Newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation

Sustainability of GECCO 2023

By Leonor Reigada, João Vasconcelos, Marco Rodrigues, João Pereira, Sara Silva and Carla Silva

Prior to the COVID-19 pandemic’s constraints on travel, academics frequently traveled over 150,000 kilometers per person per year for conferences, board meetings, collaborations, fieldwork, seminars, and lectures [1]. It is therefore unsurprising that academic air travel is one of the primary sources of greenhouse gas (GHG) emissions for universities. The COVID-19 pandemic has forced a significant shift in how we gather and connect, with massive annual conferences and small society meetings moving online. This transition has presented challenges and opportunities, but it has also highlighted the potential of virtual gatherings to bridge geographical and social divides.

GECCO and COVID-19

As with most events amid the COVID-19 pandemic, a change was observed in GECCO from face-
to-face mode to virtual (online) and, after, to hybrid mode (a combination of face-to-face and
online modes), as shown in Figure 1.

In GECCO’s history, since COVID-19, there has been a strong concern about sustainability, and
measures were enforced: avoiding plastic bottles, distributing free local collective transport
passes, reusing the participant identification plastic and lanyard, and encouraging physical
participation of nearby countries and virtual participation of faraway countries. In 2023, a task
force to compute GECCO’s CO₂eq metric was created, and it was decided to use the internationally
recognized GHG protocol [2] and publicly available spreadsheet of the UK [3], based on ISO
14040/44/67 [4–6]. Those methodologies were adapted to events and other countries, to start
creating a solid methodology to compute the GHG emissions and track GECCO’s sustainability
evolution throughout time.

Online mode advantages and disadvantages

Figure 2 wraps up the Strengths, Weaknesses, Opportunities, and Threats (SWOT) of online conferences as perceived at the time of COVID-19 and registered in several documents [1, 7–9].

Not every event organizer committee learned from the pandemic. They returned to normality, even with the positive aspects identified of online mode, and pursued the old, traditional, face-to-face gathering as opposed to the innovative format of hybridization. Despite the empirical knowledge that aviation plays an important role in aggravating global warming emissions, in face-to-face events, well-informed individuals on climate change are often frequent flyers which is highly contradictory. They often continue to combine attendance at conferences with tourism in a foreign country, aggravating the environmental impacts [10].

Measuring sustainability

Measuring sustainability is not a straightforward process since it has social, economic, and environmental aspects embedded. We can also see how intricate sustainability aspects are in the 17 United Nations Sustainable Development Goals (UNSDG) that are related to each other and have more than 150 targets to monitor and observe the evolution of! [11]

There are several efforts to bring a sustainability metric forward in conference events, mostly related to GHG emissions with global warming impacts (expressed in CO₂ equivalents, i.e., having all gases with global warming potential combined and converted into CO₂), under the UNSDG 13 – Climate Action.

The problem is that even with only one metric, CO₂eq, there are differences in approaches that make comparison of absolute numbers of different events deceiving. Nevertheless, the typical range of the numbers and the identification of the “HOT-SPOT”, causing most of the impact, could be assessed. If we grab two handfuls of scientific studies, related to different international conferences, we can observe the face-to-face, online, and hybrid modes typical number ranges, and “HOT-SPOT” identification, Fig. 3, with hybrid-mode GECCO 2023 represented.

Hybrid events’ carbon footprint is highly dependent on the percentage of online participants and their country of origin (air travel distance). Air travel is clearly the “HOT-SPOT” for face-to-face and hybrid events.

On average, the carbon footprint of the face-to-face mode is 400 times higher than the online mode. Which is a considerable difference.

GECCO 2023 had 1094 kg CO₂eq/in-person participant and 19 kg CO₂eq/online participant, i.e., an overall impact of 890 kg CO₂eq/participant, 86% due to transport (mostly international flights with distances higher than 5,000 km). Even so, going hybrid, GECCO 2023 avoided the emissions of 200 tons of CO₂eq to the atmosphere. To “offset” the hybrid mode GECCO 2023 emissions would require the activity of 24,350 trees for a year (an area equivalent to 70 football fields).

