Explore the vast range of titles available.

When exploited in remote sensing analysis, a reliable change rule with transfer ability can detect changes accurately and be applied widely. However, in practice, the complexity of land cover changes makes it difficult to use only one change rule or change feature learned from a given multi-temporal dataset to detect any other new target images without applying other learning processes. In this study, we consider the design of an efficient change rule having transferability to detect both binary and multi-class changes. The proposed method relies on an improved Long Short-Term Memory (LSTM) model to acquire and record the change information of long-term sequence remote sensing data. In particular, a core memory cell is utilized to learn the change rule from the information concerning binary changes or multi-class changes. Three gates are utilized to control the input, output and update of the LSTM model for optimization. In addition, the learned rule can be applied to detect changes and transfer the change rule from one learned image to another new target multi-temporal image. In this study, binary experiments, transfer experiments and multi-class change experiments are exploited to demonstrate the superiority of our method. Three contributions of this work can be summarized as follows: (1) the proposed method can learn an effective change rule to provide reliable change information for multi-temporal images; (2) the learned change rule has good transferability for detecting changes in new target images without any extra learning process, and the new target images should have a multi-spectral distribution similar to that of the training images; and (3) to the authors’ best knowledge, this is the first time that deep learning in recurrent neural networks is exploited for change detection. In addition, under the framework of the proposed method, changes can be detected under both binary detection and multi-class change detection.

Deep learning has triggered the current rise of artificial intelligence and is the workhorse of today’s machine intelligence. Numerous success stories have rapidly spread all over science, industry and society, but its limitations have only recently come into focus. In this Perspective we seek to distil how many of deep learning’s failures can be seen as different symptoms of the same underlying problem: shortcut learning. Shortcuts are decision rules that perform well on standard benchmarks but fail to transfer to more challenging testing conditions, such as real-world scenarios. Related issues are known in comparative psychology, education and linguistics, suggesting that shortcut learning may be a common characteristic of learning systems, biological and artificial alike. Based on these observations, we develop a set of recommendations for model interpretation and benchmarking, highlighting recent advances in machine learning to improve robustness and transferability from the lab to real-world applications. Deep learning has resulted in impressive achievements, but under what circumstances does it fail, and why? The authors propose that its failures are a consequence of shortcut learning, a common characteristic across biological and artificial systems in which strategies that appear to have solved a problem fail unexpectedly under different circumstances.

Coastal habitats, such as seagrasses, mangroves, rocky and coral reefs, salt marshes, and kelp forests, sustain many key fish and invertebrate populations around the globe. Our understanding of how animals use these broadly defined habitat types is typically derived from a few well-studied regions and is often extrapolated to similar habitats elsewhere. As a result, a working understanding of their habitat importance is often based on information derived from other regions and environmental contexts. Contexts such as tidal range, rainfall, and local geomorphology may fundamentally alter animal–habitat relationships, and there is growing evidence that broadly defined habitat types such as “mangroves” or “salt marsh” may show predictable spatial and temporal variation in habitat function in relation to these environmental drivers. In the present article, we develop a framework for systematically examining contextual predictability to define the geographic transferability of animal–habitat relationships, to guide ongoing research, conservation, and management actions in these systems.

