Search for: All records

Abstract BackgroundThe frequency and intensity of droughts are expected to increase under global change, driven by anthropogenic climate change and water diversion. Precipitation is expected to become more episodic under climate change, with longer and warmer dry spells, although some areas might become wetter. Diversion of freshwater from lakes and rivers and groundwater pumping for irrigation of agricultural fields are lowering water availability to wild plant populations, increasing the frequency and intensity of drought. Given the importance of seasonal changes and extremes in soil moisture to influence plant reproduction, and because the majority of plants are flowering plants and most of them depend on pollinators for seed production, this review focuses on the consequences of drought on different aspects of reproduction in animal-pollinated angiosperms, emphasizing interactions among drought, flowering and pollination. ScopeVisual and olfactory traits play crucial roles in attracting pollinators. Drought-induced floral changes can influence pollinator attraction and visitation, together with pollinator networks and flowering phenology, with subsequent effects on plant reproduction. Here, we review how drought influences these different aspects of plant reproduction. We identify knowledge gaps and highlight areas that would benefit from additional research. ConclusionsVisual and olfactory traits are affected by drought, but their phenotypic responses can vary with floral sex, plant sex, population and species. Ample phenotypic plasticity to drought exists for these traits, providing an ability for a rapid response to a change in drought frequency and intensity engendered by global change. The impact of these drought-induced changes in floral traits on pollinator attraction, pollen deposition and plant reproductive success does not show a clear pattern. Drought affects the structure of plant–pollinator networks and can modify plant phenology. The impact of drought on plant reproduction is not always negative, and we need to identify plant characteristics associated with these more positive responses. 

While prior federated learning (FL) methods mainly consider client heterogeneity, we focus on the Federated Domain Generalization (DG) task, which introduces train-test heterogeneity in the FL context. Existing evaluations in this field are limited in terms of the scale of the clients and dataset diversity. Thus, we propose a Federated DG benchmark that aim to test the limits of current methods with high client heterogeneity, large numbers of clients, and diverse datasets. Towards this objective, we introduce a novel data partition method that allows us to distribute any domain dataset among few or many clients while controlling client heterogeneity. We then introduce and apply our methodology to evaluate 14 DG methods, which include centralized DG methods adapted to the FL context, FL methods that handle client heterogeneity, and methods designed specifically for Federated DG on 7 datasets. Our results suggest that, despite some progress, significant performance gaps remain in Federated DG, especially when evaluating with a large number of clients, high client heterogeneity, or more realistic datasets. Furthermore, our extendable benchmark code will be publicly released to aid in benchmarking future Federated DG approaches. 

Distribution matching can be used to learn invariant representations with applications in fairness and robustness. Most prior works resort to adversarial matching methods but the resulting minimax problems are unstable and challenging to optimize. Non-adversarial likelihood-based approaches either require model invertibility, impose constraints on the latent prior, or lack a generic framework for distribution matching. To overcome these limitations, we propose a non-adversarial VAE-based matching method that can be applied to any model pipeline. We develop a set of alignment upper bounds for distribution matching (including a noisy bound) that have VAE-like objectives but with a different perspective. We carefully compare our method to prior VAE-based matching approaches both theoretically and empirically. Finally, we demonstrate that our novel matching losses can replace adversarial losses in standard invariant representation learning pipelines without modifying the original architectures—thereby significantly broadening the applicability of non-adversarial matching methods. 

Answering counterfactual queries has important applications such as explainability, robustness, and fairness but is challenging when the causal variables are unobserved and the observations are non-linear mixtures of these latent variables, such as pixels in images. One approach is to recover the latent Structural Causal Model (SCM), which may be infeasible in practice due to requiring strong assumptions, e.g., linearity of the causal mechanisms or perfect atomic interventions. Meanwhile, more practical ML-based approaches using naive domain translation models to generate counterfactual samples lack theoretical grounding and may construct invalid counterfactuals. In this work, we strive to strike a balance between practicality and theoretical guarantees by analyzing a specific type of causal query called domain counterfactuals, which hypothesizes what a sample would have looked like if it had been generated in a different domain (or environment). We show that recovering the latent SCM is unnecessary for estimating domain counterfactuals, thereby sidestepping some of the theoretic challenges. By assuming invertibility and sparsity of intervention, we prove domain counterfactual estimation error can be bounded by a data fit term and intervention sparsity term. Building upon our theoretical results, we develop a theoretically grounded practical algorithm that simplifies the modeling process to generative model estimation under autoregressive and shared parameter constraints that enforce intervention sparsity. Finally, we show an improvement in counterfactual estimation over baseline methods through extensive simulated and image-based experiments. 

Pollinators are critical for food production and ecosystem function. Although native pollinators are thought to be declining, the evidence is limited. This first, taxonomically diverse assessment for mainland North America north of Mexico reveals that 22.6% (20.6 to 29.6%) of the 1,579 species in the best-studied vertebrate and insect pollinator groups have elevated risk of extinction. All three pollinating bat species are at risk and bees are the insect group most at risk (best estimate, 34.7% of 472 species assessed, range 30.3 to 43.0%). Substantial numbers of butterflies (19.5% of 632 species, range 19.1 to 21.0%) and moths (16.1% of 142 species, range 15.5 to 19.0%) are also at risk, with flower flies (14.7% of 295 species, range 11.5 to 32.9%), beetles (12.5% of 18 species, range 11.1 to 22.2%), and hummingbirds (0% of 17 species) more secure. At-risk pollinators are concentrated where diversity is highest, in the southwestern United States. Threats to pollinators vary geographically: climate change in the West and North, agriculture in the Great Plains, and pollution, agriculture, and urban development in the East. Woodland, shrubland/chaparral, and grassland habitats support the greatest numbers of at-risk pollinators. Strategies for improving pollinator habitat are increasingly available, and this study identifies species, habitats, and threats most in need of conservation actions at state, provincial, territorial, national, and continental levels. 

Colorado is home to an incredibly rich community of native pollinating insects that contribute to the state’s economy and enhance Coloradans’ quality of life through the irreplaceable role they play in ecosystems. The pollination services these essential insects provide are at the heart of a healthy environment, contributing to our agricultural production and food systems, and relied upon by flowering plants across the state. In turn, flowering plants support the state’s wildlife, add color to the beautiful landscapes that we all treasure, and provide the basis for healthy functioning ecosystems. Despite their central importance, however, to date, no comprehensive assessment of the health of the state’s native pollinating insects has been conducted. Recognizing this need for coordinated state-level efforts to better understand the status and health of our native pollinating insects, Senate Bill 22-199, the Native Pollinating Insects Protection Study, was passed by the State Legislature and signed into law by Governor Jared Polis in May 2022. The Colorado Department of Natural Resources subsequently commissioned this study, awarded to a collaborative team of pollinator researchers, managers, and conservationists. The study was coordinated by Colorado State University Extension, in collaboration with the Xerces Society for Invertebrate Conservation and the University of Colorado Museum of Natural History, and in cooperation with leading experts in native pollinating insect ecology, management, and conservation.