Figure 3. (a) Typical ranges of values for GHG emissions: face-to-face; hybrid and online events. (b) Contributions to GHG emissions [12–18]. GECCO 2023 values are calculated by the authors.

Takeaway lesson

Unraveling more innovation in international conferences is possible. In the future, the hybridization of conferences [18] could take a new mode, featuring a network of physical hubs distributed throughout the globe. These hubs would serve as local gathering points for participants, reducing the need for air travel, fostering in-person interactions as well as maintaining the online positive aspects (Fig.2). Such a “Hybrid-with-Hubs” approach would expand the conference’s audience even further, but surely would require a significant organizational effort.

The takeaway lesson here: as in all hurdles in our life there’s a positive side to it! A pandemic paved the way for more innovative, environmentally friendly, and inclusive conferences!!!

References

1. Jack, T.; Glover, A. Online conferencing in the midst of COVID-19: an “already existing experiment” in academic internationalization without air travel. Sustain. Sci. Pract. Policy 2021, 17, 292–304.
2. Price, M. Science. April 2020,.
3. Bonifati, A.; Guerrini, G.; Lutz, C.; Martens, W.; Mazilu, L.; Paton, N.W.; Vaz Salles, M.A.; Scholl, M.H.; Zhou, Y. Holding a Conference Online and Live due to Covid-19. ACM SIGMOD Rec. 2021, 49, 28–32.
4. McEvoy, N.L.; Trapani, J.; Tume, L.N. The changing face of scientific conferences: Face to face, online, or a hybrid model? Nurs. Crit. Care 2022, 27, 7–9.
5. Gössling, S. Global environmental consequences of tourism. Glob. Environ. Chang. 2002, 12, 283–302.
6. UN TRANSFORMING OUR WORLD: THE 2030 AGENDA FOR SUSTAINABLE DEVELOPMENT. A/RES/70/1.; 2015;
7. WBCSD and WRI A Corporate Accounting and Reporting Standard. Revised edition; 2013; ISBN 1-56973-568-9.
8. UKDEFRA 2021 Government Greenhouse Gas Conversion Factors for Company Reporting. Methodology Paper for Conversion factors Final Report.; 2021;
9. ISO BS EN ISO 14047: Environmental management – Life cycle assessment – Illustrative examples on how to apply ISO 14044 to impact assessment situations. Int. Organ. Stand. 2012.
10. ISO. Environmental management – Life cycle assessment – Principles and framework. Geneva, Switzerland: International Organization for Standardization. 2006.
11. ISO 14067:2018 Greenhouse gases — Carbon footprint of products — Requirements and guidelines for quantification.
12. Cavallin Toscani, A.; Atasu, A.; Van Wassenhove, L.N.; Vinelli, A. Life cycle assessment of in‐person, virtual, and hybrid academic conferences: New evidence and perspectives. J. Ind. Ecol. 2023, 27, 1461–1475.
13. Astudillo, M.F.; AzariJafari, H. Estimating the global warming emissions of the LCAXVII conference: connecting flights matter. Int. J. Life Cycle Assess. 2018, 23, 1512–1516.
14. Periyasamy, A.G.; Singh, A.; Ravindra, K. Carbon Emissions from Virtual and Physical Modes of Conference and Prospects for Carbon Neutrality: An Analysis From India. Air, Soil Water Res. 2022, 15, 117862212210932.
15. Wortzel, J.R.; Stashevsky, A.; Wortzel, J.D.; Mark, B.; Lewis, J.; Haase, E. Estimation of the Carbon Footprint Associated With Attendees of the American Psychiatric Association Annual Meeting. JAMA Netw. Open 2021, 4, e2035641.
16. Stroud, J.T.; Feeley, K.J. Responsible academia: optimizing conference locations to minimize greenhouse gas emissions. Ecography (Cop.). 2015, 38, 402–404.
17. Burtscher, L.; Barret, D.; Borkar, A.P.; Grinberg, V.; Jahnke, K.; Kendrew, S.; Maffey, G.; McCaughrean, M.J. The carbon footprint of large astronomy meetings. Nat. Astron. 2020, 4, 823–825.
18. Tao, Y.; Steckel, D.; Klemeš, J.J.; You, F. Trend towards virtual and hybrid conferences may be an effective climate change mitigation strategy. Nat. Commun. 2021, 12, 7324.

