Efficient and Diverse Generative Robot Designs using Evolution and Intrinsic Motivation

Leni K. Le Goff; Sim\'on C. Smith

REVIEW 3 major objections 6 minor 31 references

Efficient and Diverse Generative Robot Designs using Evolution and Intrinsic Motivation

T0 review · 3 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Homeokinesis makes evolved robot designs outperform fixed-controller baselines on every tested task.

desk verdict Homeokinesis as a cheap evaluation proxy in morpho-evolution is a real idea, but the downstream superiority claim is confounded by asymmetric selection thresholds and a missing MEL baseline. read the letter →

arxiv 2411.18423 v2 pith:IQZTR6AP submitted 2024-11-27 cs.RO

Leni K. Le Goff , Simón C. Smith This is my paper · ORCID

classification cs.RO

keywords morpho-evolutionhomeokinesisintrinsicmotivationgenerativerobotdesignevolutionaryalgorithmstask-agnosticdiversity

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that replacing the expensive controller-learning phase in evolutionary robot design with a homeokinetic controller—an intrinsic-motivation rule that makes any embodied robot explore its surroundings within seconds—lets evolution generate robot bodies faster, in greater diversity, and with better downstream performance. The authors compare their method, MEHK, against morpho-evolution with a fixed random controller (MEFC) on four tasks: hill climbing, flat and rough locomotion, and object manipulation. MEHK-generated designs score higher in all four tasks, reach higher exploration scores during generation, and produce a more even spread of component types, using the same wall-clock time as the baseline. The practical stake is that co-evolving morphology and control can scale to larger design spaces without the compute cost that has blocked prior morpho-evolution-with-learning systems.

What carries the argument

The load-bearing object is the homeokinetic controller, a pseudo-linear controller updated by stochastic gradient descent on the time-loop error $\mathrm{TLE} = \| L^{-1} E \|^2$, where $E$ is the forward-model prediction error and $L$ the Jacobian of the sensorimotor loop. Minimising the TLE balances predictable behaviour against sensitive behaviour, producing exploratory motion in a few seconds with no task reward. In MEHK this controller replaces the learning phase: each design's fitness is the fraction of 64 arena cells visited during 20 minutes of homeokinetic exploration, and evolution runs through the asynchronous AME algorithm on CPPN-encoded bodies.

What would settle it

Evolve designs with MEHK but replace the homeokinetic score with a random score that matches the same distribution of exploration values, keeping the Pareto selection and NCMA-ES training identical; if downstream performance does not drop, the homeokinetic signal itself is not carrying the result.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that homeokinesis, a parameter-free intrinsic-motivation controller that maximises predictability and sensitivity via the time-loop error, is an effective and cheap evaluation policy for morpho-evolution. By scoring each robot body on how much of an 8-by-8 arena it covers in 20 minutes of homeokinetic exploration, the evolutionary loop can filter for functionally viable bodies—ones that move and interact with the environment—without task-specific training. From 10,000 generated designs, the authors select three per replicate from a Pareto front of exploration fitness and morphological sparsity, train Elman-network controllers with NCMA-ES, and find that the MEHK bodies outperform MEFC bodies on all four downstream tasks, with statistical significance at p<$10^{-4}$. The method generates 10,000 robots in about 40 hours on 64 CPUs, which the authors contrast with the 1,152 CPUs and 4,000 designs reported for a deep-RL morpho-evolution baseline.

Load-bearing premise

The 20-minute exploration score under homeokinesis is a valid proxy for whether a robot body will train well on unrelated later tasks; the authors themselves note the score has minimal impact on downstream performance.

Editorial extensions

If this is right

MEHK designs outperform MEFC designs in all four downstream tasks, so the homeokinetic exploration score is a viable cheap proxy for body viability in morpho-evolution.
The method generates and evaluates 10,000 robot designs in about 40 hours on 64 CPUs, substantially lowering the compute barrier compared to deep-RL morpho-evolution approaches.
MEHK produces a more diverse mixture of component types than MEFC, countering the premature convergence that learning-based evaluation tends to induce.
The Pareto selection between exploration fitness and morphological sparsity yields both generalist robots and specialists that excel at single tasks.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The paper's own observation that the exploration score has minimal impact on downstream performance suggests the advantage of MEHK may come primarily from the diversity of body morphologies it produces rather than from homeokinesis selecting intrinsically 'good' bodies; a selection mechanism preserving the same diversity might perform equally well.
The same morpho-evolution framework could serve as a testbed for comparing other intrinsic-motivation signals, such as empowerment or predictive information, as cheap viability filters for robot bodies.
The fixed-controller baseline's random weights bias it toward jointed bodies, so part of MEFC's weakness may stem from a controller-design mismatch rather than the absence of intrinsic motivation; a stronger baseline would use a controller tuned per morphology without full task-specific learning.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague,

