pith. sign in

arxiv: 2604.12482 · v2 · submitted 2026-04-14 · 💻 cs.RO · cs.AI

Social Learning Strategies for Evolved Virtual Soft Robots

K. Ege de Bruin , Kyrre Glette , Kai Olav Ellefsen , Giorgia Nadizar , Eric Medvet This is my paper

Pith reviewed 2026-05-10 15:07 UTC · model grok-4.3

classification 💻 cs.RO cs.AI
keywords social learningevolutionary roboticssoft robotscontroller optimizationteacher selectionmorphology transferbody-brain co-optimizationvirtual robots
0
0 comments X

The pith

Social learning from peers lets evolved soft robots optimize their controllers faster than learning independently.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper tests whether robots can speed up their own controller learning by drawing on parameters already optimized by other robots in the same evolving population. In joint body-and-brain optimization of virtual soft robots, each robot normally tunes its own control parameters through a separate learning process. The authors instead let robots select teachers among their peers, inherit those parameters, and continue adapting them. Across four tasks and environments, this social approach reaches higher performance than the baseline of learning from scratch when both methods are given identical computational budgets. Results also indicate that drawing from several teachers tends to produce steadier gains, while the best single-teacher rule remains unsettled.

Core claim

We introduce a social learning approach in which robots can exploit optimized parameters from their peers to accelerate their own brain optimization. Our results confirm the effectiveness of building on others' experience, as social learning clearly outperforms learning from scratch under equivalent computational budgets. In addition, while the optimal teacher selection strategy remains open, our findings suggest that incorporating knowledge from multiple teachers can yield more consistent and robust improvements.

What carries the argument

The social learning framework that lets each robot inherit and adapt control parameters from selected peer robots, with teacher choice based on morphology similarity or other criteria.

If this is right

  • Social learning reduces the evaluations needed to reach effective controllers during coupled body-brain evolution.
  • Learning from multiple teachers produces more consistent performance than learning from any single teacher.
  • Transfer works best when teachers share morphological traits with the learner because of the tight body-brain coupling.
  • The performance advantage holds across four distinct tasks and environments.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The method could lower the total cost of discovering good robot designs by letting the population share the expense of controller search.
  • Similar reuse of learned parameters might apply to other coupled design problems where agents have related structures.
  • Hardware tests would show whether the simulation-based transfer survives real-world variations in morphology and sensing.
  • Making teacher selection dynamic during evolution rather than fixed in advance could increase the gains further.

Load-bearing premise

Control parameters optimized by one robot contain transferable value for other robots, especially those with similar morphologies, without causing negative transfer.

What would settle it

An experiment in which robots using social learning reach no higher task performance than independent learners when both are restricted to the same total number of controller evaluations.

Figures

Figures reproduced from arXiv: 2604.12482 by Eric Medvet, Giorgia Nadizar, Kai Olav Ellefsen, K. Ege de Bruin, Kyrre Glette.

Figure 1
Figure 1. Figure 1: Overview of the social learning (SL) approach. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: A snapshot of four episodes where a VSR performs [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 4
Figure 4. Figure 4: The evolution of the population diversity, averaged [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 3
Figure 3. Figure 3: Distribution of the quality𝑞 ★ of the best-performing VSR in the four tasks for the eight approaches. Stars above the boxes show significant differences: for example, a red star on top of a gray box indicates that IL is significantly different from Best-One. clear motivations for this missed observation. We hypothesize that the diversity of evolved bodies is not large enough to make Similar-* more convenie… view at source ↗
Figure 8
Figure 8. Figure 8: Distribution of the quality𝑞 ★ of the best-performing VSR subjected to the re-learning of the brain with a large budget for BO. namely the Best-Many variant, when co-optimizing the body and the brain of a VSR: robots obtained this way are in general better and their bodies are not less diverse, nor “potentially” worse. We also measured the performance of VSRs subjected to brain re-learning, but using anoth… view at source ↗
Figure 5
Figure 5. Figure 5: Body features for all the optimized VSRs. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 7
Figure 7. Figure 7: A selection of the optimized VSR bodies, one per [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 9
Figure 9. Figure 9: Distribution of the quality𝑞 ★ of the best-performing VSR optimized on a source task (row of plots) subjected to the re-learning of the brain with a large budget for BO and for another destination task (groups of boxes). 0 100 200 300 400 500 0 50 100 Learning iteration Quality 𝑞 ★ No-BO IL SL [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Performance over learning iterations for three [PITH_FULL_IMAGE:figures/full_fig_p008_10.png] view at source ↗
read the original abstract