Aim: To synthesize the species distribution modelling (SDM) literature to inform which variables have been used in MaxEnt models for different taxa and to quantify how frequently they have been important for species' distributions. Location: Global. Methods: We conducted a quantitative synthesis analysing the contribution of over 400 distinct environmental variables to 2040 MaxEnt SDMs for nearly 1900 species representing over 300 families. Environmental variables were grouped into 24 related factors and results were analysed by examining the frequency with which variables were found to be most important, the mean contribution of each variable (at various taxonomic levels), and using TrueSkill™, a Bayesian skill rating system. Results: Precipitation, temperature, bathymetry, distance to water and habitat patch characteristics were the most important variables overall. Precipitation and temperature were analysed most frequently and one of these variables was often the most important predictor in the model (nearly 80% of models, when tested). Notably, distance to water was the most important variable in the highest proportion of models in which it was tested (42% of 225 models). For terrestrial species, precipitation, temperature and distance to water had the highest overall contributions, whereas for aquatic species, bathymetry, precipitation and temperature were most important. Main conclusions: Over all MaxEnt models published, the ability to discriminate occurrence from reference sites was high (average AUC = 0.92). Much of this discriminatory ability was due to temperature and precipitation variables. Further, variability (temperature) and extremes (minimum precipitation) were the most predictive. More generally, the most commonly tested variables were not always the most predictive, with, for instance, 'distance to water' infrequently tested, but found to be very important when it was. Thus, the results from this study summarize the MaxEnt SDM literature, and can aid in variable selection by identifying underutilized, but potentially important variables, which could be incorporated in future modelling efforts.

Deep neural networks have been shown to produce incorrect predictions when imperceptible perturbations are introduced into the clean input. This phenomenon has garnered significant attention and extensive research in 2D images. However, related work on point clouds is still in its infancy. Current methods suffer from issues such as generated point outliers and poor attack generalization. Consequently, it is not feasible to rely solely on overall or geometry-aware attacks to generate adversarial samples. In this paper, we integrate adversarial transfer networks with the geometry-aware method to introduce adversarial loss into the attack target. A state-of-the-art autoencoder is employed, and sensitivity maps are utilized. We use the autoencoder to generate a sufficiently deceptive mask that covers the original input, adjusting the critical subset through a geometry-aware trick to distort the point cloud gradient. Our proposed approach is quantitatively evaluated in terms of the attack success rate (ASR), imperceptibility, and transferability. Compared to other baselines on ModelNet40, our method demonstrates an approximately 38% improvement in ASR for black-box transferability query attacks, with an average query count of around 7.84. Comprehensive experimental results confirm the superiority of our method.

Background Numerous frameworks for supporting, evaluating and reporting patient and public involvement in research exist. The literature is diverse and theoretically heterogeneous. Objectives To identify and synthesize published frameworks, consider whether and how these have been used, and apply design principles to improve usability. Search strategy Keyword search of six databases; hand search of eight journals; ancestry and snowball search; requests to experts. Inclusion criteria Published, systematic approaches (frameworks) designed to support, evaluate or report on patient or public involvement in health‐related research. Data extraction and synthesis Data were extracted on provenance; collaborators and sponsors; theoretical basis; lay input; intended user(s) and use(s); topics covered; examples of use; critiques; and updates. We used the Canadian Centre for Excellence on Partnerships with Patients and Public (CEPPP) evaluation tool and hermeneutic methodology to grade and synthesize the frameworks. In five co‐design workshops, we tested evidence‐based resources based on the review findings. Results Our final data set consisted of 65 frameworks, most of which scored highly on the CEPPP tool. They had different provenances, intended purposes, strengths and limitations. We grouped them into five categories: power‐focused; priority‐setting; study‐focused; report‐focused; and partnership‐focused. Frameworks were used mainly by the groups who developed them. The empirical component of our study generated a structured format and evidence‐based facilitator notes for a “build your own framework” co‐design workshop. Conclusion The plethora of frameworks combined with evidence of limited transferability suggests that a single, off‐the‐shelf framework may be less useful than a menu of evidence‐based resources which stakeholders can use to co‐design their own frameworks.

Species distribution models (SDMs) are basic tools in ecology, biogeography, and biodiversity. The usefulness of SDMs has expanded beyond the realm of ecological sciences, and their application in other research areas is currently frequent, e.g., spatial epidemiology. In any research area, the principal interest in these models resides in their capacity to predict species response in new scenarios, i.e., the models’ transferability. Although the transferability of SDMs has been the subject of interest for many years, only in the 2000s did this topic gain particular attention. This article reviews the concept of the transferability of SDMs to new spatial scenarios, temporal periods, and (or) spatial resolutions, along with the potential constraints of the model’s transferability, and more specifically: (i) the type of predictors and multicollinearity, (ii) the model complexity, and (iii) the species’ intrinsic traits. Finally, we describe a practicable analytical protocol to be assessed before transferring a model to a new scenario. This protocol is based on three fundamental pillars: the environmental equilibrium of the species with the environment, the environmental similarity between the new scenario, and the areas used to model parametrisation and the correlation structure among predictors.