The genuinely new piece here is using homeokinesis as a fast, task-agnostic evaluation proxy during morpho-evolution, including exteroceptive sensors. The MEHK vs MEFC comparison is replicated 30 times, and the exploration-phase results look solid: homeokinesis covers more of the arena and produces different morphological biases than a fixed random network. The paper also makes a useful practical point: 10,000 designs in roughly 40 hours on 64 CPUs is much cheaper than deep-RL MEL, which the authors fairly attribute to prior work.

The soft spots sit in the downstream comparison. The abstract says "compared with current MEL methods," but the experiments only include MEFC, a fixed-controller baseline; no learning-based MEL baseline appears. More importantly, the selection step before NCMA-ES training uses different fitness thresholds for the two methods: 0.4 for MEHK and 0.2 for MEFC. Since MEHK's exploration scores are systematically higher, the 0.4 cut is less restrictive and the two selected sets are not drawn from comparable regions of the fitness distribution. The authors themselves note in Section IV-B that exploration score has minimal impact on downstream performance, which undercuts the idea that the higher threshold selects for trainability. The asymmetry could instead favor morphological traits, such as the wheel-heavy designs shown in Fig. 6, that make the downstream tasks easier. There is no matched-threshold control and no threshold sensitivity analysis. That is a real confound, not a nitpick.

To the authors' credit, the proxy assumption itself is honestly discussed: exploration is necessary but not sufficient, and the paper says so. I do not think that alone sinks it, but it does mean the headline claim that MEHK designs "score higher in all the tasks" is not yet established as a property of homeokinesis per se. The missing code and data also make it hard to verify the cost and reproducibility claims.

Who is this for? Evolutionary robotics researchers, especially those working on morpho-evolution, will want to read it. It deserves a serious referee: the method is clearly described, the core comparison is replicated, and the threshold confound is addressable with additional controls. I would not cite the downstream superiority claim as established, but I might cite the approach as a cost-reducing alternative worth testing.

Referee Report

3 major / 6 minor

Summary. This paper proposes MEHK, a morpho-evolution algorithm that evaluates robot designs using a homeokinetic controller to generate exploratory behavior instead of training a task-specific controller during evolution. The authors compare MEHK to MEFC, a morpho-evolution baseline with a fixed feed-forward controller, on four downstream tasks after selecting three designs per replicate via a fitness threshold and a Pareto-based sparsity selection. They report that MEHK-generated designs achieve higher task scores on all four tasks, similar wall-clock time (about 40 hours for 10,000 designs on 64 CPUs), and different morphological diversity compared to MEFC. The manuscript includes 30 replicates per condition and reports Brunner-Munzel tests with p-values below 1e-4 on most tasks.

Significance. If the result holds, MEHK offers a computationally cheap method for generating robot morphologies without per-design learning, potentially scaling to larger design spaces. The paper is, to my knowledge, the first to combine homeokinesis with CPPN-based morpho-evolution and to include exteroceptive sensors in this framework. Strengths include the clean 30-replicate protocol, the use of statistical testing, and the analysis of morphological distributions. However, the central comparison is burdened by asymmetric selection thresholds between MEHK and MEFC, and the claimed comparison with 'current MEL methods' is not actually performed; the only baseline is a fixed-controller evolution. These issues require careful re-analysis and re-framing before the headline claims can be accepted.

major comments (3)