Optimizing the body and brain of a robot is a coupled challenge: the morphology determines what control strategies are effective, while the control parameters influence how well the morphology performs. This joint optimization can be done through nested loops of evolutionary and learning processes, where the control parameters of each robot are learned independently. However, the control parameters learned by one robot may contain valuable information for others. Thus, we introduce a social learning approach in which robots can exploit optimized parameters from their peers to accelerate their own brain optimization. Within this framework, we systematically investigate how the selection of teachers, deciding which and how many robots to learn from, affects performance, experimenting with virtual soft robots in four tasks and environments. In particular, we study the effect of inheriting experience from morphologically similar robots due to the tightly coupled body and brain in robot optimization. Our results confirm the effectiveness of building on others' experience, as social learning clearly outperforms learning from scratch under equivalent computational budgets. In addition, while the optimal teacher selection strategy remains open, our findings suggest that incorporating knowledge from multiple teachers can yield more consistent and robust improvements.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper introduces a social learning framework for joint body-brain optimization in virtual soft robots, allowing individual agents to inherit and build upon control parameters optimized by peer robots rather than learning from scratch. It systematically examines teacher selection strategies (including morphologically similar teachers and multiple-teacher approaches) across four tasks and environments, reporting that social learning yields clear performance gains over independent optimization under matched computational budgets, with multi-teacher inheritance providing more consistent results.

Significance. If the empirical claims hold after addressing experimental gaps, the work would be significant for evolutionary robotics: it provides concrete evidence that social transfer can reduce the cost of coupled morphology-control search, a long-standing bottleneck. The focus on morphological similarity as a transfer heuristic is a useful contribution, and the finding that multiple teachers improve robustness is actionable for designing collective optimization systems.

major comments (2)
  1. [Results] Results section (and abstract): the central claim that social learning 'clearly outperforms learning from scratch under equivalent computational budgets' is load-bearing but unsupported by reported statistical tests, number of independent runs, effect sizes, or explicit baseline comparisons. Without these, it is impossible to determine whether observed gains exceed variance or whether any negative-transfer episodes inflated the effective budget for social learners.
  2. [Methods / Teacher Selection] Teacher selection and transfer analysis: the assumption that parameters optimized for one morphology transfer net-positively to similar morphologies is central, yet the manuscript provides no quantification of negative transfer (e.g., success rate of transferred initializations, extra evaluations needed to recover from poor starts, or cases where transfer increased total evaluations). If negative transfer occurs even for 'morphologically similar' teachers, the budget-equivalence claim does not hold.
minor comments (2)
  1. [Introduction] The abstract and introduction use 'social learning' without a concise formal definition or pseudocode; a short algorithmic box would clarify the distinction between teacher selection, parameter inheritance, and subsequent local optimization.
  2. [Figures and Tables] Figure captions and tables should explicitly state the number of robots, generations, and evaluation budget per condition to allow direct comparison of computational cost.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback highlighting the need for stronger statistical support and explicit transfer analysis. We address each major comment below and will revise the manuscript to incorporate additional reporting and analysis from our existing experimental data.

read point-by-point responses

  1. Referee: [Results] Results section (and abstract): the central claim that social learning 'clearly outperforms learning from scratch under equivalent computational budgets' is load-bearing but unsupported by reported statistical tests, number of independent runs, effect sizes, or explicit baseline comparisons. Without these, it is impossible to determine whether observed gains exceed variance or whether any negative-transfer episodes inflated the effective budget for social learners.

    Authors: We agree that the presentation of results can be strengthened with explicit statistical details. Our experiments used multiple independent runs per condition across all tasks; we will report the exact number of runs, include standard deviations, effect sizes, and appropriate statistical tests (e.g., Wilcoxon rank-sum or t-tests with p-values) comparing social learning to the individual-learning baseline under matched evaluation budgets. We will also add a brief discussion of how the budget is computed (total controller evaluations) and note that inheritance replaces random initialization without increasing the total count. This will allow readers to assess variance and any potential negative-transfer effects on effective cost. revision: yes

  2. Referee: [Methods / Teacher Selection] Teacher selection and transfer analysis: the assumption that parameters optimized for one morphology transfer net-positively to similar morphologies is central, yet the manuscript provides no quantification of negative transfer (e.g., success rate of transferred initializations, extra evaluations needed to recover from poor starts, or cases where transfer increased total evaluations). If negative transfer occurs even for 'morphologically similar' teachers, the budget-equivalence claim does not hold.

