Research map / Graph learning / GNN architectures
My Body is a Cage: the Role of Morphology in Graph-Based Incompatible Control
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use Amorpheus instead of SMP | zero-shot generalisation · walker-3-main | average return: 666.24 (133.66) vs 175.65 (157.38) | Helped |
| Use Amorpheus instead of SMP | zero-shot generalisation · walker-6-main | average return: 1171.35 (832.91) vs 729.26 (135.60) | Helped |
| Use Amorpheus instead of SMP | zero-shot generalisation · humanoid-2d-7-left-leg | average return: 2821.22 (1340.29) vs 2158.29 (785.33) | Helped |
| Use Amorpheus instead of SMP | zero-shot generalisation · humanoid-2d-8-right-knee | average return: 2717.21 (624.80 ) vs 327.93 (125.75) | Helped |
| Use Amorpheus instead of SMP | zero-shot generalisation · cheetah-3-balanced | average return: 474.82 (74.05) vs 156.16 (33.00) | Helped |
| Use Amorpheus instead of SMP | zero-shot generalisation · cheetah-5-back | average return: 3417.72 (306.84) vs 3820.77 (301.95) | Hurt |
| Use Amorpheus instead of SMP | zero-shot generalisation · cheetah-6-front | average return: 5081.71 (391.08) vs 6019.07 (506.55) | Hurt |
| Replace morphology topology with star topology (ablation) | SMP · SMP continuous-control benchmarks | average return | No clear effect |
| Replace morphology topology with line topology (ablation) | SMP · SMP continuous-control benchmarks | average return | Hurt |
| Replace morphology topology with star topology (ablation) | NerveNet · Walkers | average return | No clear effect |
| Replace morphology topology with fully connected topology (ablation) | NerveNet · Walkers | learning speed | Hurt |
| Replace morphology topology with line topology (ablation) | NerveNet · Walkers | average return | No clear effect |
Closest papers by meaning
- Universal Morphology Control via Contextual Modulation (ICML 2023 · similarity 0.89)
- A Transfer Approach Using Graph Neural Networks in Deep Reinforcement Learning (AAAI 2024 · similarity 0.85)
- $\texttt{G1}$: Teaching LLMs to Reason on Graphs with Reinforcement Learning (NeurIPS 2025 · similarity 0.84)
- Graph Agent Network: Empowering Nodes with Inference Capabilities for Adversarial Resilience (AAAI 2025 · similarity 0.84)
- Can Graph Learning Improve Planning in LLM-based Agents? (NeurIPS 2024 · similarity 0.84)
- LLM Enhancers for GNNs: An Analysis from the Perspective of Causal Mechanism Identification (ICML 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.