Robots that can adapt like animals
Open accessThe DOI matches its Crossref record; 39 of the 40 references it cites were found in a registry or an index.
Abstract
Robots have transformed many industries, most notably manufacturing, and have the power to deliver tremendous benefits to society, such as in search and rescue, disaster response, health care and transportation. They are also invaluable tools for scientific exploration in environments inaccessible to humans, from distant planets to deep oceans. A major obstacle to their widespread adoption in more complex environments outside factories is their fragility. Whereas animals can quickly adapt to injuries, current robots cannot 'think outside the box' to find a compensatory behaviour when they are damaged: they are limited to their pre-specified self-sensing abilities, can diagnose only anticipated failure modes, and require a pre-programmed contingency plan for every type of potential damage, an impracticality for complex robots. A promising approach to reducing robot fragility involves having robots learn appropriate behaviours in response to damage, but current techniques are slow even with small, constrained search spaces. Here we introduce an intelligent trial-and-error algorithm that allows robots to adapt to damage in less than two minutes in large search spaces without requiring self-diagnosis or pre-specified contingency plans. Before the robot is deployed, it uses a novel technique to create a detailed map of the space of high-performing behaviours. This map represents the robot's prior knowledge about what behaviours it can perform and their value. When the robot is damaged, it uses this prior knowledge to guide a trial-and-error learning algorithm that conducts intelligent experiments to rapidly discover a behaviour that compensates for the damage. Experiments reveal successful adaptations for a legged robot injured in five different ways, including damaged, broken, and missing legs, and for a robotic arm with joints broken in 14 different ways. This new algorithm will enable more robust, effective, autonomous robots, and may shed light on the principles that animals use to adapt to injury.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source doesn’t support the claim as written.
In the citing paper
“We render images using the MuJoCo Physics Engine's [45] built-in renderer. This renderer is not intended to be photo-realistic, and physically plausible choices of textures and lighting are not needed.”
From Domain randomization for transferring deep neural networks from simulation to the real world (2017) · cites this paper as [45]
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“under Quality-Diversity optimization (Pugh et al., 2016), the search space is divided into behavioral niches to evolve diverse solutions, a paradigm that challenged the dominant approach of training a single agent, as it proved more robust in real-world settings (Cully et al., 2015).”
From Dynamics of niche construction in adaptable populations evolving in diverse environments (2023) · cites this paper as Cully et al., 2015
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Similarly, extensive reliance on simulation and offline training further distances learning from the full variability and constraint structure of real sensorimotor interaction (Jakobi et al., 1995; Cully et al., 2015).”
From Toward Enactive Artificial Intelligence (2026) · cites this paper as Cully et al., 2015
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Gaussian Processes for Machine Learning
Rasmussen, C. E., & Williams, C. K. I. (2005) · doi:10.7551/mitpress/3206.001.0001 · Cited by 11,705
Corke, P., Roberts, J., Cunningham, J., & Hainsworth, D. (2008) · Springer Handbook of Robotics · doi:10.1007/978-3-540-30301-5_50 · Cited by 3,543
A survey of robot learning from demonstration
Argall, B. D., Chernova, S., Veloso, M., & Browning, B. (2009) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2008.10.024 · Cited by 3,227
Reinforcement learning in robotics: A survey
Kober, J., Bagnell, J. A., & Peters, J. (2013) · The International Journal of Robotics Research · doi:10.1177/0278364913495721 · Cited by 3,160 · cited 9×
“The key insight is that, whereas current learning algorithms either start with no knowledge of the search space 17 or with minimal knowledge from a few human demonstrations 17,18 , animals better understand the space of possible behaviors and their value from previous experience 19 , enabling…”
BackgroundMethodResultInfluentialDiagnosis and Fault-Tolerant Control
Blanke, M., Kinnaert, M., Lunze, J., & Staroswiecki, M. (2003) · Springer eBooks · doi:10.1007/978-3-662-05344-7 · Cited by 2,789
Nonlinear Programming: Sequential Unconstrained Minimization Techniques
Fiacco, A. V., & McCormick, G. P. (1968) · Defense Technical Information Center (DTIC) · Cited by 2,262
Stanley: The robot that won the DARPA Grand Challenge
Thrun, S., Montemerlo, M., Dahlkamp, H., Stavens, D., Aron, A., Diebel, J., Fong, P., Gale, J., Halpenny, M., Hoffmann, G., et al. (2006) · Journal of Field Robotics · doi:10.1002/rob.20147 · Cited by 2,237 · cited 2×
“3C, undam-aged: 24 [16; 41] seconds, 3 [2; 5] physical trials, repaired: 29 [16; 82] seconds, 3.5 [2; 10] trials).”