About the Authors

Leonor Reigada (left), João Vasconcelos, Marco Rodrigues (right), Bachelor’s students, FCUL, Portugal

We are Energy and Environment Eng. Bachelor students in the Faculty of Sciences, University of Lisbon. Our Bachelor’s research project supervised by Professor Carla Silva was to create a first approach to a harmonized carbon footprint calculator for conferences.

João Pereira, MSc student, FCUL, Portugal

I am an Energy and Environment Eng. MSc student in the Faculty of Sciences, University of Lisbon, under the supervision of Professor Carla Silva. My thesis evaluates the carbon footprint of several face-to-face and hybrid events, assesses the representativeness of conference surveys, and proposes a simplified approach to minimize the environmental impact of future events through a face-to-face HUB network for hybrid events.

Sara Silva, GECCO 2023 data provider and reviewer

Sara Silva is an Associate Professor at the Faculty of Sciences of the University of Lisbon and a researcher at LASIGE, Portugal. Her main research area is evolutionary machine learning applied in remote sensing and in bioinformatics. In 2018 she received the EvoStar Award for Outstanding Contribution to Evolutionary Computation in Europe by the Society for the Promotion of Evolutionary Computation in Europe and its Surroundings (SPECIES). In 2023 she was General Chair of the Genetic and Evolutionary Computation Conference (GECCO), held in Lisbon, Portugal, in hybrid mode. She is member of the executive boards of both SPECIES and ACM SIGEVO. She created the MATLAB Genetic Programming Toolbox (GPLAB) and co-authored the recently published book “Lectures on Intelligent Systems” by Springer.

Carla Silva, Sustainability team supervisor and Coordinator of the Master program Energy and Environment Eng. FCUL, Portugal

Dr. Carla Silva holds a PhD degree in Mechanical Engineering in 2005 from University of Lisbon (UL), Portugal. She was a senior researcher at the Institute of Mechanical Engineering 2008 to 2015. Since 2016 she is Professor at the Department of Geographic, Geophysics and Energy Engineering of the Faculty of Sciences of University of Lisbon. She teaches, in the 1^st and 2^nd cycles of Energy and Environment Eng. and in the 3^rd cycle in the Doctoral program Sustainable Energy Systems. She has successfully supervised 60 MSc and 10 PhD students. She is a researcher at Instituto Dom Luiz (IDL) since 2016 with several publications. Areas of interest are: complex system energy and emission analysis; biorefinery; circular economy; carbon footprint; life cycle assessment; indicators for sustainable mobility.

---

Searching the Genetic Programming Bibliography

By W. B. Langdon

Abstract

Since February 2024 the genetic programming bibliography has supported free text queries to find GP papers via http://gpbib.cs.ucl.ac.uk/gp-search/

Background

The Genetic Programming bibliography aims to cover all papers, books, PhD theses, etc., on genetic programming. It was started by John Koza and first published in 1994 as Appendix F of his second GP book [1]. I took on the bibliography, and was invited by Pete Angeline to provide it as an appendix to the second “Advances in Genetic Programming” book, which he co-edited with Kim Kinnear (published in 1996 by MIT Press [2]). Since then the bibliography has been available online, firstly by FTP and now via HTTP. Whilst Steve Gustafson was based in the UK, he assisted in the maintenance and growth of the bibliography, including designing its home page. For almost all of the first twenty years it was hosted by The Birmingham University School of Computer Science. Then in the summer of 2019 it moved to its new host http://gpbib.cs.ucl.ac.uk, back in the Computer Science department of University College, London. Since the autumn of 2019, with the support of Jason Moore, the Perelman School of Medicine in the University of Pennsylvania has hosted a mirror of the bibliography and supporting web pages: http://gpbib.pmacs.upenn.edu. The ACM’s Special Interest Group on Evolutionary Computation (SIGEVO)’s newsletter SIGEVOlution [4, 5] and the Genetic Programming and Evolvable Machines and other journals have published articles on the GP bibliography [3, 7, 8, 6] or analysis of data derived from it [9, 11, 10].

