Esther Rolf
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Esther Rolf has a total of 57 co-authors. Here are the top co-authors by number of shared publications:
- Ian Bolliger
Reask
3 shared publications - Konstantin Klemmer
Department of Computer Science Department of Computer Science & Center for Urban Science and Progress University of Warwick The Alan Turing Institute $ New York University
3 shared publications - Marc Rußwurm
Wageningen University
3 shared publications - Solomon Hsiang
Stanford Doerr School of Sustainability, Stanford University, Stanford, California, USA
3 shared publications - Andrew Hultgren
Energy Policy Institute, University of Chicago, Chicago, USA
2 shared publications - Benjamin Recht
Department of Electrical Engineering and Computer Sciences, University of California, Berkeley, USA
2 shared publications
Latest Publications
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Magazine
- p. 2-15
Earth Embeddings: Toward artificial intelligence-centric representations of our planet
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Magazine
- p. 2-15
Open AccessEarth Embeddings: Toward artificial intelligence-centric representations of our planet
- Open Access
- 10 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-2
Earth Embeddings: Harnessing the Information in Earth Observation Data with Machine Learning
- 10 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-2
Open AccessEarth Embeddings: Harnessing the Information in Earth Observation Data with Machine Learning
- Open Access
- 16 October 2024
- Published by IOS Press
- p. 3284-3291
Combining Diverse Information for Coordinated Action: Stochastic Bandit Algorithms for Heterogeneous Agents
Stochastic multi-agent multi-armed bandits typically assume that the rewards from each arm follow a fixed distribution, regardless of which agent pulls the arm. However, in many real-world settings, rewards can depend on the sensitivity of each agent to their environment. In medical screening, disease detection rates can vary by test type; in preference matching, rewards can depend on user preferences; and in environmental sensing, observation quality can vary across sensors. Since past work does not specify how to allocate agents of heterogeneous but known sensitivity of these types in a stochastic bandit setting, we introduce a UCB-style algorithm, Min-Width, which aggregates information from diverse agents. In doing so, we address the joint challenges of (i) aggregating the rewards, which follow different distributions for each agent-arm pair, and (ii) coordinating the assignments of agents to arms. Min-Width facilitates efficient collaboration among heterogeneous agents, exploiting the known structure in the agents’ reward functions to weight their rewards accordingly. We analyze the regret of Min-Width and conduct pseudo-synthetic and fully synthetic experiments to study the performance of different levels of information sharing. Our results confirm that the gains to modeling agent heterogeneity tend to be greater when the sensitivities are more varied across agents, while combining more information does not always improve performance.
- 16 October 2024
- Published by IOS Press
- p. 3284-3291
Open AccessCombining Diverse Information for Coordinated Action: Stochastic Bandit Algorithms for Heterogeneous Agents
Stochastic multi-agent multi-armed bandits typically assume that the rewards from each arm follow a fixed distribution, regardless of which agent pulls the arm. However, in many real-world settings, rewards can depend on the sensitivity of each agent to their environment. In medical screening, disease detection rates can vary by test type; in preference matching, rewards can depend on user preferences; and in environmental sensing, observation quality can vary across sensors. Since past work does not specify how to allocate agents of heterogeneous but known sensitivity of these types in a stochastic bandit setting, we introduce a UCB-style algorithm, Min-Width, which aggregates information from diverse agents. In doing so, we address the joint challenges of (i) aggregating the rewards, which follow different distributions for each agent-arm pair, and (ii) coordinating the assignments of agents to arms. Min-Width facilitates efficient collaboration among heterogeneous agents, exploiting the known structure in the agents’ reward functions to weight their rewards accordingly. We analyze the regret of Min-Width and conduct pseudo-synthetic and fully synthetic experiments to study the performance of different levels of information sharing. Our results confirm that the gains to modeling agent heterogeneity tend to be greater when the sensitivities are more varied across agents, while combining more information does not always improve performance.