BackgroundInfluentialBrochu, E., Cora, V. M., & de Freitas, N. (2010) · arXiv (Cornell University) · doi:10.48550/arxiv.1012.2599 · Cited by 2,011
Bayesian integration in sensorimotor learning
Körding, K. P., & Wolpert, D. M. (2004) · Nature · doi:10.1038/nature02169 · Cited by 2,004 · cited 1×
“…optimization procedure followed by our robot appears similar to the technique employed by humans when they optimize an unknown function 23 , and there is strong evidence that animal brains learn probability distributions, combine them with prior knowledge, and act as Bayesian optimizers 25,26 .”
MethodControl of mental activities by internal models in the cerebellum
Ito, M. (2008) · Nature Reviews Neuroscience · doi:10.1038/nrn2332 · Cited by 1,356 · cited 1×
“A final parallel is that the simulator and Gaussian process components of Intelligent Trial and Error are two forms of predictive models, which are known to exist in animals 29,12 .”
Research works printed but not linked: 2
- Siciliano, B. & Khatib, O. Springer Handbook of Robotics (Springer, 2008). · doi:10.1007/978-3-540-30301-5
- Kluger, J. & Lovell, J. Apollo 13 (Mariner Books, 2006).
Cited by
Items 701–800 of 1,007
Harnessing Game-Inspired Content Creation for Intuitive Generative Design and Optimization
Villaggi, L., Stoddart, J., & Gaier, A. (2022) · Springer eBooks · doi:10.1007/978-3-031-13249-0_13 · Cited by 5
Quality-diversity for Synthesizer Sound Matching
Masuda, N., & Saito, D. (2023) · Journal of Information Processing · doi:10.2197/ipsjjip.31.220 · Cited by 3
Open-Ended Diverse Solution Discovery with Regulated Behavior Patterns for Cross-Domain Adaptation
Xu, K., Ma, Y., Wei, B., & Li, W. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i9.26257 · Cited by 3
Evolving Dynamic Locomotion Policies in Minutes
Chatzilygeroudis, K. I., Tsakonas, C. G., & Vrahatis, M. N. (2023) · International Conference on Information, Intelligence, Systems & Applications (IISA) · doi:10.1109/iisa59645.2023.10345937 · Cited by 3
“It is worth noting as well, that the baseline is trained using a parameterized open-loop periodic signal (similar to [9], [10]) and position control.”
BackgroundInfluentialQuality Diversity for Visual Pre-Training
Chavhan, R., Gouk, H., Li, D., & Hospedales, T. M. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00496 · Cited by 6
“Access to a suite of meaningfully diverse high-quality solutions often has substantial robustness benefits that are widely exploited in robotics community [20].”
BackgroundMethodSearch Space Illumination of Robot Swarm Parameters for Trustworthy Interaction
Wilson, J., & Hauert, S. (2024) · Springer proceedings in advanced robotics · doi:10.1007/978-3-031-51497-5_13 · Cited by 3
A Hierarchical Approach to Evolving Behaviour-Trees for Swarm Control
Montague, K., Hart, E., & Paechter, B. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-56852-7_12 · Cited by 3
Covariance Matrix Adaptation MAP-Annealing: Theory and Experiments
Zhao, S., Tjanaka, B., Fontaine, M. C., & Nikolaidis, S. (2025) · ACM Transactions on Evolutionary Learning and Optimization · doi:10.1145/3665336 · Cited by 5
Evolving Generalist Controllers to Handle a Wide Range of Morphological Variations
Triebold, C., & Yaman, A. (2024) · Genetic and Evolutionary Computation Conference (GECCO) · doi:10.1145/3638529.3654116 · Cited by 3
“For example, physical damages can cause changes in the shape of the robot that the trained ANN controller is not capable of handling [5].”
BackgroundInfluentialOn the robustness of lexicase selection to contradictory objectives
Shahbandegan, S., & Dolson, E. L. (2024) · Genetic and Evolutionary Computation Conference (GECCO) · doi:10.1145/3638529.3654215 · Cited by 3
“For example, taking inspiration from [3], we could imagine that we are evolving gaits for a six-legged robot.”
BackgroundReconfigurable Robot Identification from Motion Data
Hu, Y., Wang, Y., Liu, R., Shen, Z., & Lipson, H. (2024) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10801809 · Cited by 4
“To overcome these limitations, there’s a growing interest in developing self-modeling robots [5]–[7].”
Fast and modular whole-body Lagrangian dynamics of legged robots with changing morphology
Farghdani, S., Abdelrahman, O., & Chhabra, R. (2026) · Nonlinear Dynamics · doi:10.1007/s11071-026-12594-6 · Cited by 3
“Because of its accessible documentation and API, OpenDE is widely used to simulate MLRs dynamics [89, 90].”
