Search for: All records

Humans excel at efficiently navigating through crowds without collision by focusing on specific visual regions relevant to navigation. However, most robotic visual navigation methods rely on deep learning models pre-trained on vision tasks, which prioritize salient objects—not necessarily relevant to navigation and potentially misleading. Alternative approaches train specialized navigation models from scratch, requiring significant computation. On the other hand, self-supervised learning has revolutionized computer vision and natural language processing, but its application to robotic navigation remains underexplored due to the difficulty of defining effective self-supervision signals. Motivated by these observations, in this work, we propose a Self-Supervised Vision-Action Model for Visual Navigation Pre-Training (VANP). Instead of detecting salient objects that are beneficial for tasks such as classification or detection, VANP learns to focus only on specific visual regions that are relevant to the navigation task. To achieve this, VANP uses a history of visual observations, future actions, and a goal image for self-supervision, and embeds them using two small Transformer Encoders. Then, VANP maximizes the information between the embeddings by using a mutual information maximization objective function. We demonstrate that most VANP-extracted features match with human navigation intuition. VANP achieves comparable performance as models learned end-to-end with half the training time and models trained on a large-scale, fully supervised dataset, i.e., ImageNet, with only 0.08% data. 

Abstract In adverse environments, the number of fertilizable female gametophytes (FGs) in plants is reduced, leading to increased survival of the remaining offspring. How the maternal plant perceives internal growth cues and external stress conditions to alter FG development remains largely unknown. We report that homeostasis of the stress signaling molecule nitric oxide (NO) plays a key role in controlling FG development under both optimal and stress conditions. NO homeostasis is precisely regulated by S-nitrosoglutathione reductase (GSNOR). Prior to fertilization, GSNOR protein is exclusively accumulated in sporophytic tissues and indirectly controls FG development in Arabidopsis (Arabidopsis thaliana). In GSNOR null mutants, NO species accumulated in the degenerating sporophytic nucellus, and auxin efflux into the developing FG was restricted, which inhibited FG development, resulting in reduced fertility. Importantly, restoring GSNOR expression in maternal, but not gametophytic tissues, or increasing auxin efflux substrate significantly increased the proportion of normal FGs and fertility. Furthermore, GSNOR overexpression or added auxin efflux substrate increased fertility under drought and salt stress. These data indicate that NO homeostasis is critical to normal auxin transport and maternal control of FG development, which in turn determine seed yield. Understanding this aspect of fertility control could contribute to mediating yield loss under adverse conditions.