Beyond Viewpoint Generalization: What Multi-View Demonstrations Offer and How to Synthesize Them for Robot Manipulation?

Abstract

Does multi-view demonstration truly improve robot manipulation, or merely enhance cross-view robustness? We present a systematic study quantifying the performance gains, scaling behavior, and underlying mechanisms of multi-view data for robot manipulation. Controlled experiments show that, under both fixed and randomized backgrounds, multi-view demonstrations consistently improve single-view policy success and generalization. Performance varies non-monotonically with view coverage, revealing effective regimes rather than a simple "more is better" trend. Notably, multi-view data breaks the scaling limitation of single-view datasets and continues to raise performance ceilings after saturation. Mechanistic analysis shows that multi-view learning promotes manipulation-relevant visual representations, better aligns the action head with the learned feature distribution, and reduces overfitting. Motivated by the importance of multi-view data and its scarcity in large-scale robotic datasets, as well as the difficulty of collecting additional viewpoints in real world settings, we propose RoboNVS, a geometry-aware self-supervised framework that synthesizes novel-view videos from monocular inputs. The generated data consistently improves downstream policies in both simulation and real-world environments.

Figures (21)

• Figure 1: Experimental setup overview. Camera layout, example multi-view observations, and the Clean/Random environment modes.
• Figure 2: Comparison between single-view and multi-view training under clean and random settings. Multi-view training improves performance across tasks, with stronger gains under the random setting.
• Figure 3: Scaling behavior under increasing viewpoint diversity. Both VA and VLA exhibit performance gains under moderate multi-view augmentation, while excessive view expansion does not yield monotonic improvement.
• Figure 4: Scaling behavior of multi-view demonstrations under fixed 0$^\circ$ evaluation. Each subplot shows success rate versus demonstrations per added view ($10\!\rightarrow\!50$) for one task. The black dashed line indicates the single-view baseline trained on origin view.
• Figure 5: Scaling laws of single-view vs. multi-view training on two manipulation tasks. Multi-view data continues to improve performance when single-view training saturates, achieving comparable or higher performance with substantially less scene diversity.