Polkowski, L. (2017) · Lecture notes in computer science · doi:10.1007/978-3-319-60837-2 · Cited by 3
Distributed Adaptive Locomotion Learning in ModRED Modular Self-reconfigurable Robot
Dutta, A., Dasgupta, P., & Nelson, C. A. (2018) · Springer proceedings in advanced robotics · doi:10.1007/978-3-319-73008-0_24 · Cited by 3
Automating the Incremental Evolution of Controllers for Physical Robots
Faíña, A., Jacobsen, L. T., & Risi, S. (2017) · Artificial Life · doi:10.1162/artl_a_00226 · Cited by 3
“[8], which could limit the number of evaluations that have to be performed in the real world.”
BackgroundXie, B., & Maciejewski, A. A. (2017) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2017.2715879 · Cited by 3
“on optimizing robot configurations in anticipation of failures [17], while others on the self-repair after a failure occurs [18].”
BackgroundClonal plasticity: an autonomic mechanism for multi-agent systems to self-diversify
Nallur, V., & Clarke, S. (2017) · Autonomous Agents and Multi-Agent Systems · doi:10.1007/s10458-017-9380-x · Cited by 4
“Due to real-time response requirements, other phenotypic diversity mechanisms such asMAP-Elites [106] use pre-computation to explore the possible behaviour space before deployment.”
BackgroundMethodTrivun, D., Dindo, H., & Lačević, B. (2017) · International Conference on Information, Communication and Automation Technologies (ICAT) · doi:10.1109/icat.2017.8171613 · Cited by 5
“Also, a novel approach is based on trial-and-error algorithm after mapping the robot’s prior knowledge [4].”
BackgroundGras, G., Seneci, C., Γιαταγάνας, Π., & Yang, G. (2018) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2018.8460588 · Cited by 3
“Bayesian optimization has been successfully used in a number of mobile robotic applications such as gait optimization for bipedal locomotion [5], gait adaptation after incurring damage in a 6-legged robot [6], gaze planning [7], and path planning [8].”
MethodZhang, Y., Luo, X., Zhu, C. L., Lu, W., & Wang, X. (2019) · Tehnicki vjesnik - Technical Gazette · doi:10.17559/tv-20190104090943 · Cited by 3
Constructive Policy: Reinforcement Learning Approach for Connected Multi-Agent Systems
Raza, S. J. A., & Lin, M. (2019) · IEEE International Conference on Automation Science and Engineering (CASE) · doi:10.1109/coase.2019.8843223 · Cited by 3
“Agents that possess intricate geometrical structure (like bioinspired robots or soft robots) cannot be explicitly modeled due to their redundant kinematic properties and due to larger pool of possible sub-optimal solutions that make it difficult to converge towards most optimal solution [14] [15].”
BackgroundA Gradient-Based Algorithm to Deceive Deep Neural Networks
Xie, T., & Li, Y. (2019) · Communications in Computer and Information Science · doi:10.1007/978-3-030-36808-1_7 · Cited by 3
Robots Solving the Urgent Problems by Themselves: A Review
Guo, D., Fu, L. H., & Wang, L. (2019) · IEEE International Conference on Consumer Electronics - Taiwan (ICCE-Taiwan) · doi:10.1109/icce-tw46550.2019.8991947 · Cited by 3
“In order to reduce the time of learning process, Antoine Cully [7] proposed an intelligent trial-and-error (IT&E) algorithm which allowed the robot adapt to damage within less than two minutes in large search space without a self-diagnosis or pre-designated contingency plan.”
MethodMOVEMENT ALGORITHM OF AN AUTONOMOUS ROBOT-HEXAPOD FOR MOVING IN NARROW CLOSED SPACES
Platov, I. M., & Pavlovskyi, O. (2021) · Bulletin of Kyiv Polytechnic Institute Series Instrument Making · doi:10.20535/1970.61(1).2021.237103 · Cited by 3
Reprint of: Application cases of biological transformation in manufacturing technology
Bergs, T., Schwaneberg, U., Barth, S., Hermann, L., Grunwald, T., Mayer, S., Biermann, F., & Sözer, N. (2021) · CIRP journal of manufacturing science and technology · doi:10.1016/j.cirpj.2021.06.008 · Cited by 3
Phenotypic Niching Using Quality Diversity Algorithms
Hagg, A. (2021) · Natural computing series · doi:10.1007/978-3-030-79553-5_12 · Cited by 6
Past Visions of Artificial Futures:
Taylor, T., & Dorin, A. (2018) · Conference on Artificial Life (ALIFE) · doi:10.1162/isal_a_00022 · Cited by 3
“The idea of self-repairing machines is present in the work of Eliot, Forster, Campbell [ The Machine ] and Bernal, and this is indeed a theme in current evolutionary robotics research, e.g. (Bongard et al., 2006), (Cully et al., 2015).”