Dignity is a fundamental value in human life, in human social functioning, as well as in the legal order. It is inalienable, inherent and inviolable. In Poland, human dignity is guaranteed by Article 30 of the Constitution of the Republic of Poland. The indicated regulation determines the position of a person in the legal system, grants him subjectivity and takes away the right of state power to determine his status. Dignity is possessed by every human being, regardless of age, gender, race, religion. It is also possessed by children. The Polish Constitution ensures the protection of children’s rights while safeguarding their dignity, which should not be abused or restricted by others. It should be a fundamental value during childhood. Granting subjective status to children means that those responsible for the actual realization of their rights are limited in their actions, while at the same time they are obliged to respect them.

Background The effectiveness of complex interventions, as well as their success in reaching relevant populations, is critically influenced by their implementation in a given context. Current conceptual frameworks often fail to address context and implementation in an integrated way and, where addressed, they tend to focus on organisational context and are mostly concerned with specific health fields. Our objective was to develop a framework to facilitate the structured and comprehensive conceptualisation and assessment of context and implementation of complex interventions. Methods The Context and Implementation of Complex Interventions (CICI) framework was developed in an iterative manner and underwent extensive application. An initial framework based on a scoping review was tested in rapid assessments, revealing inconsistencies with respect to the underlying concepts. Thus, pragmatic utility concept analysis was undertaken to advance the concepts of context and implementation. Based on these findings, the framework was revised and applied in several systematic reviews, one health technology assessment (HTA) and one applicability assessment of very different complex interventions. Lessons learnt from these applications and from peer review were incorporated, resulting in the CICI framework. Results The CICI framework comprises three dimensions—context, implementation and setting—which interact with one another and with the intervention dimension. Context comprises seven domains (i.e., geographical, epidemiological, socio-cultural, socio-economic, ethical, legal, political); implementation consists of five domains (i.e., implementation theory, process, strategies, agents and outcomes); setting refers to the specific physical location, in which the intervention is put into practise. The intervention and the way it is implemented in a given setting and context can occur on a micro, meso and macro level. Tools to operationalise the framework comprise a checklist, data extraction tools for qualitative and quantitative reviews and a consultation guide for applicability assessments. Conclusions The CICI framework addresses and graphically presents context, implementation and setting in an integrated way. It aims at simplifying and structuring complexity in order to advance our understanding of whether and how interventions work. The framework can be applied in systematic reviews and HTA as well as primary research and facilitate communication among teams of researchers and with various stakeholders.

Ecological forecasting provides a powerful set of methods for predicting short‐ and long‐term change in living systems. Forecasts are now widely produced, enabling proactive management for many applied ecological problems. However, despite numerous calls for an increased emphasis on prediction in ecology, the potential for forecasting to accelerate ecological theory development remains underrealized. Here, we provide a conceptual framework describing how ecological forecasts can energize and advance ecological theory. We emphasize the many opportunities for future progress in this area through increased forecast development, comparison and synthesis. Our framework describes how a forecasting approach can shed new light on existing ecological theories while also allowing researchers to address novel questions. Through rigorous and repeated testing of hypotheses, forecasting can help to refine theories and understand their generality across systems. Meanwhile, synthesizing across forecasts allows for the development of novel theory about the relative predictability of ecological variables across forecast horizons and scales. We envision a future where forecasting is integrated as part of the toolset used in fundamental ecology. By outlining the relevance of forecasting methods to ecological theory, we aim to decrease barriers to entry and broaden the community of researchers using forecasting for fundamental ecological insight.