    Authors: We acknowledge that the current manuscript does not provide a dedicated quantification of negative transfer. In revision we will add an analysis (drawing on the existing run data) reporting: (i) the fraction of transfers where inherited parameters yielded higher initial fitness than random initialization, (ii) the distribution of post-transfer starting performance, and (iii) any observed cases requiring extra evaluations to recover. Results will be broken down by single-teacher (morphologically similar) vs. multi-teacher conditions. If negative transfer is present, we will discuss its impact on the budget-equivalence argument and note it as a limitation. revision: yes

Circularity Check

0 steps flagged

Empirical study with no derivation chain or self-referential reductions

full rationale

The paper reports experimental comparisons of social learning versus independent learning-from-scratch in evolved soft robots across four tasks. All performance claims rest on measured outcomes under matched evaluation budgets rather than any first-principles derivation, fitted parameter renamed as prediction, or self-citation that supplies a uniqueness theorem. No equations, ansatzes, or theoretical steps appear; the transferability assumption is tested directly via morphology-similarity conditions and multiple-teacher variants, leaving the results falsifiable by the reported simulations themselves.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides no explicit free parameters, axioms, or invented entities; all details on implementation, fitness functions, and transfer mechanisms are absent.

pith-pipeline@v0.9.0 · 5505 in / 1062 out tokens · 29543 ms · 2026-05-10T15:07:25.883964+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

61 extracted references · 61 canonical work pages

  1. [1]

    Alberto Bartoli, Marco Catto, Andrea De Lorenzo, Eric Medvet, and Jacopo Talamini. 2020. Mechanisms of social learning in evolved artificial life. InArtificial Life Conference Proceedings 32. MIT Press, 190–198

    work page 2020

  2. [2]

    Yoav Benjamini and Yosef Hochberg. 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing.Journal of the Royal statistical society: series B (Methodological)57, 1 (1995), 289–300

    work page 1995

  3. [3]

    Jagdeep Bhatia, Holly Jackson, Yunsheng Tian, Jie Xu, and Wojciech Matusik

    work page

  4. [4]

    Evolution gym: A large-scale benchmark for evolving soft robots.Advances in Neural Information Processing Systems34 (2021), 2201–2214

    work page 2021

  5. [5]

    Nicolas Bredeche and Nicolas Fontbonne. 2022. Social learning in swarm robotics. Philosophical Transactions of the Royal Society B: Biological Sciences377, 1843 (2022)

    work page 2022

  6. [6]

    Konstantinos Chatzilygeroudis, Vassilis Vassiliades, Freek Stulp, Sylvain Calinon, and Jean-Baptiste Mouret. 2019. A survey on policy search algorithms for Social learning for Evolved VSRs learning robot controllers in a handful of trials.IEEE Transactions on Robotics36, 2 (2019), 328–347

    work page 2019

  7. [7]

    Nick Cheney, Josh Bongard, Vytas Sunspiral, and Hod Lipson. 2016. On the Dif- ficulty of Co-Optimizing Morphology and Control in Evolved Virtual Creatures. Proceedings of the Artificial Life Conference 2016 (ALIFE XV)(2016), 226–234

    work page 2016

  8. [8]

    Nick Cheney, Josh Bongard, Vytas SunSpiral, and Hod Lipson. 2018. Scalable co-optimization of morphology and control in embodied machines.Journal of The Royal Society Interface15, 143 (2018), 20170937

    work page 2018

  9. [9]

    Nick Cheney, Robert MacCurdy, Jeff Clune, and Hod Lipson. 2014. Unshackling evolution: evolving soft robots with multiple materials and a powerful generative encoding.ACM SIGEVOlution7, 1 (2014), 11–23

    work page 2014

  10. [10]

    Antoine Cully and Yiannis Demiris. 2017. Quality and diversity optimization: A unifying modular framework.IEEE Transactions on Evolutionary Computation 22, 2 (2017), 245–259

    work page 2017

  11. [11]

    Ege de Bruin. 2026. Replication Data for: Social Learning Strategies for Evolved Virtual Soft Robots. doi:10.18710/7BR3NZ

    work page doi:10.18710/7br3nz 2026

  12. [12]