BackgroundExploration and generalization in vast spaces
Wu, C. M., Schulz, E., Speekenbrink, M., Nelson, J. D., & Meder, B. (2017) · Max Planck Digital Library · Cited by 9
Coevo: a collaborative design platform with artificial agentsPreprint
Serra, G., & Miralles, D. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1904.13333 · Cited by 3
“To address this topic, evolutionary computing [6] has been demonstrated as a powerful tool to explore solution space and generate valid proposals [7] [8] [9].”
MethodKutsuzawa, K., Kusano, H., Kume, A., & Yamaguchi, S. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1910.03253 · Cited by 3
“cGANs do not require designing the characteristics of the latent space by hand, unlike other methods [25].”
BackgroundSelf-Referential Quality Diversity Through Differential Map-ElitesPreprint
Choi, T. J., & Togelius, J. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.04964 · Cited by 3
“MAP-Elites is a QD algorithm that has seen use in domains as different as robot locomotion planning [2, 16], game level generation [1, 7, 10, 11], game strategy characterization [5], and workforce scheduling [21].”
MethodSchoepp, S., Taghian, M., Miwa, S., Mitsuka, Y., Golestan, S., & Zai͏̈ane, O. R. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2407.15283 · Cited by 5
“Cully et al. (2014) used a robot simulator to construct a low-dimensional behaviour performance map.”
Forestier, S. (2019) · theses.fr (ABES) · doi:10.70675/abd58967z79d0z4df2zbe94z307821e9b628 · Cited by 7
Speculative Rollback Correction for Quality-Diverse Web Agent ImitationPreprint
Hao, L., Lin, H., Li, H., Liu, Z., Yang, Z., Hao, H., Huang, D., Yang, H., Ge, H., Xie, M. J., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2606.12485 · Cited by 3
“…replay buffering: successful trajectories compete within coarse behavioral niches, so training data can preserve multiple verified solution modes instead of only the shortest or most teacher-like path [Mouret and Clune, 2015, Pugh et al., 2016, Lehman and Stanley, 2011, Cully et al., 2015].”
AI Finds A WayPreprint
Dharna, A., Lu, C., Sullivan, R., Lehman, J., Krakovna, V., & Clune, J. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2608.23875 · Cited by 3
Self-citationDynamics of niche construction in adaptable populations evolving in diverse environments
Nisioti, E., & Moulin-Frier, C. (2023) · The 2023 Conference on Artificial Life · doi:10.1162/isal_a_00668 · Cited by 2
Alternating Optimisation and Quadrature for Robust Control
Paul, S., Chatzilygeroudis, K. I., Ciosek, K., Mouret, J.-B., Osborne, M. A., & Whiteson, S. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.11687 · Cited by 2
“We applied ALOQ to learn the optimal policy given these damage probabilities, but restricted the search to the policies in the archive created by (Cully et al. 2015).”
BackgroundMethodInfluentialSelf-citationExploration-Driven Reinforcement Learning for Avionic System Fault Detection (Experience Paper)
Tolguenec, P.-A. L., Rachelson, E., Besse, Y., Teichteil-Königsbuch, F., Schneider, N., Waeselynck, H., & Wilson, D. G. (2024) · ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA) · doi:10.1145/3650212.3680331 · Cited by 4
“The evolutionary community has thoroughly investigated the question of massive exploration and the balance with exploitation through concepts like Novelty Search[33–35] and Quality Diversity [16, 17].”
GeMuCo: Generalized Multisensory Correlational Model for Body Schema Learning
Kawaharazuka, K., Okada, K., & Inaba, M. (2024) · IEEE Robotics & Automation Magazine · doi:10.1109/mra.2024.3415111 · Cited by 2
“For Change Adaptability , existing methods rarely incorporate information about gradual model changes directly into the models [15], [16].”
MethodMatsubara, E., Yagi, S., Goto, Y., Yamamori, S., & Morimoto, J. (2025) · Artificial Life and Robotics · doi:10.1007/s10015-024-00984-1 · Cited by 2
“To date, research has been conducted on damage detection and recovery in autonomous mobile robots, particularly in the context of multi-legged robots [3–5].”
Safe Human–Robot Collaboration Using Multimodal Perception and Robot Learning
Xu, W., Yao, B., Xiang, Y., Ji, Z., & Ai, H. (2025) · Springer eBooks · doi:10.1007/978-3-031-82170-7_10 · Cited by 2
Diverse Prompts: Illuminating the Prompt Space of Large Language Models with MAP-Elites
Santos, G. M., Julia, R. M. D. S., & do Nascimento, M. Z. (2025) · IEEE Congress on Evolutionary Computation (CEC) · doi:10.1109/cec65147.2025.11043024 · Cited by 2
“Widely applied in robotics [24], reinforcement learning [25], and creative computing [26], QD algorithms are particularly suited for exploring diverse prompt structures to uncover both high-performing and distinct designs.”