[Section III-B (Robotic design selection)] The selection step uses asymmetric thresholds: '0.4 for MEHK and 0.2 for MEFC.' Since Fig. 5 shows MEHK exploration scores are systematically higher (median 0.8 vs 0.45), these thresholds select different percentiles of the two distributions. The paper's own Section IV-B states that exploration score has 'minimal impact' on downstream performance, which means the higher threshold does not have an obvious mechanism for selecting more trainable designs. The observed downstream advantage could therefore be caused by the threshold asymmetry or by morphological correlates (e.g., more wheels and fewer limbs in MEHK, per Fig. 6) rather than by homeokinesis acting as a viability filter. The authors should provide a threshold-sensitivity analysis with matched thresholds, or compare downstream performance of all generated designs without threshold filtering.
[Abstract and Section I (Introduction)] The abstract claims comparison with 'current MEL methods,' but the only in-paper baseline is MEFC, a morpho-evolution with a fixed controller and no learning phase (Section II-A-c). This overstates the evidence: the efficiency argument relies on a comparison with prior work (Gupta et al. [6]) rather than an in-paper MEL baseline. Moreover, Fig. 5 shows MEHK and MEFC use the same wall-clock time, so the abstract's 'quickly generated compared to morpho-evolution with static parameters' is not supported by the reported experiment. Please either add a genuine MEL baseline (e.g., AME with CMA-ES evaluation) or carefully revise the claims and abstract to state that the comparison is against fixed-controller morpho-evolution.
[Section I and Section IV-B] The paper motivates homeokinesis as a proxy for design viability: 'homeokinesis is a proxy towards optimal designs, filtering designs unsuitable for movement' (Section I) and 'The exploration task biases the designs for functional combinations...' (Section III-A). Yet Section IV-B states that 'the exploration score... has minimal impact on the performance of the downstream tasks.' The mechanism by which MEHK improves downstream performance is therefore not established, and the downstream advantage could be a side-effect of selection on morphological traits. I recommend reporting the correlation between exploration fitness and downstream task scores within each condition, and comparing random subsamples with matched thresholds, to directly test the proxy-validity claim.

minor comments (6)

[Figure 2 caption] The caption contains a typo: 'homoekinesis' should be 'homeokinesis.'
[Acknowledgements] The word 'Unver' should likely be 'University' in 'Edinburgh Napier Unver.'
[Section III-B] The sparsity score uses 15 nearest neighbours; no justification or sensitivity analysis is provided for this choice.
[Section IV-B] The hill-climbing advantage is modest (medians around 0.4 vs 0.5) and only a tail of MEHK robots performs substantially better; reporting effect sizes or confidence intervals alongside p-values would strengthen the conclusions.
[Section IV-A] The statement that MEHK 'can generate a wider diversity of designs than MEFC when comparing components' needs reconciliation with Fig. 7, which shows MEFC has higher chassis diversity; the claims about diversity are therefore more nuanced than the text suggests.
[Section III-C and reproducibility] The paper states 'The code will be available upon acceptance' but no code or detailed simulation parameters (physics engine, timestep, sensor noise model) are provided; for a computational paper, shipping code or a complete parameter listing would substantially improve reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the downstream comparison is an independent empirical evaluation, not a reduction to inputs; the asymmetric selection threshold is a correctness confound, not a circular step.

full rationale

The derivation chain is self-contained and does not reduce to its own inputs. The paper evolves designs using a homeokinesis-driven exploration score, selects a small subset, and then independently trains those designs on four downstream tasks with NCMA-ES. The downstream task scores are never used during evolution, selection, or the homeokinetic fitness computation, so no fitted parameter is renamed as a prediction. The stated claim that 'the exploration score has minimal impact on the performance of the downstream tasks' (Section IV-B) further confirms that the downstream results are not a relabeling of the selection fitness. The self-citations to AME, NCMA-ES, ARE, and homeokinetic learning are prior building blocks with external published content; invoking them as components of the pipeline is not load-bearing circularity because the novelty and the reported comparison are empirical outcomes of the combined system. The asymmetric selection thresholds of 0.4 for MEHK and 0.2 for MEFC (Section III-B) are a legitimate methodological concern and could confound the comparison, but confounding is not circularity: the selected MEHK designs are still trained by a separate optimization process, and no downstream score is equal by construction to the exploration fitness. The paper is therefore best assessed as having no significant circularity, with the threshold asymmetry flagged as a correctness risk rather than as a circular step.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The method depends on prior autonomous components (AME, ARE design space, homeokinesis) and on the paper's proxy assumption. The targets of the downstream tasks are never used during evolution, so no target-fitting circularity was found; the main free choices are the selection thresholds and the NCMA-ES annealing schedule.

