Rice cultivation supplies half the world's population with staple food, while also being a major driver of freshwater depletion--consuming roughly a quarter of global freshwater--and accounting for ~48% of greenhouse gas emissions from croplands. In regions like Punjab, India, where groundwater levels are plummeting at 41.6 cm/year, adopting water-saving rice farming practices is critical. Direct-Seeded Rice (DSR) and Alternate Wetting and Drying (AWD) can cut irrigation water use by 20–40% without hurting yields, yet lack of spatial data on adoption impedes effective adaptation policy and climate action. We present a machine learning framework to bridge this data gap by monitoring sustainable rice farming at scale. In collaboration with agronomy experts and a large-scale farmer training program, we obtained ground-truth data from ~1,400 fields across Punjab. Leveraging this partnership, we developed a novel dimensional classification approach that decouples sowing and irrigation practices, achieving F1 scores of 0.8 and 0.74 respectively, solely employing Sentinel-1 satellite imagery. Explainability analysis reveals that DSR classification is robust while AWD classification depends primarily on planting schedule differences, as Sentinel-1's 12-day revisit frequency cannot capture the higher frequency irrigation cycles characteristic of AWD practices. Applying this model across 3 million fields reveals spatial heterogeneity in adoption at the state level, highlighting gaps and opportunities for policy targeting. Our district-level adoption rates correlate well with government estimates (Spearman's=0.69 and Rank Biased Overlap=0.77). This study provides policymakers and sustainability programs a powerful tool to track practice adoption, inform targeted interventions, and drive data-driven policies for water conservation and climate mitigation at regional scale.

Crop field boundaries are foundational datasets for agricultural monitoring and assessments but are expensive to collect manually. Machine learning (ML) methods for automatically extracting field boundaries from remotely sensed images could help realize the demand for these datasets at a global scale. However, current ML methods for field instance segmentation lack sufficient geographic coverage, accuracy, and generalization capabilities. Further, research on improving ML methods is restricted by the lack of labeled datasets representing the diversity of global agricultural fields. We present Fields of The World (FTW)---a novel ML benchmark dataset for agricultural field instance segmentation spanning 24 countries on four continents (Europe, Africa, Asia, and South America). FTW is an order of magnitude larger than previous datasets with 70,462 samples, each containing instance and semantic segmentation masks paired with multi-date, multi-spectral Sentinel-2 satellite images. We provide results from baseline models for the new FTW benchmark, show that models trained on FTW have better zero-shot and fine-tuning performance in held-out countries than models that aren't pre-trained with diverse datasets, and show positive qualitative zero-shot results of FTW models in a real-world scenario -- running on Sentinel-2 scenes over Ethiopia.

Geographic information is essential for modeling tasks in fields ranging from ecology to epidemiology. However, extracting relevant location characteristics for a given task can be challenging, often requiring expensive data fusion or distillation from massive global imagery datasets. To address this challenge, we introduce Satellite Contrastive Location-Image Pretraining (SatCLIP). This global, general-purpose geographic location encoder learns an implicit representation of locations by matching CNN and ViT inferred visual patterns of openly available satellite imagery with their geographic coordinates. The resulting SatCLIP location encoder efficiently summarizes the characteristics of any given location for convenient use in downstream tasks. In our experiments, we use SatCLIP embeddings to improve performance on nine diverse geospatial prediction tasks including temperature prediction, animal recognition, and population density estimation. Across tasks, SatCLIP consistently outperforms alternative location encoders and shows promise for improving geographic domain adaptation. These results demonstrate the potential of vision-location models to learn meaningful representations of our planet from the vast, varied, and largely untapped modalities of geospatial data.

