REVIEW 4 major objections 7 minor 35 references
Multi-Task Multi-Agent Reinforcement Learning via Skill Graphs
T0 review · 4 major / 7 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read By embedding environments, tasks, and skills in a shared vector space and scoring every skill against a query, the paper shows that one hierarchical controller can handle adversarial and cooperative tasks in the same library and…
desk verdict A plausible modular skill-graph system for unrelated multi-agent tasks, but the headline comparison against MAPPO rests on a metric artifact (ρ_succ=100% by definition). read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The skill graph is a knowledge graph whose entities are environments, tasks, and skills and whose relations are "environment to skill" and "task to skill". TransH embeds each relation as a hyperplane with a translation vector and scores a triple by $\exp(-\lambda \| (h - w_r^\top h w_r) + d_r - (b - w_r^\top b w_r)\|)$, supporting one-to-many and many-to-many relations. A query is scored by multiplying the task-skill and environment-skill scores; the score tier decides direct selection, weighted combination, or further training. The construction is independent of the low-level policies, which is what gives the high-level module its generalization and lets unrelated tasks share one library.
What would settle it
Take a pair of queries with identical feature vectors but different true dynamics, for example two tasks that share the reported boundary type, speed limits, attack radius, and perception range but differ in robot mass or in an unlisted environmental factor. If the skill graph still rates an old skill near 1 and the execution fails, the hand-built features are not sufficient to make embedding distance track real skill suitability, and the ranking mechanism would not transfer to unseen queries.
Extended reading notes
Core claim
The central discovery is that a knowledge-graph embedding of skills turns multi-task multi-agent learning into a retrieval-and-refinement problem. The paper builds a skill graph with environment, task, and skill entities and relations $e\to s$ and $t\to s$, learns their representations with TransH under a loss that mixes positive, negative, and soft samples, and then answers a query $(e_{\mathrm{new}}, t_{\mathrm{new}})$ by computing $S = S(t, r_{t\to s}, s) \cdot S(e, r_{e\to s}, s)$ for every skill. Depending on where the score falls relative to thresholds $\alpha_{\mathrm{high}}$ and $\alpha_{\mathrm{low}}$, the controller chooses the top skill, combines the actions of top skills with normalized weights, or further trains the top skill with RL. This mechanism, with low-level policies trained by a local-critic version of MADDPG, is what lets adversarial and cooperative tasks live in the same skill library and lets new but similar tasks be solved faster than training from scratch.
Load-bearing premise
The whole ranking rests on the assumption that the hand-built feature vectors for environments and tasks make embedding-space distance track which skill will actually work.
Editorial extensions
If this is right
- A team trained once on a library of flocking and combat skills can be reused for a new mission that mixes both; the skill graph supplies the right skill for each stage without retraining the high level.
- When a new task resembles two stored skills, the robot action is a normalized weighted sum of those skills' actions, giving a cheap adaptation path that does not require new reinforcement learning.
- When the new task falls outside the stored distribution, warm-starting from the top-ranked skill reaches a working policy in fewer episodes than training from scratch, as shown in the reward curves.
- Because the graph's construction is independent of the low-level policies, changes to low-level parameters such as initial counts or leader speed do not degrade high-level decision success, whereas the hierarchical MAPPO baseline's success rate drops as the skill count grows and parameters vary.
- The high-level decision layer never needs centralized observations; the low-level local-critic MADDPG handles 50-agent tasks, so the method is compatible with distributed robot swarms.
Reading between the lines
- Editorial extension: the score margin between the top-ranked and second-ranked skills could be used as a confidence signal for online mode switching, a use the paper does not explore.
- Editorial extension: the combination step assumes that blending actions by normalized weights is safe; a natural stress test is blending an attack skill with a flocking skill and checking whether the mixed behavior preserves both objectives.
- Editorial extension: the paper's own conclusion notes that totally unseen scenarios yield low scores and require new skills; this suggests the graph should be able to grow by inserting newly trained skills, turning one-shot adaptation into a lifelong learning process.