free parameters (4)

exploration fitness selection threshold = 0.4 for MEHK, 0.2 for MEFC
Section III-B: different thresholds for the two methods before selecting designs for downstream training can bias the comparison.
novelty weight increment in NCMA-ES = 0.05
Section II-B: manually chosen annealing step that controls the transition from novelty search to task optimization; affects downstream training but not generation.
number of nearest neighbours in sparsity score = 15
Section III-B: fixed constant used to compute the morphological diversity score.
hidden neurons in controller networks = 6
Section II-C: fixed architecture for MEFC and NCMA-ES controllers; not varied across experiments.

assumptions (4)

domain assumption Homeokinesis as described in Der and Martius produces effective exploratory behavior in arbitrary robot morphologies within seconds.
The fitness of every evolved design is computed from this behavior; if homeokinesis works only on some morphologies, the evaluation is distorted.
ad hoc to paper The coverage score over an 8 by 8 arena is a sufficient proxy for design viability on unseen downstream tasks.
Section III-A defines fitness as visited cells; Section IV-B admits that exploration score has minimal impact on downstream performance, so this proxy is not independently validated.
standard math The inverse of the Jacobian L exists for the sensorimotor loop, as required by the time-loop-error update in Eq. 1.
Footnote 1: the TLE definition assumes an inverse of L exists; the paper applies an input shift and Taylor expansion to compute it.
ad hoc to paper The fixed-controller baseline MEFC is representative of 'morpho-evolution with static parameters' and is an adequate stand-in for the claimed comparison against MEL.
Section II-A-c: MEFC draws random weights and does not learn; the abstract says 'current MEL methods' but the experiments only test this static baseline.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Efficient and Diverse Generative Robot Designs using Evolution and Intrinsic Motivation." pith.science (2026). https://pith.science/paper/IQZTR6AP

@misc{pith2026241118423,
  author       = {Pith},
  title        = {Pith review of: Efficient and Diverse Generative Robot Designs using Evolution and Intrinsic Motivation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IQZTR6AP}},
  note         = {Machine review of arXiv:2411.18423}
}

read the original abstract

Methods for generative design of robot physical configurations can automatically find optimal and innovative solutions for challenging tasks in complex environments. The vast search-space includes the physical design-space and the controller parameter-space, making it a challenging problem in machine learning and optimisation in general. Evolutionary algorithms (EAs) have shown promising results in generating robot designs via gradient-free optimisation. Morpho-evolution with learning (MEL) uses EAs to concurrently generate robot designs and learn the optimal parameters of the controllers. Two main issues prevent MEL from scaling to higher complexity tasks: computational cost and premature convergence to sub-optimal designs. To address these issues, we propose combining morpho-evolution with intrinsic motivations. Intrinsically motivated behaviour arises from embodiment and simple learning rules without external guidance. We use a homeokinetic controller that generates exploratory behaviour in a few seconds with reduced knowledge of the robot's design. Homeokinesis replaces costly learning phases, reducing computational time and favouring diversity, preventing premature convergence. We compare our approach with current MEL methods in several downstream tasks. The generated designs score higher in all the tasks, are more diverse, and are quickly generated compared to morpho-evolution with static parameters.

Figures

Figures reproduced from arXiv: 2411.18423 by the authors.

**Figure 1.** Robots generated by our approach MEHK. Exploration values refer to the evaluation phase during morpho-evolution. The other values correspond [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

**Figure 2.** Asynchronous Morpho-Evolution. Our approach, MEHK, uses [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Five components for the robot’s design. a) Design-space: The robotic design-space follows the ARE project [30], [31]. A robot comprises 5 pre-designed components assembled on a free-formed voxel-based chassis ( [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Figures from the paper (5 more)

**Figure 4.** Figure 4: a) Arena used for the evaluation of the robots during MEHK generative phase. b), c) and d), are the downstream tasks environments. For clarity, d) only shows the beginning of the arena. into target velocity for each wheel or goal position for the joints. For MEFC and N…

**Figure 5.** Figure 5: Comparison of MEHK (blue) and MEFC (orange). On the left, the [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]