Lifelong Active Inference of Gait Control
Szadkowski, R., & Faigl, J. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2025.3579814 · Cited by 2
“The body can also get damaged [5], heated [6], or dirtied that together with terramechanical properties of the terrain [7], [8], [9] a ff ect the robot-environment interaction during terrain traversal.”
Behaviour Diversity in a Walking and Climbing Centipede-Like Virtual Creature
Norstein, E. S., Yasui, K., Kano, T., Ishiguro, A., & Glette, K. (2025) · Artificial Life · doi:10.1162/artl_a_00476 · Cited by 2
Vector Quantized-Elites: Unsupervised and Problem-Agnostic Quality-Diversity Optimization
Tsakonas, C. G., & Chatzilygeroudis, K. I. (2025) · IEEE Transactions on Evolutionary Computation · doi:10.1109/tevc.2025.3631786 · Cited by 6
Limits of Lamarckian Evolution Under Pressure of Morphological Novelty
Muff, J. R., Miras, K., & Eiben, A. E. (2026) · IEEE Congress on Evolutionary Computation (CEC) · doi:10.1109/cec69861.2026.11727260 · Cited by 2
“Novelty search and quality-diversity algorithms explicitly encourage exploration of diverse body plans, often leading to more creative, adaptable, or evolvable designs [14], [15], [16].”
Hierarchical surrogate modeling for illumination algorithms
Hagg, A. (2017) · Genetic and Evolutionary Computation Conference Companion (GECCO Companion) · doi:10.1145/3067695.3082495 · Cited by 2
“Today, optimization of high-dimensional problems, such as in computational fluid dynamics [8] or robotics [7], can be extremely expensive in terms of computational effort.”
BackgroundMethodLow-noise OGV design for broadband noise using bayesian optimisation
Paruchuri, C. C., Vellanki, P., Kalyan, A., Coupland, J., & Joseph, P. (2019) · AIAA/CEAS Aeroacoustics Conference · doi:10.2514/6.2019-2489 · Cited by 2
Qi, R., Tang, Y., & Zhang, K. (2021) · Industrial Robot the international journal of robotics research and application · doi:10.1108/ir-03-2020-0057 · Cited by 2
Efficient Dynamics Estimation With Adaptive Model Sets
Ratner, E., Bajcsy, A. V., Fong, T., Tomlin, C. J., & Dragan, A. D. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3060415 · Cited by 2
“Finally, Cully et al. leverage an offline precomputation to enable the system to detect and compensate for failures (e.g. the loss of an actuator) [11]; while no new models are learned online, their system is still capable of adapting to novel situations at execution time.”
BackgroundSurvivable Robotic Control through Guided Bayesian Policy Search with Deep Reinforcement Learning
Raza, S. J. A., Dastider, A., & Lin, M. (2021) · IEEE International Conference on Automation Science and Engineering (CASE) · doi:10.1109/case49439.2021.9551472 · Cited by 2
Nakashima, S., Kawaharazuka, K., Nishiura, M., Asano, Y., Kakiuchi, Y., Okada, K., Kawasaki, K., & Inaba, M. (2021) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48506.2021.9561531 · Cited by 2
“A hexapod introduced learning-based recovery from fractures [1].”
BackgroundEvolving neural networks through bio-inspired parent selection in dynamic environments
Sunagawa, J., Yamaguchi, R., & Nakaoka, S. (2022) · Biosystems · doi:10.1016/j.biosystems.2022.104686 · Cited by 2
Out of Time: On the Constrains that Evolution in Hardware Faces When Evolving Modular Robots
Moreno, R., & Faíña, A. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-02462-7_42 · Cited by 4
Huang, L., & Zou, Z. (2022) · Winter Simulation Conference · doi:10.1109/wsc57314.2022.10015308 · Cited by 2
“Although being the most prevailing form of robots in construction and having achieved success, pre-programmed robots suffer from the deficiency of being unable to adapt to unexpected scenarios (Cully et al. 2015).”
BackgroundChen, Z., Gao, F., Xi, Q., & Zhao, Y. (2023) · Lecture notes in mechanical engineering · doi:10.1007/978-981-19-9398-5_116 · Cited by 2
Curiosity Creates Diversity in Policy Search
Tolguenec, P.-A. L., Rachelson, E., Besse, Y., & Wilson, D. G. (2022) · ACM Transactions on Evolutionary Learning and Optimization · doi:10.1145/3605782 · Cited by 2
Biemer, C. F. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment · doi:10.1609/aiide.v19i1.27540 · Cited by 2
“With these grids, they tested how quickly a grid can be updated for a new agent with the intelligent trial-and-error algorithm (Cully et al. 2015).”
MethodDiversifying Policies With Non-Markov Dispersion to Expand the Solution Space
Qu, B., Cao, X., Chang, Y., Tsang, I. W.-H., & Ong, Y.-S. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3455257 · Cited by 2
“The MAP-Elites algorithms [50], [51] solve this problem by discretizing the behavior description space into a grid of cells.”
