Self-Supervised Interpretable Sensorimotor Learning via Latent Functional Modularity

arXiv:2403.18947v1 Announce Type: new
Abstract: We introduce MoNet, a novel method that combines end-to-end learning with modular network architectures for self-supervised and interpretable sensorimotor learning. MoNet is composed of three functionally distinct neural modules: Perception, Planning, and Control. Leveraging its inherent modularity through a cognition-guided contrastive loss function, MoNet efficiently learns task-specific decision-making processes in latent space, without requiring task-level supervision. Moreover, our method incorporates an online post-hoc explainability approach, which enhances the interpretability of the end-to-end inferences without a …

abstract architectures arxiv cognition control cs.lg cs.ro function functional loss modular modules network novel perception planning through type via