- Editorial extension: because the ranking depends on hand-crafted features, a testable improvement is to learn task and environment features end-to-end from observations instead of specifying vectors like $(y, L)$ and $(v_{\max}, v_{\min}, \Delta h, n_o, r_{\mathrm{att}})$ by hand.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes a hierarchical multi-task multi-agent reinforcement learning (MT-MARL) architecture in which a knowledge-graph-style "skill graph" serves as the high-level module and a local-critic MADDPG variant serves as the low-level module. Environments, tasks, and skills are represented as entities in a graph embedded with TransH; given a query environment and task, the graph scores all skills and selects, combines, or further trains the highest-scoring ones. The authors claim the method handles unrelated tasks (adversarial and cooperative), transfers knowledge better than standard hierarchical RL, and outperforms hierarchical MAPPO baselines. The paper reports simulation experiments with 50-agent teams and a qualitative six-robot real-world demonstration, and it makes code and videos publicly available.
Significance. If the comparative claims were fully substantiated, the paper would be a useful contribution: it addresses the under-studied setting of unrelated tasks in MT-MARL, provides a modular graph-based selection mechanism, and includes both simulation and real-robot demonstrations. The availability of code and videos is a clear strength, as is the explicit acknowledgment in Section VII that entirely unseen scenarios require learning from scratch. However, the central comparative claim against hierarchical MAPPO is not currently supported by the reported evidence. The skill graph's ρ_succ=100% is asserted on the basis of its definition rather than measured under the same success criterion used for the baseline, and the baseline receives less informative observations than the skill graph. These issues are load-bearing because the abstract and introduction cite "outperforms the latest hierarchical MAPPO" as the main empirical result. The underlying idea remains plausible, and the missing evidence is in principle obtainable, so the paper is best treated as requiring major revision rather than rejection.
major comments (4)
- [V-C.2] The claim that the skill graph achieves ρ_succ=100% is definitional, not empirical. The text states that the skill graph selects skills based on their scores without requiring a perfect score (i.e., 1), resulting in ρ_succ=100%; under that definition, any selector that always returns a highest-scoring skill would also have 100% success. The ρ_succ reported for high-level MAPPO in Table II is a genuine accuracy measure: a decision is successful only if the selected skill matches the expected one, with a -1 reward for mismatch. Unless the skill graph is evaluated under the same matching-based criterion on the same set of test queries, the paper's headline claim that the proposed method outperforms hierarchical MAPPO is unsupported. Please report the skill graph's ρ_succ under the identical success definition, including variance over seeds, and define the expected skill for each query in the combination and further-training regimes.
- [V-C.1 and V-B.2] The comparison between Scheme 1 and Scheme 2 is confounded by input information. The high-level MAPPO observes only o_g = [x_rg, n_g, n_r] (relative team-center position and team sizes), whereas the skill graph receives the full ground-truth query: the environment features (y, L) and the task features (v_max, v_min, Δh, n_o, r_att) or (v_max, v_min, d_ref, r_perc). The skill graph therefore has privileged information about the exact task parameters, while the baseline must infer the task from trajectories. The two frameworks are also not controlled at the low level: local-MADDPG versus MAPPO can change the distribution of states observed by the high level. These differences alone could explain the reported gap. Please either provide the same task-feature inputs to the high-level MAPPO baseline, or evaluate the skill graph from partial observations, and ideally match the low-level algorithm.
- [IV-E and Eq. (1)] The skill graph's generalization behavior rests on the assumption that hand-crafted feature vectors and the weighted similarity δ in Eq. (1) make embedding-space distance correspond to skill suitability, but no sensitivity analysis is provided for the chosen weights k_j or the thresholds α_high and α_low. Moreover, for queries that exactly match a training triple (e.g., stage 1 query (1,6)+(1,0,0.4,3) matching 'floc 3 fixed'), a score near 1 is directly enforced by the positive-sample term (S_posi - 1)^2 in Eq. (1), so the stage-1 and stage-2 results in Fig. 5(d)-(e) demonstrate retrieval of memorized samples rather than predictive generalization. The d_ref=0.6 combination case in Fig. 5(f) is the only genuinely interpolated query presented, and it is not evaluated with a quantitative success metric. Please add held-out queries and report how often the graph's top selection matches the independently verified best skill.
- [V-B.2 and VI] Several effectiveness claims are supported only by single illustrative trials. Fig. 5(h) reports that further training of the highest-scoring skill requires less training time and better sampling efficiency than training from scratch, but no reward curves with numerical values, no number of seeds, and no definition of the convergence criterion are given. Similarly, the real-world experiment in Section VI is described through snapshots only; there are no quantitative metrics such as task completion rate, time to convergence, or number of repeated trials. These claims should either be quantified with statistics or explicitly labeled as qualitative demonstrations.
minor comments (7)
- [IV] The section heading contains a typo: 'Feature Constructon' should be 'Feature Construction'.