MethodLearning Dynamic Tasks on a Large-scale Soft Robot in a Handful of Trials
Zwane, S., Cheney, D., Johnson, C. C., Luo, Y., Bekiroglu, Y., Killpack, M. D., & Deisenroth, M. P. (2024) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10802122 · Cited by 2
“Similarly, [13] confined the actuation of each leg of a spider robot to five discrete settings.”
Yang, Y., Li, J.-C., Wang, F., Lin, J., Jhih-Hao, L., & Shih, J.-W. (2024) · IEEE International Conference on Cyborg and Bionic Systems (CBS) · doi:10.1109/cbs61689.2024.10860652 · Cited by 2
Gong, H., & Zhang, Y. (2025) · Materials Today Communications · doi:10.1016/j.mtcomm.2025.112513 · Cited by 2
A Quality Diversity Approach to Evolving Model Rockets
Schrum, J., & Crosby, C. O. (2025) · Genetic and Evolutionary Computation Conference (GECCO) · doi:10.1145/3712256.3726338 · Cited by 2
Qin, H., Xiang, Y., Han, Y., Wang, Y., Chen, Q., & Zhou, P. (2025) · Scientia Sinica Informationis · doi:10.1360/ssi-2025-0113 · Cited by 2
RL + Transformer = A General-Purpose Problem Solver
Rentschler, M., & Roberts, J. (2025) · Workshop for Research on Agent Language Models (REALM) · doi:10.18653/v1/2025.realm-1.29 · Cited by 4
Veloz, T. (2025) · Frontiers in Cognition · doi:10.3389/fcogn.2025.1618381 · Cited by 2
“…and Multi-dimensional Archive of Phenotypic Elites (MAP-Elites) demonstrate how maintaining behavioral diversity rather than focusing solely on performance leads to more robust and creative solutions that better resemble biological adaptability (Mouret and Clune, 2015; Cully et al., 2015).”
Modular Neural Control for Object Transportation of a Bio-inspired Hexapod Robot
Sørensen, C. T. L., & Manoonpong, P. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-43488-9_7 · Cited by 2
“Such a structure is also considered as a major advantage [12], compared to many other controllers [1], since it is able to deal with transferring and scaling issues; i.”
BackgroundMethodAn artificial immune system for fault tolerant control of an over-actuated rover
Kidd, R. B., & Crane, C. (2016) · International Conference on Control, Automation and Systems (ICCAS) · doi:10.1109/iccas.2016.7832349 · Cited by 2
“For a simple hexapod robot, researchers automating the generation of different gaits required 20 million iterations over 2 weeks to generate 13 thousand different fault models [2].”
MethodTaylor, T., & Dorin, A. (2018) · Monash University Research Portal (Monash University) · doi:10.48550/arxiv.1806.01322 · Cited by 2
“The idea of self-repairing machines is present in the work of Eliot, Forster, Campbell [ The Machine ] and Bernal, and this is indeed a theme in current evolutionary robotics research, e.g. (Bongard et al., 2006), (Cully et al., 2015).”
BackgroundA Generic Decentralized Gait Generator Architecture for Statically Stable Motion of Crawling Robots
Malik, A. (2019) · IEEE International Conference on Robotic Computing (IRC) · doi:10.1109/irc.2019.00124 · Cited by 2
“Some have resorted to more modern techniques involving deep learning [10], [11], reinforcement learning [12] or a combination of both [13] to yield amazing results.”
BackgroundComparing reliability of grid-based quality-diversity algorithms using artificial landscapes
Cazenille, L. (2019) · Genetic and Evolutionary Computation Conference Companion (GECCO Companion) · doi:10.1145/3319619.3321895 · Cited by 3
“These algorithms are particularly successful in evolutionary robotics problems [3], [5], [6], either by improving diversity to overcome deceptive search spaces [7], or by identifying and exploiting the generated repertoire of solutions [3].”
BackgroundProbabilistic Inferences on Quadruped Robots: An Experimental Comparison
Zhu, J., Rong, C., & Rosendo, A. (2019) · IEEE International Conference on Robotics and Biomimetics (ROBIO) · doi:10.1109/robio49542.2019.8961870 · Cited by 2
“BO and CMA-ES were previously used to solve blackbox optimizations [5], [14], [15], [16], [17], and these two methods rely on a probabilistic approach to find an optimal solution.”
MethodSurvivable Hyper-Redundant Robotic Arm with Bayesian Policy Morphing
Raza, S. J. A., Dastider, A., & Lin, M. (2020) · IEEE International Conference on Automation Science and Engineering (CASE) · doi:10.1109/case48305.2020.9216963 · Cited by 2
“The key idea of our baseline study can be described as an incremental control policy learning termed as selfmodelling in [23], which is actually a Gaussian distribution of its MDP learnt offline.”