    K Ege de Bruin, Kyrre Glette, and Kai Olav Ellefsen. 2025. Generational Replace- ment and Learning for High-Performing and Diverse Populations in Evolvable Robots. In2025 IEEE Symposium on Computational Intelligence in Artificial Life and Cooperative Intelligent Systems (ALIFE-CIS). IEEE, 1–7

    work page 2025

  13. [13]

    K Ege de Bruin, Kyrre Glette, and Kai Olav Ellefsen. 2025. Integrating Sample Inheritance into Bayesian Optimization for Evolutionary Robotics. InArtificial Life Conference Proceedings 37, Vol. 2025. MIT Press One Rogers Street, Cambridge, MA 02142-1209, USA journals-info . . . , 38

    work page 2025

  14. [14]

    Aguston E Eiben, Nicolas Bredeche, Mark Hoogendoorn, Jürgen Stradner, Jon Timmis, Andy Tyrrell, and Alan Winfield. 2013. The triangle of life: Evolving robots in real-time and real-space. InEuropean conference on artificial life (ECAL- 2013). 1–8

    work page 2013

  15. [15]

    Kai Olav Ellefsen. 2013. Balancing the Costs and Benefits of Learning Ability.. In ECAL. 292–299

    work page 2013

  16. [16]

    David Eriksson, Michael Pearce, Jacob Gardner, Ryan D Turner, and Matthias Poloczek. 2019. Scalable global optimization via local Bayesian optimization. Advances in neural information processing systems32 (2019)

    work page 2019

  17. [17]

    Andrés Faíña, Francisco Bellas, Fernando López-Peña, and Richard J Duro. 2013. EDHMoR: Evolutionary designer of heterogeneous modular robots.Engineering Applications of Artificial Intelligence26, 10 (2013), 2408–2423

    work page 2013

  18. [18]

    Marcus W Feldman, Kenichi Aoki, and Jochen Kumm. 1996. Individual versus so- cial learning: evolutionary analysis in a fluctuating environment.Anthropological Science104, 3 (1996), 209–231

    work page 1996

  19. [19]

    Agrim Gupta, Silvio Savarese, Surya Ganguli, and Li Fei-Fei. 2021. Embodied intelligence via learning and evolution.Nature Communications12, 1 (Oct. 2021), 5721

    work page 2021

  20. [20]

    Kazuaki Harada and Hitoshi Iba. 2024. Lamarckian Co-design of Soft Robots via Transfer Learning. InProceedings of the Genetic and Evolutionary Computation Conference. 832–840

    work page 2024

  21. [21]

    Emma Hart and Léni K Le Goff. 2022. Artificial evolution of robot bodies and control: on the interaction between evolution, learning and culture.Philosophical Transactions of the Royal Society B377, 1843 (2022)

    work page 2022

  22. [22]

    Jacqueline Heinerman, Dexter Drupsteen, and Agoston Endre Eiben. 2015. Three- fold adaptivity in groups of robots: the effect of social learning. InProceedings of the 2015 annual conference on genetic and evolutionary computation. 177–183

    work page 2015

  23. [23]

    Jonathan Hiller and Hod Lipson. 2011. Automatic design and manufacture of soft robots.IEEE Transactions on Robotics28, 2 (2011), 457–466

    work page 2011

  24. [24]

    Milan Jelisavcic, Kyrre Glette, Evert Haasdijk, and A. E. Eiben. 2019. Lamarckian Evolution of Simulated Modular Robots.Frontiers in Robotics and AI6 (Feb. 2019)

    work page 2019

  25. [25]

    Kirthevasan Kandasamy, Jeff Schneider, and Barnabás Póczos. 2015. High dimen- sional Bayesian optimisation and bandits via additive models. InInternational conference on machine learning. PMLR, 295–304

    work page 2015

  26. [26]

    Kevin N Laland. 2004. Social learning strategies.Learning & behavior32, 1 (2004), 4–14

    work page 2004

  27. [27]

    Tomczak, Diederik M

    Gongjin Lan, Matteo De Carlo, Fuda Van Diggelen, Jakub M. Tomczak, Diederik M. Roijers, and A.E. Eiben. 2021. Learning directed locomotion in modular robots with evolvable morphologies.Applied Soft Computing111 (Nov. 2021), 107688

    work page 2021

  28. [28]

