Given current anthropogenic alterations to many ecosystems and communities, it is becoming increasingly important to consider whether and how organisms can cope with changing resources. Metabolic rate, because it represents the rate of energy expenditure, may play a key role in mediating the link between resource conditions and performance and thereby how well organisms can persist in the face of environmental change. Here, we focus on the role that energy metabolism plays in determining organismal responses to changes in food availability over both short-term ecological and longer-term evolutionary timescales. Using a meta-analytical approach encompassing multiple species, we find that individuals with a higher metabolic rate grow faster under high food levels but slower once food levels decline, suggesting that the association between metabolism and life-history traits shifts along resource gradients. We also find that organisms can cope with changing resource availability through both phenotypic plasticity and genetically based evolutionary adaptation in their rates of energy metabolism. However, the metabolic rates of individuals within a population and of species within a lineage do not all respond in the same manner to changes in food availability. This diversity of responses suggests that there are benefits but also costs to changes in metabolic rate. It also underscores the need to examine not just the energy budgets of organisms within the context of metabolic rate but also how energy metabolism changes alongside other physiological and behavioural traits in variable environments.

No abstract

No abstract

Despite growing interest in object detection, very few works address the extremely practical problem of crossdomain robustness especially for automative applications. In order to prevent drops in performance due to domain shift, we introduce an unsupervised domain adaptation method built on the foundation of faster-RCNN with two domain adaptation components addressing the shift at the instance and image levels respectively and apply a consistency regularization between them. We also introduce a family of adaptation layers that leverage the squeeze excitation mechanism called SE Adaptors to improve domain attention and thus improves performance without any prior requirement of knowledge of the new target domain. Finally, we incorporate a center loss in the instance and image level representations to improve the intra-class variance. We report all results with Cityscapes as our source domain and Foggy Cityscapes as the target domain outperforming previous baselines. code: https://github.com/shreyasrajesh/DA-Object-Detection

No abstract