- [Fig. 4 caption] The caption contains garbled text ('0.6refdd' and inconsistent d_ref formatting) and the mapping of subfigures (c)-(f) to periodic versus fixed boundaries is unclear; please clean up the caption and ensure the labels match the images.
- [Table I] The table header repeats 'Adve/Floc' in a way that makes it difficult to tell which hyperparameters belong to local-MADDPG and which to MAPPO; please reformat with clearer column groupings.
- [V-C.1] The text says 'dim' indicates the number of low-level skills, but Table II rows range from dim=2 to dim=12; please clarify how these skill sets were constructed and how the 'expected' skill for the -1 reward is determined in each case.
- [Eq. (1) and V-B.1] The weight notation k_j is not fully defined: the text gives values such as k_1,t=0, k_2,t=0, ..., but does not explain the mapping from j to the attributes of the feature vectors, nor how k_j values for environment features interact with those for task features.
- [III-B] The thresholds α_high and α_low are used in the utilization procedure but are only assigned concrete values in Section V-B.1; please either define them in Section III-B or state that they are application-specific hyperparameters.
- [Related work] The skill graph idea is credited to reference [24] (RSG), but the paper does not discuss how the present construction differs from RSG; a brief comparison would help position the contribution.
Circularity Check
The headline comparison rests on a definitionally perfect success rate, and two 'inference' examples replay training samples; no load-bearing self-citation chain is involved.
-
self definitional
[Section V-C.2, 'Results analysis' (comparison of skill graph vs. high-level MAPPO)]
"Second, the skill graph selects skills based on their scores without requiring a perfect score, (i.e. 1), resulting in ρsucc = 100%."
In V-C.1, ρsucc is defined as nsucc/ntotal, where nsucc counts 'successful decisions.' For high-level MAPPO, success is operationalized as matching the expected skill, which makes Table II a genuine accuracy measure (about 88-95%). For the skill graph, no comparable count is reported: the 100% figure follows from the definition that any selection of a highest-scoring skill is successful. Under this definition, any argmax selector—even a random one that always emits some skill—would achieve ρsucc = 100%, so the metric cannot distinguish the skill graph from a trivial baseline.
-
fitted input called prediction
[Section V-B.2, 'Results analysis' (skill graph application, stages 1 and 2)]
"The input environment + task query is (1,6) + (1,0,0.4,3), and the output (see Fig. 5(d)) shows that the skill “floc 3 fixed” achieves the highest score ( 0.97), exceeding the threshold αhigh = 0.95, and is therefore selected for stage 1."
The query (1,6)+(1,0,0.4,3) is exactly the positive training triple for the skill 'floc 3 fixed': all eight flocking tasks and two environments listed in V-B.1 produce 32 skills, and III-A says factual triples constructed from known environment-task-skill matchings serve as positive samples. The loss in Eq. (1) contains (Sposi − 1)^2, so the training objective explicitly drives this triple's score toward 1. The reported 0.97 is therefore a fitted value for a training sample, not an independent prediction or inference. The same holds for stage 2, whose query matches 'adve 2 fixed' exactly.
full rationale
The paper's comparative claim against hierarchical MAPPO is not supported by a measurement: ρsucc is defined in V-C.1 as nsucc/ntotal, with MAPPO's success judged by matching the expected skill, but the skill graph's ρsucc = 100% in V-C.2 is asserted because the graph always returns a highest-scoring skill. That makes the success metric trivially satisfied and voids the stated outperformance comparison. Separately, the stage-1 and stage-2 demonstrations of 'selection' use exact positive training triples, so their near-1 scores are consequences of the supervised loss in Eq. (1), not predictions. The stage-3 weighted-combination example and the real-world experiment use queries outside the training distribution and constitute genuinely non-circular evidence for the method's transfer capability; these prevent the entire paper from being vacuous. No load-bearing self-citation circularity was found: the skill graph concept in [24] is cited as prior work but is not used here to forbid alternatives or as a uniqueness theorem, and the TransH construction and empirical demonstrations are implemented in this paper. The central definitional metric issue, however, means the headline comparison reduces by construction.
