC and their quantification
- Landscape analysis and XAI
- Contributions of EC to XAI in general
- Use of EC to generate explainable/interpretable models
- XAI in real-world applications of EC
- Possible interplay between XAI and EC theory
- Applications of existing XAI methods to EC
- Novel XAI methods for EC
- Legal and ethical considerations
- Case studies / applications of EC & XAI technologies
Papers will be double blind reviewed by members of our technical programme committee.
Authors can submit short contributions including position papers of up to 4 pages and regular contributions of up to 8 pages following in each category the GECCO paper formatting guidelines. Software demonstrations will also be welcome.
- Submission opening:
February 11, 2022
- Submission deadline:
April 11, 2022
April 25, 2022
May 2, 2022
- Presenter mandatory registration:
May 2, 2022
- Workshop: TBC
Workshop papers must be submitted using the GECCO submission system (https://ssl.linklings.net/conferences/gecco/). After login, the authors need to select the “Workshop Paper” submission form. In the form, the authors must select the workshop they are submitting to. To see a sample of the “Workshop Paper” submission form, go to GECCO’s submission system and select “Sample Submission Forms”. Submitted papers must not exceed 8 pages (excluding references) and are required to be in compliance with the GECCO 2022 Papers Submission Instructions (https://gecco-2022.sigevo.org/Paper-Submission-Instructions). It is recommended to use the same templates as the papers submitted to the main tracks. Each paper submitted to this workshop will be rigorously reviewed in a double-blind review process. In other words, authors should not know who the reviewers of their work are and reviewers should not know who the authors are. To this end, the following information is very important: Submitted papers should be ANONYMIZED. This means that they should NOT contain any element that may reveal the identity of their authors. This includes author names, affiliations, and acknowledgments. Moreover, any references to any of the author’s own work should be made as if the work belonged to someone else. All accepted papers will be presented at the ECXAI workshop and appear in the GECCO 2022 Conference Companion Proceedings. By submitting a paper, the author(s) agree that, if their paper is accepted, they will:
- Submit a final, revised, camera-ready version to the publisher on or before the camera-ready deadline.
- Register at least one author by the author registration deadline to participate in the conference.
- Provide a pre-recorded version of the talk and be present during its online transmission (which will occur during the days of the conference) to answer questions from the (online) audience.
As a published ACM author, you and your co-authors are subject to all ACM Publications Policies (https://www.acm.org/publications/policies/toc), including ACM’s new Publications Policy on Research Involving Human Participants and Subjects (https://www.acm.org/publications/policies/research-involving-human-participants-and-subjects).
Organisers (in alphabetical order)
- Jaume Bacardit, Newcastle University, UK, firstname.lastname@example.org
- Alexander Brownlee, University of Stirling, UK, email@example.com
- Stefano Cagnoni, University of Parma, Italy, firstname.lastname@example.org
- Giovanni Iacca, University of Trento, Italy, email@example.com
- John McCall, Robert Gordon University, UK, firstname.lastname@example.org
- David Walker, University of Plymouth, UK, email@example.com
Technical Programme Committee
- Mauro Castelli
- Matthew Craven
- Alberto Franzin
- Julie Jacques
- Ed Keedwell
- Benjamin Lacroix
- Eric Medvet
- Fabrício Olivetti de França
- Roman Šenkeřík
- Ryan Urbanowicz
- Marco Virgolin
- Sean Walton
- Ciprian Zavoianu