BackgroundMethodInfluentialReinforcement Learning Applied to Hexapod Robot Locomotion: An Overview
Jorge, E., Gorrostieta-Hurtado, E., Emilio, V.-S., & Ramos-Arreguín, J. M. (2020) · Communications in computer and information science · doi:10.1007/978-3-030-62554-2_14 · Cited by 2
Xu, W., Liu, S., Kong, W., Wang, C., Deng, H., & Mei, T. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3062814 · Cited by 2
“Bayesian optimization [28], Online learning control [29] and model based control [30] techniques may be incorporated.”
BackgroundAn Analysis of Phenotypic Diversity in Multi-Solution Optimization
Hagg, A., Preuß, M., Wessing, S., Asteroth, A., & Bäck, T. (2020) · Publication Server of Bonn-Rhein-Sieg University of Applied Sciences (Bonn-Rhein-Sieg University of Applied Sciences) · Cited by 2
Objective Sampling Strategies for Generalized Locomotion Behavior with Lexicase Selection
Moore, J. M., & Stanton, A. (2021) · Conference on Artificial Life (ALIFE) · doi:10.1162/isal_a_00398 · Cited by 2
“…encompass: the ability to learn and react to environmental information across multiple unique environments (Lehman et al., 2013), the capability to adapt and reconfigure due to damage (Kriegman et al., 2019), and the expression of multiple locomotive behaviors in one robot (Cully et al., 2015).”
BackgroundOpen-Ended Search for Environments and Adapted Agents Using MAP-Elites
Norstein, E. S., Ellefsen, K. O., & Glette, K. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-02462-7_41 · Cited by 2
“In this case it would be interesting to either design the behaviour dimensions of the map so that they correlate with the skills necessary to solve the environments, or use methods for adapting through trial and error [9].”
Dynamics-aware novelty search with behavior repulsion
Xu, K., Ma, Y., & Li, W. (2022) · Genetic and Evolutionary Computation Conference (GECCO) · doi:10.1145/3512290.3528761 · Cited by 2
“Different local optima in the policy space can correspond to substantially different strategies, and discovering a diverse set of policies can be critical for efficient adaptation [20] and tasks with several deceptive local optima [21].”
BackgroundMethodAdaptation of Walking-Patterns to Damage for the Six-Legged Walking Robot LAURON V
Plasberg, C., Roennau, A., & Dillmann, R. (2022) · IEEE Transactions on Medical Robotics and Bionics · doi:10.1109/tmrb.2022.3185425 · Cited by 2
An Efficient Evaluation Mechanism for Evolutionary Reinforcement Learning
Wu, X., Zhu, Q., Lin, Q., Li, J., Chen, J., & Ming, Z. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-13870-6_4 · Cited by 2
Bongard, J. (2023) · Artificial Life and Robotics · doi:10.1007/s10015-023-00872-0 · Cited by 3
“, [6]) that likewise adopted injury as a proxy for studying how intelligent agents can grapple with morphological and control changes.”
BackgroundIlluminating the property space in crystal structure prediction using Quality-Diversity algorithms
Wolinska, M., Walsh, A., & Cully, A. (2024) · Digital Discovery · doi:10.1039/d4dd00054d · Cited by 4
Self-citationAlternating Optimisation and Quadrature for Robust Reinforcement Learning.Preprint
Paul, S., Ciosek, K., Osborne, M. A., & Whiteson, S. (2016) · arXiv (Cornell University) · Cited by 2
Discovering Evolutionary Stepping Stones through Behavior DominationPreprint
Meyerson, E., & Miikkulainen, R. P. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1704.05554 · Cited by 2
“[6, 25].”
Emotion in Future Intelligent Machines.Preprint
Belkaid, M., & Pessoa, L. (2020) · arXiv (Cornell University) · Cited by 2
A Simple Approach to Continual Learning by Transferring Skill ParametersPreprint
Zentner, K. R., Julian, R., Puri, U., Zhang, Y., & Sukhatme, G. S. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2110.10255 · Cited by 2
“…Nagabandi et al. [2019], Chatzilygeroudis and Mouret [2018], Ha and Schmidhuber [2018], Dasari et al. [2019], Chatzilygeroudis et al. [2018], Cully et al. [2015], Kaushik et al. [2020], Merel et al. [2019], Rastogi et al. [2018], Jeong et al. [2019], and goal-conditioned RL Agrawal et al.…”
BackgroundBiologically Inspired Robot Body Models and Self-Calibration
Hoffmann, M. (2021) · Springer eBooks · doi:10.1007/978-3-642-41610-1_201-1 · Cited by 4
Exploring Game Design through Human-AI Collaboration
Alvarez, A. (2022) · doi:10.24834/isbn.9789178773084 · Cited by 8
AI-driven Automated Discovery Tools Reveal Diverse Behavioral Competencies of Biological NetworksPreprint
Etcheverry, M., Moulin-Frier, C., Oudeyer, P., & Levin, M. G. (2024) · eLife · doi:10.7554/elife.92683.2 · Cited by 3
AI-driven Automated Discovery Tools Reveal Diverse Behavioral Competencies of Biological NetworksPreprint
Etcheverry, M., Moulin-Frier, C., Oudeyer, P., & Levin, M. G. (2024) · eLife · doi:10.7554/elife.92683.3 · Cited by 3
Grillotti, L., Coiffard, L., Pang, O., Faldor, M., & Cully, A. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2508.19172 · Cited by 3
“The approach is based on the methodology introduced in Cully et al. [3].”