Assumptions & free parameters
free parameters (4)
- α_high =
0.95
- α_low =
0.85
- λ =
3
- k_j weights for soft-sample similarity δ =
k1,t..k5,t = 0,0,0,3,1; k1,e=0.95, k2,e=0.05
assumptions (4)
- domain assumption TransH knowledge-graph embedding can represent environment/task/skill relationships with the scoring function S.
- domain assumption The hand-crafted feature vectors (Section IV-E) are sufficient descriptors of environments and tasks for skill suitability.
- ad hoc to paper Weighted feature similarity δ reflects task/environment relatedness, so soft samples should have score ≤ 1-δ.
- domain assumption The designed reward functions for adversarial and flocking tasks induce the intended low-level skills.
Cite this review
Pith. "Pith review of Multi-Task Multi-Agent Reinforcement Learning via Skill Graphs." pith.science (2026). https://pith.science/paper/UAGSDGML
@misc{pith2026250706690,
author = {Pith},
title = {Pith review of: Multi-Task Multi-Agent Reinforcement Learning via Skill Graphs},
year = {2026},
howpublished = {\url{https://pith.science/paper/UAGSDGML}},
note = {Machine review of arXiv:2507.06690}
}
read the original abstract
Multi-task multi-agent reinforcement learning (MT-MARL) has recently gained attention for its potential to enhance MARL's adaptability across multiple tasks. However, it is challenging for existing multi-task learning methods to handle complex problems, as they are unable to handle unrelated tasks and possess limited knowledge transfer capabilities. In this paper, we propose a hierarchical approach that efficiently addresses these challenges. The high-level module utilizes a skill graph, while the low-level module employs a standard MARL algorithm. Our approach offers two contributions. First, we consider the MT-MARL problem in the context of unrelated tasks, expanding the scope of MTRL. Second, the skill graph is used as the upper layer of the standard hierarchical approach, with training independent of the lower layer, effectively handling unrelated tasks and enhancing knowledge transfer capabilities. Extensive experiments are conducted to validate these advantages and demonstrate that the proposed method outperforms the latest hierarchical MAPPO algorithms. Videos and code are available at https://github.com/WindyLab/MT-MARL-SG
Figures
Figures from the paper (3 more)
Reference graph
Works this paper leans on
-
[24]
Unlock Reliable Skill Inference for Quadruped Adaptive Behavior by Skill Graph
H. Zhang, et al., “RSG: Fast learning adaptive skills for quadruped robots by skill graph,” arXiv preprint arXiv:2311.06015 , 2023
work page Pith review arXiv 2023
-
[1]
Grandmaster level in starcraft ii using multi-agent reinforcement learning,
O. Vinyals, et al., “Grandmaster level in starcraft ii using multi-agent reinforcement learning,” nature, vol. 575, no. 7782, pp. 350–354, 2019
work page 2019
-
[2]
On reinforcement learning for full-length game of starcraft,
Z.-J. Pang, R.-Z. Liu, Z.-Y . Meng, Y . Zhang, Y . Yu, and T. Lu, “On reinforcement learning for full-length game of starcraft,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2019, pp. 4691–4698
work page 2019
-
[3]
T. Fan, P. Long, W. Liu, and J. Pan, “Distributed multi-robot collision avoidance via deep reinforcement learning for navigation in complex scenarios,” The International Journal of Robotics Research, vol. 39, no. 7, pp. 856–892, 2020
work page 2020
-
[4]
Decentralized multi-agent pursuit using deep reinforcement learning,
C. De Souza, R. Newbury, A. Cosgun, P. Castillo, B. Vidolov, and D. Kuli ´c, “Decentralized multi-agent pursuit using deep reinforcement learning,” IEEE Robotics and Automation Letters , vol. 6, no. 3, pp. 4552–4559, 2021
work page 2021
-
[5]
D. Johnson, G. Chen, and Y . Lu, “Multi-agent reinforcement learning for real-time dynamic production scheduling in a robot assembly cell,” IEEE Robotics and Automation Letters , vol. 7, no. 3, pp. 7684–7691, 2022
work page 2022
-
[6]
Discovering generalizable multi-agent coordination skills from multi-task offline data,