**Figure 7.** Figure 7: Exploration scores of each robot (dots) with respect to the dimension [PITH_FULL_IMAGE:figures/full_fig_p005_7.png]

**Figure 8.** Figure 8: Best performance scores of each robot for the different tasks. Each [PITH_FULL_IMAGE:figures/full_fig_p006_8.png]

**Figure 9.** Figure 9: MEHK results for the 4 tasks over the number of sensors (top left), [PITH_FULL_IMAGE:figures/full_fig_p006_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 29 canonical work pages

[6]

Embodied intel- ligence via learning and evolution,

A. Gupta, S. Savarese, S. Ganguli, and L. Fei-Fei, “Embodied intel- ligence via learning and evolution,” Nature communications, vol. 12, no. 1, p. 5721, 2021

2021
[1]

Art and the science of generative ai,

Z. Epstein, A. Hertzmann, I. of Human Creativity, M. Akten, H. Farid, J. Fjeld, M. R. Frank, M. Groh, L. Herman, N. Leach, et al., “Art and the science of generative ai,” Science, vol. 380, no. 6650, pp. 1110– 1111, 2023

work page 2023
[2]

Generative design: an explorative study,

F. Buonamici, M. Carfagni, R. Furferi, Y . V olpe, L. Governi, et al. , “Generative design: an explorative study,”Computer-Aided Design and Applications, vol. 18, no. 1, pp. 144–155, 2020

work page 2020
[3]

Data-efficient co-adaptation of morphology and behaviour with deep reinforcement learning,

K. S. Luck, H. B. Amor, and R. Calandra, “Data-efficient co-adaptation of morphology and behaviour with deep reinforcement learning,” in Conference on Robot Learning . PMLR, 2020, pp. 854–869

work page 2020
[4]

Reinforcement learning for freeform robot design,

M. Li, D. Matthews, and S. Kriegman, “Reinforcement learning for freeform robot design,” in 2024 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2024, pp. 8799–8806

work page 2024
[5]

Efficient automatic design of robots,

D. Matthews, A. Spielberg, D. Rus, S. Kriegman, and J. Bongard, “Efficient automatic design of robots,” Proceedings of the National Academy of Sciences , vol. 120, no. 41, p. e2305180120, 2023

work page 2023
[7]

Morpho evolution with learning using a controller archive as an inheritance mechanism,

L. K. Le Goff, E. Buchanan, E. Hart, A. E. Eiben, W. Li, M. De Carlo, A. F. Winfield, M. F. Hale, R. Woolley, M. Angus, et al. , “Morpho evolution with learning using a controller archive as an inheritance mechanism,” IEEE Transactions on Cognitive and Developmental Systems, vol. 15, no. 2, pp. 507–517, 2022

work page 2022
[8]

Lamarckian evolution of simulated modular robots,

M. Jelisavcic, K. Glette, E. Haasdijk, and A. Eiben, “Lamarckian evolution of simulated modular robots,” Frontiers in Robotics and AI , vol. 6, p. 9, 2019

work page 2019

Show all 31 references

[9]

Evaluation of frameworks that combine evolution and learning to design robots in complex morphological spaces,

W. Li, E. Buchanan, L. K. Le Goff, E. Hart, M. F. Hale, M. De Carlo, R. Wooley, A. F. Winfield, J. Timmis, and A. M. Tyrrell, “Evaluation of frameworks that combine evolution and learning to design robots in complex morphological spaces,” IEEE Transactions on Evolutionary Comp...

2023
[10]

Scalable co- optimization of morphology and control in embodied machines,

N. Cheney, J. Bongard, V . SunSpiral, and H. Lipson, “Scalable co- optimization of morphology and control in embodied machines,” Journal of The Royal Society Interface , vol. 15, no. 143, p. 20170937, 2018

2018
[11]

Investigating premature convergence in co- optimization of morphology and control in evolved virtual soft robots,

A. Mertan and N. Cheney, “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, 2024, pp. 38–55

2024
[12]

Real-world evolution adapts robot morphology and control to hard- ware limitations,

T. F. Nygaard, C. P. Martin, E. Samuelsen, J. Torresen, and K. Glette, “Real-world evolution adapts robot morphology and control to hard- ware limitations,” in Proceedings of the Genetic and Evolutionary Computation Conference, 2018, pp. 125–132