    Léni K Le Goff, Edgar Buchanan, Emma Hart, Agoston E Eiben, Wei Li, Matteo De Carlo, Alan F Winfield, Matthew F Hale, Robert Woolley, Mike Angus, et al

    work page

  29. [29]

    Morpho evolution with learning using a controller archive as an inheritance mechanism.IEEE Transactions on Cognitive and Developmental Systems15, 2 (2022), 507–517

    work page 2022

  30. [30]

    Julie Legrand, Seppe Terryn, Ellen Roels, and Bram Vanderborght. 2023. Recon- figurable, multi-material, voxel-based soft robots.IEEE Robotics and Automation Letters8, 3 (2023), 1255–1262

    work page 2023

  31. [31]

    Thomas Liao, Grant Wang, Brian Yang, Rene Lee, Kristofer Pister, Sergey Levine, and Roberto Calandra. 2019. Data-efficient learning of morphology and controller for a microrobot. In2019 International Conference on Robotics and Automation (ICRA). IEEE, 2488–2494

    work page 2019

  32. [32]

    Hod Lipson and Jordan B Pollack. 2000. Automatic design and manufacture of robotic lifeforms.Nature406, 6799 (2000), 974–978

    work page 2000

  33. [33]

    Jie Luo, Karine Miras, Jakub Tomczak, and Agoston E. Eiben. 2023. Enhancing robot evolution through Lamarckian principles.Scientific Reports13, 1 (Nov. 2023), 21109

    work page 2023

  34. [34]

    Stuurman, Jakub M

    Jie Luo, Aart C. Stuurman, Jakub M. Tomczak, Jacintha Ellers, and Agoston E. Eiben. 2022. The Effects of Learning in Morphologically Evolving Robot Systems. Frontiers in Robotics and AI9 (May 2022), 797393

    work page 2022

  35. [35]

    Henry B Mann and Donald R Whitney. 1947. On a test of whether one of two random variables is stochastically larger than the other.The annals of mathematical statistics(1947), 50–60

    work page 1947

  36. [36]

    Eric Medvet, Alberto Bartoli, Andrea De Lorenzo, and Giulio Fidel. 2020. Evolu- tion of distributed neural controllers for voxel-based soft robots. InProceedings of the 2020 Genetic and Evolutionary Computation Conference. 112–120

    work page 2020

  37. [37]

    Eric Medvet and Francesco Rusin. 2022. Impact of morphology variations on evolved neural controllers for modular robots. InItalian Workshop on Artificial Life and Evolutionary Computation. Springer, 266–277

    work page 2022

  38. [38]

    Jessica Mégane, Eric Medvet, Nuno Lourenço, and Penousal Machado. 2024. Grammar-Based Evolution of Polyominoes. InEuropean Conference on Genetic Programming (Part of EvoStar). Springer, 56–72

    work page 2024

  39. [39]

    Alican Mertan and Nick Cheney. 2023. Modular controllers facilitate the co- optimization of morphology and control in soft robots. InProceedings of the Genetic and Evolutionary Computation Conference. 174–183

    work page 2023

  40. [40]

    Alican Mertan and Nick Cheney. 2024. Investigating Premature Convergence in Co-optimization of Morphology and Control in Evolved Virtual Soft Robots. In European Conference on Genetic Programming (Part of EvoStar). Springer

    work page 2024

  41. [41]

    Alican Mertan and Nick Cheney. 2024. Towards Multi-Morphology Controllers with Diversity and Knowledge Distillation. InProceedings of the Genetic and Evolutionary Computation Conference. 367–376

    work page 2024

  42. [42]

    Alican Mertan and Nick Cheney. 2025. Controller Distillation Reduces Fragile Brain-Body Co-Adaptation and Enables Migrations in MAP-Elites. InProceedings of the Genetic and Evolutionary Computation Conference. 158–166

    work page 2025

  43. [43]

    Alican Mertan and Nick Cheney. 2025. Evolutionary Brain-Body Co-Optimization Consistently Fails to Select for Morphological Potential.arXiv preprint arXiv:2508.17464(2025)

    work page arXiv 2025

  44. [44]

    Karine Miras, Matteo De Carlo, Sayfeddine Akhatou, and A. E. Eiben. 2020. Evolving-Controllers Versus Learning-Controllers for Morphologically Evolv- able Robots. InApplications of Evolutionary Computation. Vol. 12104. Springer International Publishing, Cham, 86–99

    work page 2020

  45. [45]