F. Zhang, C. Jia, Y .-C. Li, L. Yuan, Y . Yu, and Z. Zhang, “Discovering generalizable multi-agent coordination skills from multi-task offline data,” in Proceedings of the International Conference on Learning Representations, 2023, pp. 1–24
work page 2023
-
[7]
A survey on multi-task learning,
Y . Zhang and Q. Yang, “A survey on multi-task learning,” IEEE transactions on knowledge and data engineering , vol. 34, no. 12, pp. 5586–5609, 2021
2021
Show all 35 references
-
[8]
Deep decentralized multi-task multi-agent reinforcement learning under partial observability,
S. Omidshafiei, J. Pazis, C. Amato, J. P. How, and J. Vian, “Deep decentralized multi-task multi-agent reinforcement learning under partial observability,” inProceedings of the International Conference on Machine Learning, 2017, pp. 2681–2690
2017
-
[9]
Multi-task deep reinforcement learning with popart,
M. Hessel, H. Soyer, L. Espeholt, W. Czarnecki, S. Schmitt, and H. Van Hasselt, “Multi-task deep reinforcement learning with popart,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2019, pp. 3796–3803
2019
-
[10]
Conflict-averse gradient descent for multi-task learning,
B. Liu, X. Liu, X. Jin, P. Stone, and Q. Liu, “Conflict-averse gradient descent for multi-task learning,” in Proceedings of the Advances in Neural Information Processing Systems , 2021, pp. 18 878–18 890
2021
-
[11]
A deep hierarchical approach to lifelong learning in minecraft,
C. Tessler, S. Givony, T. Zahavy, D. Mankowitz, and S. Mannor, “A deep hierarchical approach to lifelong learning in minecraft,” in Proceedings of the AAAI conference on Artificial Intelligence , 2017, pp. 1553–1561
2017
-
[12]
Lifelong federated reinforcement learning: A learning architecture for navigation in cloud robotic systems,
B. Liu, L. Wang, and M. Liu, “Lifelong federated reinforcement learning: A learning architecture for navigation in cloud robotic systems,” IEEE Robotics and Automation Letters , vol. 4, no. 4, pp. 4555–4562, 2019
2019
-
[13]
Actor-mimic: Deep multitask and transfer reinforcement learning,
E. Parisotto, J. L. Ba, and R. Salakhutdinov, “Actor-mimic: Deep multitask and transfer reinforcement learning,” in Proceedings of the International Conference on Learning Representations , 2016, pp. 1–16
2016
-
[14]
Policy distillation,
A. A. Rusu, et al., “Policy distillation,” arXiv preprint arXiv:1511.06295, 2015
2015 arXiv
-
[15]
Knowledge transfer in multi- task deep reinforcement learning for continuous control,
Z. Xu, K. Wu, Z. Che, J. Tang, and J. Ye, “Knowledge transfer in multi- task deep reinforcement learning for continuous control,” in Proceedings of the Advances in Neural Information Processing Systems , 2020, pp. 15 146–15 155
2020
-
[16]
PaCo: Parameter- compositional multi-task reinforcement learning,
L. Sun, H. Zhang, W. Xu, and M. Tomizuka, “PaCo: Parameter- compositional multi-task reinforcement learning,” in Proceedings of the Advances in Neural Information Processing Systems , 2022, pp. 21 495– 21 507
2022
-
[17]
Hierarchical rein- forcement learning: A comprehensive survey,
S. Pateria, B. Subagdja, A.-h. Tan, and C. Quek, “Hierarchical rein- forcement learning: A comprehensive survey,” ACM Computing Surveys (CSUR), vol. 54, no. 5, pp. 1–35, 2021
2021
-
[18]
Planning-augmented hierarchical reinforcement learning,
R. Gieselmann and F. T. Pokorny, “Planning-augmented hierarchical reinforcement learning,” IEEE Robotics and Automation Letters , vol. 6, no. 3, pp. 5097–5104, 2021
2021
-
[19]
Learning functionally decomposed hierarchies for continuous control tasks with path planning,
S. Christen, L. Jendele, E. Aksan, and O. Hilliges, “Learning functionally decomposed hierarchies for continuous control tasks with path planning,” IEEE Robotics and Automation Letters , vol. 6, no. 2, pp. 3623–3630, 2021
2021
-
[20]
Modular multitask reinforcement learning with policy sketches,
J. Andreas, D. Klein, and S. Levine, “Modular multitask reinforcement learning with policy sketches,” in Proceedings of the International Conference on Machine Learning , 2017, pp. 166–175
2017
-
[21]