2018
[13]

Evolving 3d morphology and behavior by competition,

K. Sims, “Evolving 3d morphology and behavior by competition,” Artificial life, vol. 1, no. 4, pp. 353–372, 1994

1994
[14]

Automatic design and manufacture of robotic lifeforms,

H. Lipson and J. B. Pollack, “Automatic design and manufacture of robotic lifeforms,” Nature, vol. 406, no. 6799, pp. 974–978, 2000

2000
[15]

Unshackling evolution: evolving soft robots with multiple materials and a powerful generative encoding,

N. Cheney, R. MacCurdy, J. Clune, and H. Lipson, “Unshackling evolution: evolving soft robots with multiple materials and a powerful generative encoding,” ACM SIGEVOlution, vol. 7, no. 1, pp. 11–23, 2014

2014
[16]

If it evolves it needs to learn,

A. Eiben and E. Hart, “If it evolves it needs to learn,” in Proceed- ings of the 2020 Genetic and Evolutionary Computation Conference Companion, 2020, pp. 1383–1384

2020
[17]

The effects of learning in morphologically evolving robot systems,

J. Luo, A. C. Stuurman, J. M. Tomczak, J. Ellers, and A. E. Eiben, “The effects of learning in morphologically evolving robot systems,” Frontiers in Robotics and AI , vol. 9, p. 797393, 2022

2022
[18]

On the difficulty of co-optimizing morphology and control in evolved virtual creatures,

N. Cheney, J. Bongard, V . SunSpiral, and H. Lipson, “On the difficulty of co-optimizing morphology and control in evolved virtual creatures,” in Artificial life conference proceedings . MIT Press One Rogers Street, Cambridge, MA 02142-1209, USA journals-info . . . , 2016, pp. 226–233

2016
[19]

Guided self-organization: perception–action loops of embodied systems,

N. Ay, R. Der, and M. Prokopenko, “Guided self-organization: perception–action loops of embodied systems,” Theory in Biosciences, vol. 131, no. 3, pp. 125–127, 2012

2012
[20]

Learn goal-conditioned policy with intrinsic motivation for deep reinforcement learning,

J. Liu, D. Wang, Q. Tian, and Z. Chen, “Learn goal-conditioned policy with intrinsic motivation for deep reinforcement learning,” in Proceedings of the AAAI conference on artificial intelligence , vol. 36, no. 7, 2022, pp. 7558–7566

2022
[21]

Homeokinetic reinforcement learning,

S. C. Smith and J. M. Herrmann, “Homeokinetic reinforcement learning,” in IAPR International Workshop on Partially Supervised Learning. Springer, 2011, pp. 82–91

2011
[22]

Evaluation of internal models in autonomous learning,

——, “Evaluation of internal models in autonomous learning,” IEEE Transactions on Cognitive and Developmental Systems , vol. 11, no. 4, pp. 463–472, 2018

2018
[23]

Intrinsically motivated goal exploration processes with automatic curriculum learn- ing,

S. Forestier, R. Portelas, Y . Mollard, and P.-Y . Oudeyer, “Intrinsically motivated goal exploration processes with automatic curriculum learn- ing,” Journal of Machine Learning Research , vol. 23, no. 152, pp. 1–41, 2022

2022
[24]

Task-agnostic morphology evo- lution,

D. J. Hejna, P. Abbeel, and L. Pinto, “Task-agnostic morphology evo- lution,” in 9th International Conference on Learning Representations, ICLR 2021, 2021

2021
[25]

Empowerment: A universal agent-centric measure of control,

A. S. Klyubin, D. Polani, and C. L. Nehaniv, “Empowerment: A universal agent-centric measure of control,” in 2005 ieee congress on evolutionary computation, vol. 1. IEEE, 2005, pp. 128–135

2005
[26]

Der and G

R. Der and G. Martius, The playful machine: theoretical foundation and practical realization of self-organizing robots . Springer Science & Business Media, 2012, vol. 15

2012
[27]

Improving efficiency of evolving robot designs via self-adaptive learning cycles and an asynchronous archi- tecture,