- Open Access
- 2 October 2024
- Published by Wiley
Preserving open data as a world heritage to secure our future
Addressing global challenges and taking action towards a sustainable future-in regards to ecology, biodiversity, water and food security and climate-requires both solution-driven research and a deep understanding of Earth and space systems. Trustworthy, reproducible science requires that data, including their provenance, be accessible for evaluation and further study. International collaboration and open data sharing are essential for addressing these challenges and promoting new scientific advances. Although strides have been made in improving data sharing platforms, as well as data and metadata standards, there is still work to be done to facilitate careful collection, stewardship and reuse of data across the Earth and space sciences. To achieve an ecosystem where data are consistently collected, well described, curated, shared, preserved and reused there needs to be significant continuous evolution in (i) equitable access to data and infrastructure, (ii) a change in the research culture and the scientific rewarding system to value data contributions, and (iii) standards around describing and documenting data across disciplines and domains. This statement takes the broadest interpretation of the term 'data' according to the Beijing Declaration on Research Data: data can be collected, generated or compiled by humans or machines and include (but not limited to) metadata, samples, methods, software and algorithms. Infrastructure includes hardware, intangible assets such as software and human capital. Ensure equitable access to trusted and reusable data Researchers have a responsibility to collect, document and share data-raw and processed-in an ethical manner that is as open and transparent as possible. Persistent funding is crucial for maintaining open and trusted data infrastructures, such as repositories, to facilitate ease of deposit and long-term reuse. This infrastructure supports the management and curation of data throughout the research process, ensuring that data are findable, accessible, interoperable and reusable (FAIR) for both people and machines. To meet these expectations, investment in support staff, such as data stewards and IT developers, along with research and development of data architectures is imperative.
- 2 October 2024
- Published by Wiley
Open AccessPreserving open data as a world heritage to secure our future
Addressing global challenges and taking action towards a sustainable future-in regards to ecology, biodiversity, water and food security and climate-requires both solution-driven research and a deep understanding of Earth and space systems. Trustworthy, reproducible science requires that data, including their provenance, be accessible for evaluation and further study. International collaboration and open data sharing are essential for addressing these challenges and promoting new scientific advances. Although strides have been made in improving data sharing platforms, as well as data and metadata standards, there is still work to be done to facilitate careful collection, stewardship and reuse of data across the Earth and space sciences. To achieve an ecosystem where data are consistently collected, well described, curated, shared, preserved and reused there needs to be significant continuous evolution in (i) equitable access to data and infrastructure, (ii) a change in the research culture and the scientific rewarding system to value data contributions, and (iii) standards around describing and documenting data across disciplines and domains. This statement takes the broadest interpretation of the term 'data' according to the Beijing Declaration on Research Data: data can be collected, generated or compiled by humans or machines and include (but not limited to) metadata, samples, methods, software and algorithms. Infrastructure includes hardware, intangible assets such as software and human capital. Ensure equitable access to trusted and reusable data Researchers have a responsibility to collect, document and share data-raw and processed-in an ethical manner that is as open and transparent as possible. Persistent funding is crucial for maintaining open and trusted data infrastructures, such as repositories, to facilitate ease of deposit and long-term reuse. This infrastructure supports the management and curation of data throughout the research process, ensuring that data are findable, accessible, interoperable and reusable (FAIR) for both people and machines. To meet these expectations, investment in support staff, such as data stewards and IT developers, along with research and development of data architectures is imperative.
- Open Access
SatCLIP: Global, General-Purpose Location Embeddings with Satellite Imagery
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 prediction performance on nine diverse location-dependent tasks including temperature prediction, animal recognition, and population density estimation. Across tasks, SatCLIP consistently outperforms alternative location encoders and improves geographic generalization by encoding visual similarities of spatially distant environments. 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.
Open AccessSatCLIP: Global, General-Purpose Location Embeddings with Satellite Imagery
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 prediction performance on nine diverse location-dependent tasks including temperature prediction, animal recognition, and population density estimation. Across tasks, SatCLIP consistently outperforms alternative location encoders and improves geographic generalization by encoding visual similarities of spatially distant environments. 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.
- Open Access
Resolving label uncertainty with implicit posterior models
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.
Open AccessResolving label uncertainty with implicit posterior models
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.
- Top Cited PapersOpen Access
- 20 July 2021
- Published by Springer Nature in Nature Communications
- Vol. 12 (1) , 1-11
A generalizable and accessible approach to machine learning with global satellite imagery
Combining satellite imagery with machine learning (SIML) has the potential to address global challenges by remotely estimating socioeconomic and environmental conditions in data-poor regions, yet the resource requirements of SIML limit its accessibility and use. We show that a single encoding of satellite imagery can generalize across diverse prediction tasks (e.g., forest cover, house price, road length). Our method achieves accuracy competitive with deep neural networks at orders of magnitude lower computational cost, scales globally, delivers label super-resolution predictions, and facilitates characterizations of uncertainty. Since image encodings are shared across tasks, they can be centrally computed and distributed to unlimited researchers, who need only fit a linear regression to their own ground truth data in order to achieve state-of-the-art SIML performance.