Satellite data has the potential to inspire a seismic shift for machine learning—one in which we rethink existing practices designed for traditional data modalities. As machine learning for satellite data (SatML) gains traction for its real-world impact, our field is at a crossroads. We can either continue applying ill-suited approaches, or we can initiate a new research agenda that centers around the unique characteristics and challenges of satellite data. This position paper argues that satellite data constitutes a distinct modality for machine learning research and that we must recognize it as such to advance the quality and impact of SatML research across theory, methods, and deployment. We outline research directions, critical discussion questions and actionable suggestions to transform SatML from merely an intriguing application area to a dedicated research discipline that helps move the needle on big challenges for machine learning and society.

Cropland mapping can play a vital role in addressing environmental, agricultural, and food security challenges. However, in the context of Africa, practical applications are often hindered by the limited availability of high-resolution cropland maps. Such maps typically require extensive human labeling, thereby creating a scalability bottleneck. To address this, we propose an approach that utilizes unsupervised object clustering to refine existing weak labels, such as those obtained from global cropland maps. The refined labels, in conjunction with sparse human annotations, serve as training data for a semantic segmentation network designed to identify cropland areas. We conduct experiments to demonstrate the benefits of the improved weak labels generated by our method. In a scenario where we train our model with only 33 human-annotated labels, the F1score for the cropland category increases from 0.53 to 0.84 when we add the mined negative labels.

This work presents an approach for combining household demographic and living standards survey questions with features derived from satellite imagery to predict the poverty rate of a region. Our approach utilizes visual features obtained from a single-step featurization method applied to freely available 10m/px Sentinel-2 surface reflectance satellite imagery. These visual features are combined with ten survey questions in a proxy means test (PMT) to estimate whether a household is below the poverty line. We show that the inclusion of visual features reduces the mean error in poverty rate estimates from 4.09% to 3.88% over a nationally representative out-of-sample test set. In addition to including satellite imagery features in proxy means tests, we propose an approach for selecting a subset of survey questions that are complementary to the visual features extracted from satellite imagery. Specifically, we design a survey variable selection approach guided by the full survey and image features and use the approach to determine the most relevant set of small survey questions to include in a PMT. We validate the choice of small survey questions in a downstream task of predicting the poverty rate using the small set of questions. This approach results in the best performance – errors in poverty rate decrease from 4.09% to 3.71%. We show that extracted visual features encode geographic and urbanization differences between regions.

The Landsat program is the longest-running Earth observation program in history, with 50+ years of data acquisition by 8 satellites. The multispectral imagery captured by sensors onboard these satellites is critical for a wide range of scientific fields. Despite the increasing popularity of deep learning and remote sensing, the majority of researchers still use decision trees and random forests for Landsat image analysis due to the prevalence of small labeled datasets and lack of foundation models. In this paper, we introduce SSL4EO-L, the first ever dataset designed for Self-Supervised Learning for Earth Observation for the Landsat family of satellites (including 3 sensors and 2 product levels) and the largest Landsat dataset in history (5M image patches). Additionally, we modernize and re-release the L7 Irish and L8 Biome cloud detection datasets, and introduce the first ML benchmark datasets for Landsats 4–5 TM and Landsat 7 ETM+ SR. Finally, we pre-train the first foundation models for Landsat imagery using SSL4EO-L and evaluate their performance on multiple semantic segmentation tasks. All datasets and model weights are available via the TorchGeo library, making reproducibility and experimentation easy, and enabling scientific advancements in the burgeoning field of remote sensing for a multitude of downstream applications.

Remotely sensed geospatial data are critical for applications including precision agriculture, urban planning, disaster monitoring and response, and climate change research, among others. Deep learning methods are particularly promising for modeling many remote sensing tasks given the success of deep neural networks in similar computer vision tasks and the sheer volume of remotely sensed imagery available. However, the variance in data collection methods and handling of geospatial metadata make the application of deep learning methodology to remotely sensed data nontrivial. For example, satellite imagery often includes additional spectral bands beyond red, green, and blue and must be joined to other geospatial data sources that can have differing coordinate systems, bounds, and resolutions. To help realize the potential of deep learning for remote sensing applications, we introduce TorchGeo, a Python library for integrating geospatial data into the PyTorch deep learning ecosystem. TorchGeo provides data loaders for a variety of benchmark datasets, composable datasets for generic geospatial data sources, samplers for geospatial data, and transforms that work with multispectral imagery. TorchGeo is also the first library to provide pre-trained models for multispectral satellite imagery (e.g., models that use all bands from the Sentinel-2 satellites), allowing for advances in transfer learning on downstream remote sensing tasks with limited labeled data. We use TorchGeo to create reproducible benchmark results on existing datasets and benchmark our proposed method for preprocessing geospatial imagery on the fly. TorchGeo is open source and available on GitHub: https://github.com/microsoft/torchgeo.