Online Data, Web Searches and Unavailable Traditional Queries

The whole bibliography is held in a single bibtex file gp-bibliography.bib. At the time of writing, it contains 17, 921 entries (16,696 directly related to genetic programming), 38 MBytes. Originally the intention was gp-bibliography.bib would just be used as a bibtex file with references being extracted when people wrote papers in LATEX. However uses of the bibliography have expanded. Although many bibliography tools support bibtex, copies in refer format gp-bibliography.ref and as plain text gp-bib-alpha.txt are also available.

Perhaps the most widely used interface is that each GP paper has its own HTML web page (e.g. http://gpbib.cs.ucl.ac.uk/gp-html/koza_2003_gpt.html). Almost all have hypertext links to the paper itself, its publisher and to its Google Scholar citations. Since these pages are freely available, they make attractive targets for web browsing and web searches and so the GP bibliography is easily searchable via Google and other web search engines.

For thirty years (1993–2023) The Collection of Computer Science Bibliographies (CCSB) hosted hundreds of bibliographies, including the genetic programming bibliography. It provided a number of search interfaces, including allowing online searches using Lucerne syntax. Since the GP bibliography was part of CCSB, it could be searched in a sophisticated way using the CCSB’s web interfaces. However CCSB was retired last summer and when it was switch off, its search interface for the GP bibliography was also removed.

The New Free Text Search Interface

As mentioned in the previous section, the CCSB sophisticated search interface was removed in July 2023. The goal of its replacement http://gpbib.cs.ucl.ac.uk/gp-search/ is to provide something lightweight, which is interactive, fast, immediate and without a learning curve, as shown in Figure 1.

Following the success of the co-author search interface http://gpbib.cs.ucl.ac.uk/gp-coauthors, the new free text search interface is also based in the user’s device, using JavaScript running in a web browser.

Single Web Page

The interface consists of a single HTML web page containing a web form with various input and output forms and buttons, see Figure 1.

Find Keywords

• The first input field allows the user to enter keywords to be searched for.
• Stop words are ignored, e.g. words of one letter, “in”, “the”, “from” and “model”.
• Also since every paper will automatically match “genetic” “algorithms” and programming” (after all this is a bibliography dedicated to Genetic Programming) these and “gp” are ignored.
• To avoid overly long time consuming searches, queries that match more than 1000 papers are cut short.

Limiting the number of matches

By default, up to ten hits are displayed. The “≤hits” input field allows the number of papers displayed to be changed (up to a maximum of 100).

Types of output

There are 3 types of output: brief, terse, and full, as radio buttons.

• brief. The brief display requires no further interaction from the host (or mirror) and simply displays the bibtex key of the matching GP paper and gives a hyperlink to its web page within the GP bibliography.
• terse. The terse output format shows fragments of the matching paper’s entry with the matching parts highlighted. There is also a hyperlink to the corresponding full web page. This is the default.
• full. Shows the complete bibtex of the matching GP paper. This often includes links to the paper itself, the publisher of the article and to its Google Scholar citations.

Matching by year

• The lower two input areas allow the user to specify the years they are interested in. Years are four digit numbers.
• The up buttons ^ increase the year by one. The down buttons V decrease it by one. To allow easy sweeps of papers published in a range of years, the central ^/V buttons increment or decrement both before and after years by the same amount.

Search summary statistics

The output display reflects the query (without stop words) in the order the query’s words are used, the number of matching papers displayed and the total number of matches. If (e.g. for speed) the search is truncated, so that the number of matches is not known, a + is appended. (The total number of searches is given in square brackets [ ] but it is often only useful for debugging.)