Adaptive and explainable deployment of navigation skills via hierarchical deep reinforcement learning,
K. Lee, S. Kim, and J. Choi, “Adaptive and explainable deployment of navigation skills via hierarchical deep reinforcement learning,” in Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), 2023, pp. 1673–1679
2023
-
[22]
Initial task allocation in multi-human multi-robot teams: An attention-enhanced hierarchical reinforcement learning approach,
R. Wang, D. Zhao, A. Gupte, and B.-C. Min, “Initial task allocation in multi-human multi-robot teams: An attention-enhanced hierarchical reinforcement learning approach,” IEEE Robotics and Automation Letters, vol. 9, no. 4, pp. 3451–3458, 2024
2024
-
[23]
Multi-agent reinforcement learn- ing: A selective overview of theories and algorithms,
K. Zhang, Z. Yang, and T. Ba s ¸ar, “Multi-agent reinforcement learn- ing: A selective overview of theories and algorithms,” Handbook of reinforcement learning and control , pp. 321–384, 2021
2021
-
[25]
A survey on knowledge graphs: Representation, acquisition, and applications,
S. Ji, S. Pan, E. Cambria, P. Marttinen, and S. Y . Philip, “A survey on knowledge graphs: Representation, acquisition, and applications,” IEEE Transactions on Neural Networks and Learning Systems , vol. 33, no. 2, pp. 494–514, 2021
2021
-
[26]
Knowledge graph embedding by translating on hyperplanes,
Z. Wang, J. Zhang, J. Feng, and Z. Chen, “Knowledge graph embedding by translating on hyperplanes,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2014, pp. 1112–1119
2014
-
[27]
Multi-agent actor-critic for mixed cooperative-competitive environments,
R. Lowe, Y . I. Wu, A. Tamar, J. Harb, O. Pieter Abbeel, and I. Mordatch, “Multi-agent actor-critic for mixed cooperative-competitive environments,” in Proceedings of the Advances in Neural Information Processing Systems, 2017, pp. 6379–6390
2017
-
[28]
Knowledge graph embedding: A survey of approaches and applications,
Q. Wang, Z. Mao, B. Wang, and L. Guo, “Knowledge graph embedding: A survey of approaches and applications,” IEEE Transactions on Knowledge and Data Engineering , vol. 29, no. 12, pp. 2724–2743, 2017
2017
-
[29]
A novel framework of knowledge transfer system for construction projects based on knowledge graph and transfer learning,
J. Xu, M. He, and Y . Jiang, “A novel framework of knowledge transfer system for construction projects based on knowledge graph and transfer learning,” Expert Systems with Applications , vol. 199, p. 116964, 2022
2022
-
[30]
Translating embeddings for modeling multi-relational data,
A. Bordes, N. Usunier, A. Garcia-Duran, J. Weston, and O. Yakhnenko, “Translating embeddings for modeling multi-relational data,” in Proceed- ings of the Advances in Neural Information Processing Systems , 2013, pp. 2787–2795
2013
-
[31]
Predator–prey survival pressure is sufficient to evolve swarming behaviors,
J. Li, L. Li, and S. Zhao, “Predator–prey survival pressure is sufficient to evolve swarming behaviors,” New Journal of Physics , vol. 25, no. 9, p. 092001, 2023
2023
-
[32]
Swarm robotics: Past, present, and future [point of view],
M. Dorigo, G. Theraulaz, and V . Trianni, “Swarm robotics: Past, present, and future [point of view],” Proceedings of the IEEE , vol. 109, no. 7, pp. 1152–1165, 2021
2021
-
[33]
Optimized flocking of autonomous drones in confined environments,
G. V ´as´arhelyi, C. Vir ´agh, G. Somorjai, T. Nepusz, A. E. Eiben, and T. Vicsek, “Optimized flocking of autonomous drones in confined environments,” Science Robotics, vol. 3, no. 20, p. eaat3536, 2018
2018
-
[34]
The surprising effectiveness of PPO in cooperative multi- agent games,
C. Yu, et al., “The surprising effectiveness of PPO in cooperative multi- agent games,” in Proceedings of the Advances in Neural Information Processing Systems, 2022, pp. 24 611–24 624
2022
-
[35]
Omnibot: A scalable vision-based robot swarm platform,
Z. Ma, et al., “Omnibot: A scalable vision-based robot swarm platform,” in Proceedings of the IEEE International Conference on Control & Automation (ICCA), 2024, pp. 975–980
2024
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.