    Jonas Mockus. 1974. On Bayesian methods for seeking the extremum. InPro- ceedings of the IFIP Technical Conference. 400–404

    work page 1974

  46. [46]

    Marco A Montes de Oca and Thomas Stützle. 2008. Towards incremental social learning in optimization and multiagent systems. InProceedings of the 10th annual conference companion on Genetic and evolutionary computation. 1939–1944

    work page 2008

  47. [47]

    Rodrigo Moreno and Andres Faina. 2022. Out of time: On the constrains that evolution in hardware faces when evolving modular robots. InInternational Conference on the Applications of Evolutionary Computation (Part of EvoStar). Springer, 667–682

    work page 2022

  48. [48]

    Giorgia Nadizar, Eric Medvet, and Dennis G Wilson. 2025. Enhancing Adaptabil- ity in Embodied Agents: A Multi-Quality-Diversity Approach.IEEE Transactions on Evolutionary Computation(2025)

    work page 2025

  49. [49]

    2016.Evolu- tionary robotics

    Stefano Nolfi, Josh Bongard, Phil Husbands, and Dario Floreano. 2016.Evolu- tionary robotics. Springer

    work page 2016

  50. [50]

    Ingo Paenke, Bernhard Sendhoff, Jon Rowe, and Chrisantha Fernando. 2007. On the adaptive disadvantage of Lamarckianism in rapidly changing environments. InEuropean Conference on Artificial Life. Springer, 355–364

    work page 2007

  51. [51]

    2006.How the body shapes the way we think: a new view of intelligence

    Rolf Pfeifer and Josh Bongard. 2006.How the body shapes the way we think: a new view of intelligence. MIT press

    work page 2006

  52. [52]

    Federico Pigozzi, Federico Julian Camerota Verdù, and Eric Medvet. 2023. How the morphology encoding influences the learning ability in body-brain co- optimization. InProceedings of the Genetic and Evolutionary Computation Confer- ence. 1045–1054

    work page 2023

  53. [53]

    Federico Pigozzi, Eric Medvet, Alberto Bartoli, and Marco Rochelli. 2023. Fac- tors impacting diversity and effectiveness of evolved modular robots.ACM Transactions on Evolutionary Learning3, 1 (2023), 1–33

    work page 2023

  54. [54]

    Luke Rendell, Robert Boyd, Daniel Cownden, Marquist Enquist, Kimmo Eriksson, Marc W Feldman, Laurel Fogarty, Stefano Ghirlanda, Timothy Lillicrap, and Kevin N Laland. 2010. Why copy others? Insights from the social learning strategies tournament.Science328, 5975 (2010), 208–213

    work page 2010

  55. [55]

    Robert G Reynolds. 1994. An introduction to cultural algorithms. InProceedings of the third annual conference on evolutionary programming, Vol. 24. World Scientific, 131–139

    work page 1994

  56. [56]

    Takahiro Sasaki and Mario Tokoro. 1999. Evolving learnable neural networks under changing environments with various rates of inheritance of acquired characters: comparison of Darwinian and Lamarckian evolution.Artificial Life5, 3 (1999), 203–223

    work page 1999

  57. [57]

    Karl Sims. 1994. Evolving 3D morphology and behavior by competition.Artificial life1, 4 (1994), 353–372

    work page 1994

  58. [58]

    Fuda van Diggelen, Eliseo Ferrante, and AE Eiben. 2024. Comparing Robot Controller Optimization Methods on Evolvable Morphologies.Evolutionary K. Ege de Bruin, Kyrre Glette, Kai Olav Ellefsen, Giorgia Nadizar, and Eric Medvet Computation32, 2 (2024), 105–124

    work page 2024

  59. [59]

    Ziyu Wang, Frank Hutter, Masrour Zoghi, David Matheson, and Nando De Feitas

    work page

  60. [60]

    Journal of Artificial Intelligence Research55 (2016), 361–387

    Bayesian optimization in a billion dimensions via random embeddings. Journal of Artificial Intelligence Research55 (2016), 361–387

    work page 2016

  61. [61]

    Ciyou Zhu, Richard H Byrd, Peihuang Lu, and Jorge Nocedal. 1997. Algorithm 778: L-BFGS-B: Fortran subroutines for large-scale bound-constrained optimization. ACM Transactions on mathematical software (TOMS)23, 4 (1997), 550–560

    work page 1997