We propose a method for jointly inferring labels across a collection of data samples, where each sample consists of an observation and a prior belief about the label. By implicitly assuming the existence of a generative model for which a differentiable predictor is the posterior, we derive a training objective that allows learning under weak beliefs. This formulation unifies various machine learning settings; the weak beliefs can come in the form of noisy or incomplete labels, likelihoods given by a different prediction mechanism on auxiliary input, or common-sense priors reflecting knowledge about the structure of the problem at hand. We demonstrate the proposed algorithms on diverse problems: classification with negative training examples, learning from rankings, weakly and self-supervised aerial imagery segmentation, co-segmentation of video frames, and coarsely supervised text classification.

AI for good (AI4G) projects involve developing and applying artificial intelligence (AI) based solutions to further goals in areas such as sustainability, health, humanitarian aid, and social justice. Developing and deploying such solutions must be done in collaboration with partners who are experts in the domain in question and who already have experience in making progress towards such goals. Based on our experiences, we detail the different aspects of this type of collaboration broken down into four high-level categories: communication, data, modeling, and impact, and distill eleven takeaways to guide such projects in the future. We briefly describe two case studies to illustrate how some of these takeaways were applied in practice during our past collaborations.

Longitudinal studies are vital to understanding dynamic changes of the planet, but labels (e.g., buildings, facilities, roads) are often available only for a single point in time. We propose a general model, Temporal Cluster Matching (TCM), for detecting building changes in time series of remotely sensed imagery when footprint labels are observed only once. The intuition behind the model is that the relationship between spectral values inside and outside of building’s footprint will change when a building is constructed (or demolished). For instance, in rural settings, the pre-construction area may look similar to the surrounding environment until the building is constructed. Similarly, in urban settings, the pre-construction areas will look different from the surrounding environment until construction. We further propose a heuristic method for selecting the parameters of our model which allows it to be applied in novel settings without requiring data labeling efforts (to fit the parameters). We apply our model over a dataset of poultry barns from 2016/2017 high-resolution aerial imagery in the Delmarva Peninsula and a dataset of solar farms from a 2020 mosaic of Sentinel 2 imagery in India. Our results show that our model performs as well when fit using the proposed heuristic as it does when fit with labeled data, and further, that supervised versions of our model perform the best among all the baselines we test against. Finally, we show that our proposed approach can act as an effective data augmentation strategy – it enables researchers to augment existing structure footprint labels along the time dimension and thus use imagery from multiple points in time to train deep learning models. We show that this improves the spatial generalization of such models when evaluated on the same change detection task.

Modern deep learning-based semantic segmentation models and traditional pattern matching segmentation methods demonstrate similar failure modes in mapping land cover from diverse satellite/aerial imagery. The key problem is that these models mostly respond to textures and colors which, locally, do tend to have consistent land cover labels, but may resemble very different labels in imagery acquired farther away, with a different sensor, or under new imaging conditions. One way to resolve this issue is to endow the algorithms with higher-level, human-like reasoning abilities - e.g., an awareness that houses are connected to roads with driveways, that roads connect towns, and that bridges cast shadows - and the mechanism for tracking such objects across larger areas in order to resolve ambiguity. We propose an alternative approach of human-machine collaboration for creating land cover maps, motivated by the fact that a large amount of human labor is necessary even in seemingly automated land cover mapping solutions. We build spatial ensembles of land cover models with weak supervision and treat them as a hypothesis space. Then, we build an interface where humans, aware of the higher-level structure in imagery, can act to select, apply, and refine these models locally, with the goal of minimizing the labor required to create a land cover map.