L. Le Goff and E. Hart, “Improving efficiency of evolving robot designs via self-adaptive learning cycles and an asynchronous archi- tecture,” in Proceedings of the Genetic and Evolutionary Computation Conference Companion, 2024, pp. 1607–1615

2024
[28]

Compositional pattern producing networks: A novel abstraction of development,

K. O. Stanley, “Compositional pattern producing networks: A novel abstraction of development,” Genetic programming and evolvable machines, vol. 8, pp. 131–162, 2007

2007
[29]

Sample and time efficient policy learning with cma-es and bayesian optimisation,

L. K. Le Goff, E. Buchanan, E. Hart, A. E. Eiben, W. Li, M. De Carlo, M. F. Hale, M. Angus, R. Woolley, J. Timmis, et al. , “Sample and time efficient policy learning with cma-es and bayesian optimisation,” in Artificial Life Conference Proceedings 32 . MIT Press One Rogers St...

2020
[30]

Bootstrap- ping artificial evolution to design robots for autonomous fabrication,

E. Buchanan, L. K. Le Goff, W. Li, E. Hart, A. E. Eiben, M. De Carlo, A. F. Winfield, M. F. Hale, R. Woolley, M. Angus, et al., “Bootstrap- ping artificial evolution to design robots for autonomous fabrication,” Robotics, vol. 9, no. 4, p. 106, 2020

2020
[31]

Practical hardware for evolvable robots,

M. Angus, E. Buchanan, L. K. Le Goff, E. Hart, A. E. Eiben, M. De Carlo, A. F. Winfield, M. F. Hale, R. Woolley, J. Timmis, et al., “Practical hardware for evolvable robots,” Frontiers in Robotics and AI, vol. 10, p. 1206055, 2023

2023

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [6]

Embodied intel- ligence via learning and evolution,

A. Gupta, S. Savarese, S. Ganguli, and L. Fei-Fei, “Embodied intel- ligence via learning and evolution,” Nature communications, vol. 12, no. 1, p. 5721, 2021

2021

[2] [1]

Art and the science of generative ai,

Z. Epstein, A. Hertzmann, I. of Human Creativity, M. Akten, H. Farid, J. Fjeld, M. R. Frank, M. Groh, L. Herman, N. Leach, et al., “Art and the science of generative ai,” Science, vol. 380, no. 6650, pp. 1110– 1111, 2023

work page 2023

[3] [2]

Generative design: an explorative study,

F. Buonamici, M. Carfagni, R. Furferi, Y . V olpe, L. Governi, et al. , “Generative design: an explorative study,”Computer-Aided Design and Applications, vol. 18, no. 1, pp. 144–155, 2020

work page 2020

[4] [3]

Data-efficient co-adaptation of morphology and behaviour with deep reinforcement learning,

K. S. Luck, H. B. Amor, and R. Calandra, “Data-efficient co-adaptation of morphology and behaviour with deep reinforcement learning,” in Conference on Robot Learning . PMLR, 2020, pp. 854–869

work page 2020

[5] [4]

Reinforcement learning for freeform robot design,

M. Li, D. Matthews, and S. Kriegman, “Reinforcement learning for freeform robot design,” in 2024 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2024, pp. 8799–8806

work page 2024

[6] [5]

Efficient automatic design of robots,

D. Matthews, A. Spielberg, D. Rus, S. Kriegman, and J. Bongard, “Efficient automatic design of robots,” Proceedings of the National Academy of Sciences , vol. 120, no. 41, p. e2305180120, 2023

work page 2023

[7] [7]

Morpho evolution with learning using a controller archive as an inheritance mechanism,

L. K. Le Goff, E. Buchanan, E. Hart, A. E. Eiben, W. Li, M. De Carlo, A. F. Winfield, M. F. Hale, R. Woolley, M. Angus, et al. , “Morpho evolution with learning using a controller archive as an inheritance mechanism,” IEEE Transactions on Cognitive and Developmental Systems, vol. 15, no. 2, pp. 507–517, 2022

work page 2022

[8] [8]

Lamarckian evolution of simulated modular robots,

M. Jelisavcic, K. Glette, E. Haasdijk, and A. Eiben, “Lamarckian evolution of simulated modular robots,” Frontiers in Robotics and AI , vol. 6, p. 9, 2019

work page 2019