Ordering Results

The interface attempts to display the best matches at the top of the list of entries. This ordering is based on a heuristic which weights hits with matches that occur together more highly, matches on titles above those on keywords, above those on abstracts. Matches in long abstracts carry less weight but matches against exact words or matches which match case exactly are up weighted.

Feedback

It would be great to hear of people’s experiences with the new interface, bug reports, errors and future ideas for the search tool or for the GP bibliography in general.

I asked a number of authors of recent computer science surveys in various fields unrelated to Genetic Programming how they used other online search interfaces, such as those provided by DBLP, IEEE Xplore, the ACM Digital Library. The most useful feature seemed to be the ability to search by keywords. This is what the new interface provides. Its underlying data structures are public and so interested parties could build additional features on top of them. Mainstream digital libraries and online bibliographies include other tools, which the authors of surveys found useful but which are not included in the GP bibliography or only indirectly supported by it.

Perhaps the SIGEvolution newsletter could be a venue to enable everyone to hear of people’s experiences with the new search tool and to know of their thoughts on it or possible extensions.

Adding Your GP Papers

The Genetic Programming bibliography has long had a web interface allowing authors to provide missing GP papers, however in practice the easiest way is simply to email a bibtex description of missing GP works to me.

Acknowledgements

I would like to thank Una-May O’Reilly who first highlighted the need for a new search tool, Aymeric Blot who highlighted the problem of scaling existing web browser search interfaces from 200 papers to 20,000, Carmen Meinson who suggested stylish improvements and Farooq Karimi Zadeh, Justyna Petke, Carol Hanna, and UCL students, for testing it.

References

[1]  J. R. Koza. Genetic Programming II: Automatic Discovery of Reusable Programs. MIT Press, Cambridge Massachusetts, May 1994.

[2]  W. B. Langdon. A bibliography for genetic programming. In P. J. Angeline and K. E. Kinnear, Jr., editors, Advances in Genetic Programming 2, appendix B, pages 507–531. MIT Press, Cambridge, MA, USA, 1996.

[3]  W. B. Langdon. Genetic programming and evolvable machines: Books and other resources. Genetic Programming and Evolvable Machines, 1(1/2):165–169, Apr. 2000.

[4]  W. B. Langdon. Web usage of the GP bibliography. SIGEVOlution newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation, 1(4):16–21, Dec. 2006.

[5]  W. B. Langdon. News of the GP bibliography. SIGEVOlution newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation, 6(3-4):12–16, July 2014.

[6]  W. B. Langdon. Genetic programming and evolvable machines at 20. Genetic Programming and Evolvable Machines, 21(1-2):205–217, June 2020. Twentieth Anniversary Issue.

[7]  W. B. Langdon and S. Gustafson. Genetic programming and evolvable machines: Five years of reviews. Genetic Programming and Evolvable Machines, 6(2):221–228, June 2005.

[8] W. B. Langdon and S. M. Gustafson. Genetic programming and evolvable machines: ten years of reviews. Genetic Programming and Evolvable Machines, 11(3/4):321–338, Sept. 2010. Tenth Anniversary Issue: Progress in Genetic Programming and Evolvable Machines.

[9] W. B. Langdon, R. Poli, and W. Banzhaf. An eigen analysis of the GP community. Genetic Programming and Evolvable Machines, 9(3):171–182, Sept. 2008.

[10] M. Tomassini and L. Luthi. Empirical analysis of the evolution of a scientific collaboration network. Physica A, 385:750–764, 2007. Available online 25 July 2007.

[11] M. Tomassini, L. Luthi, M. Giacobini, and W. B. Langdon. The structure of the genetic programming collaboration network. Genetic Programming and Evolvable Machines, 8(1):97–103, Mar. 2007.