We propose incorporating human labelers in a model fine-tuning system that provides immediate user feedback. In our framework, human labelers can interactively query model predictions on unlabeled data, choose which data to label, and see the resulting effect on the model's predictions. This bi-directional feedback loop allows humans to learn how the model responds to new data. We implement this framework for fine-tuning high-resolution land cover segmentation models and compare human-selected points to points selected using standard active learning methods. Specifically, we fine-tune a deep neural network – trained to segment high-resolution aerial imagery into different land cover classes in Maryland, USA – to a new spatial area in New York, USA using both our human-in-the-loop method and traditional active learning methods. The tight loop in our proposed system turns the algorithm and the human operator into a hybrid system that can produce land cover maps of large areas more efficiently than the traditional workflows. Our framework has applications in machine learning settings where there is a practically limitless supply of unlabeled data, of which only a small fraction can feasibly be labeled through human efforts, such as geospatial and medical image-based applications.

Normalization layers have been shown to improve convergence in deep neural networks, and even add useful inductive biases. In many vision applications the local spatial context of the features is important, but most common normalization schemes including Group Normalization (GN), Instance Normalization (IN), and Layer Normalization (LN) normalize over the entire spatial dimension of a feature. This can wash out important signals and degrade performance. For example, in applications that use satellite imagery, input images can be arbitrarily large; consequently, it is nonsensical to normalize over the entire area. Positional Normalization (PN), on the other hand, only normalizes over a single spatial position at a time. A natural compromise is to normalize features by local context, while also taking into account group level information. In this pa- per, we propose Local Context Normalization (LCN): a normalization layer where every feature is normalized based on a window around it and the filters in its group. We propose an algorithmic solution to make LCN efficient for arbitrary window sizes, even if every point in the image has a unique window. LCN outperforms its Batch Normalization (BN), GN, IN, and LN counterparts for object detection, semantic segmentation, and instance segmentation applications in several benchmark datasets, while keeping performance in- dependent of the batch size and facilitating transfer learning.

We propose an iterative clustering-based label super-resolution approach and epitome-based approach to weakly supervised semantic segmentation, as well as a deep learning-based postprocessing step for land cover segmentation. An ensemble of the iterative clustering and epitome approaches with the proposed postprocessing step results in a top validation leaderboard average accuracy of 70.43%. A similar ensemble, that also considers class accuracy feedback from the leaderboard, achieves a top Track 1 leaderboard average accuracy of 57.49%.

Deep learning is increasingly used in decision-making tasks. However, understanding how neural networks produce final predictions remains a fundamental challenge. Existing work on interpreting neural network predictions for images often focuses on explaining predictions for single images or neurons. As predictions are often computed from millions of weights that are optimized over millions of images, such explanations can easily miss a bigger picture. We present Summit, an interactive system that scalably and systematically summarizes and visualizes what features a deep learning model has learned and how those features interact to make predictions. Summit introduces two new scalable summarization techniques: (1) activation aggregation discovers important neurons, and (2) neuron-influence aggregation identifies relationships among such neurons. Summit combines these techniques to create the novel attribution graph that reveals and summarizes crucial neuron associations and substructures that contribute to a model's outcomes. Summit scales to large data, such as the ImageNet dataset with 1.2M images, and leverages neural network feature visualization and dataset examples to help users distill large, complex neural network models into compact, interactive visualizations. We present neural network exploration scenarios where Summit helps us discover multiple surprising insights into a prevalent, large-scale image classifier's learned representations and informs future neural network architecture design. The Summit visualization runs in modern web browsers and is open-sourced.