InfluentialSelf-citationCreative Adaptation through Learning
Cully, A. (2015) · theses.fr (ABES) · doi:10.70675/f17594caz64c7z42b0z815czd4caac624298 · Cited by 5
“This kind of procedure is customary when algorithms start from scratch (Lizotte et al., 2007; Calandra et al., 2014; Cully et al., 2015).”
BackgroundSelf-citationLocomotion Control Framework for Snake-like Robot using Deep Reinforcement Learning
Obe Olumide, O., & Ayogu Thomas, O. (2021) · International Journal of Computer Trends and Technology · doi:10.14445/22312803/ijctt-v69i7p103 · Cited by 1
Transfer Dynamics in Emergent Evolutionary Curricula
Dharna, A., Hoover, A. K., Togelius, J., & Soros, L. B. (2022) · IEEE Transactions on Games · doi:10.1109/tg.2022.3151025 · Cited by 9
“While typical methods diversify based on genotypic representations, behavioral speciation is another promising method to preserve diversity [39], [43], [45].”
BackgroundNakashima, S., Kawaharazuka, K., Nagamatsu, Y., Shinjo, K., Miki, A., Suzuki, T., Asano, Y., Kakiuchi, Y., Okada, K., & Inaba, M. (2025) · Advanced Intelligent Systems · doi:10.1002/aisy.202500040 · Cited by 1
A study of the efficacy of generative flow networks for robotics and machine fault-adaptation
Sufiyan, Z., Golestan, S., Miwa, S., Mitsuka, Y., & Zai͏̈ane, O. R. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.111904 · Cited by 1
Demir, G. (2025) · Elsevier eBooks · doi:10.1016/b978-0-443-33740-6.00005-0 · Cited by 1
Navigating the Future: The Impact of Artificial Intelligence on Adaptive and Ethical Leadership
(2025) · Personalia Pelajar · doi:10.17576/personalia.2801.2025.03 · Cited by 1
Task-oriented hybrid learning for robotic skill acquisition and control
Gu, X., Bruneau, O., & Mehdi-Souzani, C. (2025) · Procedia CIRP · doi:10.1016/j.procir.2025.08.173 · Cited by 1
Metadata
- Authors
- Cully, A., Clune, J., Tarapore, D., & Mouret, J.-B.
- Affiliations
- Centre National de la Recherche Scientifique (FR)
- University of Wyoming (US)
- Corresponding author
- Mouret, Jean-Baptiste
- Venue
- Nature
- Published
- 2015-05-27
- Volume
- 521
- Issue
- 7553
- Pages
- 503-507
- ISSN
- 0028-0836 (ISSN-L)
- 1476-4687
- OpenAlex
- W1738827650
- Cited by
- 979
- References
- 40
- Citations by year
- 2026: 51 · 2025: 76 · 2024: 88 · 2023: 90 · 2022: 100 · 2021: 115 · 2020: 129 · 2019: 112 · 2018: 78 · 2017: 78 · 2016: 48 · 2015: 13
- FWCI
- 70.47
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- www.springer.com/tdm (tdm)
- www.springer.com/tdm (vor)
- Locations
- Nature (publishedVersion)
- PubMed (publishedVersion)
- ePrints Soton (University of Southampton) (acceptedVersion)
- arXiv (Cornell University) (submittedVersion)
- HAL (submittedVersion, other-oa)
- RePEc: Research Papers in Economics (submittedVersion)
- arXiv (Cornell University)
- Funders
- Agence Nationale de la Recherche (ANR-12-JS03-0009)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Reinforcement Learning in Robotics
- Physical Sciences › Engineering › Mechanical Engineering › Modular Robots and Swarm Intelligence
- Physical Sciences › Computer Science › Artificial Intelligence › Evolutionary Algorithms and Applications
- Keywords
- compensatory behavior, trial-and-error learning, autonomous mobile robots, legged robots, robot manipulators, prior knowledge, robot robustness, search and rescue robots, disaster response
- MeSH terms
- Adaptation, Physiological, Algorithms, Animals, Artificial Intelligence, Behavior, Animal, Dogs, Extremities, Motor Skills, Robotics, Time Factors, Biomimetics
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.