Remembering Jeffrey Horn (1963-2024)

by Kalyanmoy Deb and David E. Goldberg and W. B. Langdon

Kalyanmoy Deb and David E. Goldberg

He is remembered for his research in genetic algorithms and evolutionary computation, particularly his work in evolutionary multi-objective optimization (EMO), understanding problem difficulty, and the evolution of cooperative co-solutions. His 1994 study on niched Pareto genetic algorithm (NPGA) received more than 4,000 Google Scholar citations and is one of the three earliest EMO algorithms which propelled the EMO field forward. His 2001 application of NPGA to ground water remediation system design was one of the early practical applications of EMO algorithms. Besides his contributions in EMO and its applications, he was passionate in understanding ways to form stable niches within evolving populations to find multiple optimal solutions. He constructed deceptive and long path problems to have a deeper understanding of the working of genetic algorithms. He also contributed to evolutionary robotics, artificial life, and applied evolutionary computation to shape nesting problems and in games.

He is survived by his wife, Gabriele Burkhard, and his children Thorin and Ellissa.

W. B. Langdon

I had met Jeff several times at evolutionary computation events, eg. PPSN. The first time we met was at the first genetic programming conference organised by John Koza in Stanford University in 1996, whilst I was still a PhD student. Somehow I had wangled an invitation to the Foundations of Genetic Algorithms
workshop (the oh so poetic FOGA 4) This was being organised by Rik Belew in San Diego, the other end of California but I had not a plan to get there. In stepped Jeff, he had been at GP’96 and had a real paper to present at FOGA and offered me a lift!

Next morning onto San Diego getting the rental back more-or-less in time. Jeff was always softly spoken but I remember him being very enthusiastic about the upper peninsula where he taught after his PhD with David Goldberg.

Announcements

ACM TELO Special Issue on Explainable AI in Evolutionary Computation

TELO EICs are pleased to announce the following 2023 Featured Articles:

Combining Evolution and Deep Reinforcement Learning for Policy Search: A Survey.
This is the most systematic survey of recent works (since 2017) that combine Deep neuroevolution and Deep Reinforcement Learning. Anyone interest in those topics will find this paper to be essential reading.

Empirical analysis of PGA-MAP-Elites for Neuroevolution in Uncertain Domains.
Quality-Diversity (QD) algorithms are ideal for tackling optimization tasks that aim to produce diverse and innovative solutions, such as engineering design. PGA-MAP-Elites is a QD algorithm that overcomes the limitations of other QD algorithms by being applicable to high-dimensional spaces and uncertain environments. This paper provides a thorough empirical analysis of this powerful algorithm.

Covariance Matrix Adaptation Evolutionary Strategy with Worst-Case Ranking Approximation for Min–Max Optimization and Its Application to Berthing Control Tasks.
This paper proposes a variant of the well-known CMA-ES optimizer designed for Min-Max Black-box Optimization problems, in particular when the goal is to find the more robust solution (least risky). As an illustrative real-world application, the approach is applied to design a controller that realizes a fine control of a ship toward a target state with the least collision risk and minimum elapsed time. A candidate controller is evaluated using simulation under uncertainty. The proposed algorithm can be applied to other min-max black-box optimization problems under uncertainty.

Multi-Objective Hyperparameter Optimization in Machine Learning—An Overview.
Hyperparameter optimization is nowadays a standard step in machine learning (ML) pipelines and the basis of AutoML approaches. It is often the case that there are more than one performance metrics giving rise to a multi-objective problem. However, such approaches are often neglected in introductory texts on hyperparameter optimization. This paper provides an extensive survey of this topic that covers the domains of evolutionary algorithms and Bayesian optimization. It also illustrates the utility of multi-objective optimization in several specific ML applications, considering objectives such as operating conditions, prediction time, sparseness, fairness, interpretability, and robustness.