In this paper we propose multi-resolution data fusion methods for deep learning-based high-resolution land cover mapping from aerial imagery. The land cover mapping problem, at country-level scales, is challenging for common deep learning methods due to the scarcity of high-resolution labels, as well as variation in geography and quality of input images. On the other hand, multiple satellite imagery and low-resolution ground truth label sources are widely available, and can be used to improve model training efforts. Our methods include: introducing low-resolution satellite data to smooth quality differences in high-resolution input, exploiting low-resolution labels with a dual loss function, and pairing scarce high-resolution labels with inputs from several points in time. We train models that are able to generalize from a portion of the Northeast United States, where we have high-resolution land cover labels, to the rest of the US. With these models, we produce the first high-resolution (1-meter) land cover map of the contiguous US, consisting of over 8 trillion pixels. We demonstrate the robustness and potential applications of this data in a case study with domain experts and develop a web application to share our results. This work is practically useful, and can be applied to other locations over the earth as high-resolution imagery becomes more widely available even as high-resolution labeled land cover data remains sparse.

Developing and maintaining resilient transportation infrastructure is a key strategy for meeting several UN sustainable development goals in the face of climate change-driven extreme flooding events. We present a framework for performing data-driven vulnerability analysis for flooding on existing transportation networks, and use this analysis to inform decision-making about investments for climate adaptation. We apply this approach to study the potential impacts of severe flooding on regional mobility in Senegal, using a combination of flood hazard maps and a travel demand model based on call detail record data. We use the estimated number of infeasible trips as a direct measure of flooding-induced mobility impacts, as well as an objective for minimizing these impacts. We then compare three alternative road network upgrade strategies to assess the extent to which each strategy would preserve network functionality under a given flooding scenario. We illustrate that strategies driven solely by travel demand can lead to underinvestment in roads that are at risk of flooding, while solely focusing on repairing flooded road segments neglects the criticality of those repairs to mobility. For example, in a 100 year flooding scenario with a fixed budget, our strategy that considers both flooding and mobility data can achieve a 53% reduction in the number of infeasible trips, while a strategy that just considers flooding data achieves only a 38% reduction for the same cost. Our framework can be applied more broadly to integrate information from a variety of sources about climate hazards and potential human impacts to make better informed decisions about investments in critical infrastructure systems.

Human migration is a type of human mobility, where a trip involves a person moving with the intention of changing their home location. Predicting human migration as accurately as possible is important in city planning applications, international trade, spread of infectious diseases, conservation planning, and public policy development. Traditional human mobility models, such as gravity models or the more recent radiation model, predict human mobility flows based on population and distance features only. These models have been validated on commuting flows, a different type of human mobility, and are mainly used in modeling scenarios where large amounts of prior ground truth mobility data are not available. One downside of these models is that they have a fixed form and are therefore not able to capture more complicated migration dynamics. We propose machine learning models that are able to incorporate any number of exogenous features, to predict origin/destination human migration flows. Our machine learning models outperform traditional human mobility models on a variety of evaluation metrics, both in the task of predicting migrations between US counties as well as international migrations. In general, predictive machine learning models of human migration will provide a flexible base with which to model human migration under different what-if conditions, such as potential sea level rise or population growth scenarios.

The world food system is a globalized and inter-connected network, and prior research has shown the US is one of the most important countries within this global network. Considering this, any affects on the US food network might have far reaching implications. The objective of this paper is to show how the efficiency of the US food network can be optimized and to examine the tradeoffs between sustainability, efficiency, and resiliency in a network science based approach. To do this we use the 2012 US Commodity Flow Survey (CFS) data focusing on origin-destination flows for 5 categories of food. We interpret these origin-destination flows as a directed weighted network and use a linear programming (LP) based approach to reconfigure the flows in order to minimize the “food miles” in the network. We extend our LP to a multi-objective optimization problem in order to consider partially optimized solutions that are closer to the current food network. Finally, we calculate characteristics of the optimized networks and compare them to the original network. Our results show that the US food network has the potential for more than a 50% reduction in “food miles”, which will result in a food system that is not only more economically efficient but also with a smaller sustainability impact by reducing the corresponding transportation GHG emissions.