Forthcoming Events

---

Call for Submissions

18th International Conference on Parallel Problem Solving From Nature (PPSN 2024)

PPSN was originally designed to bring together researchers and practitioners in the field of Natural Computing, the study of computing approaches which are gleaned from natural models. Today, the conference series has evolved and welcomes works on all types of iterative optimization heuristics. Notably, we also welcome submissions on connections between search heuristics and machine learning or other artificial intelligence approaches. Submissions covering the entire spectrum of work, ranging from rigorously derived mathematical results to carefully crafted empirical studies, are invited.

Important dates

• Paper submission due: April 5, 2024
• Notification of acceptance: May 31, 2024

Graph-based Genetic Programming Workshop

Call for Papers

To be held as part of the Genetic and Evolutionary Computation Conference (GECCO)
July 14-18, 2024 (Sunday – Thursday), Melbourne, Australia (Hybrid)

We invite submissions that present recent developments in graph-based Genetic Programming. The content of the submission must be either new original work or discuss new perspectives on recently published work. Submitting work that is in an early stage or in progress is welcomed and appreciated.

Important Dates

• Submission deadline: April 8, 2024
• Notification: May 3, 2024
• Camera-ready: May 10, 2024
• Author’s mandatory registration: TBD

21th Annual (2024) “Humies” Awards

Entries are hereby solicited for awards totaling $10,000 for human-competitive results that have been produced by any form of genetic and evolutionary computation (including, but not limited to genetic algorithms, genetic programming, evolution strategies, evolutionary programming, learning classifier systems, grammatical evolution, gene expression programming, differential evolution, genetic improvement, etc.) and that have been published in the open, reviewed literature between the deadline for the previous competition and the deadline for the current competition.

Important Dates

• Friday, May 31, 2024: Deadline for entries (consisting of one TEXT file, PDF files for one or more papers, and possible “in press” documentation. Please send entries to goodman at msu dot edu
• Friday, June 14, 2024: Finalists will be notified by e-mail
• Friday, June 28, 2024: Finalists who will not be in Melbourne to present in person must
  submit a 10-minute video presentation to goodman at msu dot edu. Finalists who will present in person must submit a copy of their slides, for the advance use of the judges, to goodman at msu dot edu.
• July 14-18, 2024 (Sunday – Thursday): GECCO conference (the schedule for the Humies session is not yet final, so please check the GECCO program as it is updated for the time of the Humies session).
• Thursday, July 18, 2024: Announcement of awards at the plenary session of the GECCO conference.

16th International Conference on Evolutionary Computation Theory and Applications (ECTA 2024)

Considered a subfield of computational intelligence focused on combinatorial optimisation problems, evolutionary computation is associated with systems that use computational models of evolutionary processes as the key elements in design and implementation, i.e. computational techniques which are based to some degree on the evolution of biological life in the natural world. A number of evolutionary computational models have been proposed, including evolutionary algorithms, genetic algorithms, the evolution strategy, evolutionary programming and swarm intelligence. These techniques form the basis of several disciplines such as artificial life and evolutionary robotics.
This conference intends to be a major forum for scientists, engineers and practitioners interested in the study, analysis, design, modeling and implementation of evolvable systems, both theoretically and in a broad range of application fields.

Important Dates

• Regular Paper Submission: June 3, 2024
• Position Paper Submission: July 17, 2024

Interpretable Control Competition

Call for Submissions

To be held as part of the Genetic and Evolutionary Computation Conference (GECCO)
July 14-18, 2024 (Sunday – Thursday), Melbourne, Australia (Hybrid)

The Interpretable Control Competition is a part of GECCO 2024, challenging you to create control systems that are both effective and understandable. Opaque AI, while powerful, can be risky, especially for safety-critical applications. This competition seeks to find a balance between performance and interpretability. The competition focuses on two key goals:

1. Develop interpretable control systems for continuous and discrete tasks (Walker2D and 2048)
2. Establish methods for evaluating interpretability.

You can use any technique you like, but we encourage the use of Evolutionary Computation (EC) for either policy generation or explanation.

Join us to push the boundaries of interpretable AI for control systems!
Submission Deadline: June 13th, 2024.

See our website or Discord
(https://discord.gg/dA8jpFVa9t) for details.

About this Newsletter