- 20 July 2021
- Published by Springer Nature in Nature Communications
- Vol. 12 (1) , 1-11
Top Cited PapersOpen AccessA generalizable and accessible approach to machine learning with global satellite imagery
Combining satellite imagery with machine learning (SIML) has the potential to address global challenges by remotely estimating socioeconomic and environmental conditions in data-poor regions, yet the resource requirements of SIML limit its accessibility and use. We show that a single encoding of satellite imagery can generalize across diverse prediction tasks (e.g., forest cover, house price, road length). Our method achieves accuracy competitive with deep neural networks at orders of magnitude lower computational cost, scales globally, delivers label super-resolution predictions, and facilitates characterizations of uncertainty. Since image encodings are shared across tasks, they can be centrally computed and distributed to unlimited researchers, who need only fit a linear regression to their own ground truth data in order to achieve state-of-the-art SIML performance.
- Top Cited PapersOpen Access
- 22 August 2020
- Published by Springer Nature in Nature
- Vol. 585 (7824) , E7
Publisher Correction: The effect of large-scale anti-contagion policies on the COVID-19 pandemic
- 22 August 2020
- Published by Springer Nature in Nature
- Vol. 585 (7824) , E7
Top Cited PapersOpen AccessPublisher Correction: The effect of large-scale anti-contagion policies on the COVID-19 pandemic
- 28 July 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Robotics
- Vol. 37 (1) , 34-47
A Successive-Elimination Approach to Adaptive Robotic Source Seeking
In this article, we study an adaptive source seeking problem, in which a mobile robot must identify the strongest emitter(s) of a signal in an environment with background emissions. Background signals may be highly heterogeneous and can mislead algorithms that are based on receding horizon control. We propose $\mathtt {AdaSearch}$ , a general algorithm for adaptive source seeking in the face of heterogeneous background noise. $\mathtt {AdaSearch}$ combines global trajectory planning with principled confidence intervals in order to concentrate measurements in promising regions while guaranteeing sufficient coverage of the entire area. Theoretical analysis shows that $\mathtt {AdaSearch}$ confers gains over a uniform sampling strategy when the distribution of background signals is highly variable. Simulation experiments demonstrate that when applied to the problem of radioactive source-seeking, $\mathtt {AdaSearch}$ outperforms both uniform sampling and a receding time horizon information-maximization approach based on the current literature. We also demonstrate $\mathtt {AdaSearch}$ in hardware, providing further evidence of its potential for real-time implementation.- 28 July 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Robotics
- Vol. 37 (1) , 34-47
A Successive-Elimination Approach to Adaptive Robotic Source Seeking
In this article, we study an adaptive source seeking problem, in which a mobile robot must identify the strongest emitter(s) of a signal in an environment with background emissions. Background signals may be highly heterogeneous and can mislead algorithms that are based on receding horizon control. We propose $\mathtt {AdaSearch}$ , a general algorithm for adaptive source seeking in the face of heterogeneous background noise. $\mathtt {AdaSearch}$ combines global trajectory planning with principled confidence intervals in order to concentrate measurements in promising regions while guaranteeing sufficient coverage of the entire area. Theoretical analysis shows that $\mathtt {AdaSearch}$ confers gains over a uniform sampling strategy when the distribution of background signals is highly variable. Simulation experiments demonstrate that when applied to the problem of radioactive source-seeking, $\mathtt {AdaSearch}$ outperforms both uniform sampling and a receding time horizon information-maximization approach based on the current literature. We also demonstrate $\mathtt {AdaSearch}$ in hardware, providing further evidence of its potential for real-time implementation.- Open Access
A Successive-Elimination Approach to Adaptive Robotic Sensing
We study an adaptive source seeking problem, in which a mobile robot must identify the strongest emitter(s) of a signal in an environment with background emissions. Background signals may be highly heterogeneous and can mislead algorithms that are based on receding horizon control. We propose AdaSearch, a general algorithm for adaptive source seeking in the face of heterogeneous background noise. AdaSearch combines global trajectory planning with principled confidence intervals in order to concentrate measurements in promising regions while guaranteeing sufficient coverage of the entire area. Theoretical analysis shows that AdaSearch confers gains over a uniform sampling strategy when the distribution of background signals is highly variable. Simulation experiments demonstrate that when applied to the problem of radioactive source seeking, AdaSearch outperforms both uniform sampling and a receding time horizon information-maximization approach based on the current literature. We also demonstrate AdaSearch in hardware, providing further evidence of its potential for real-time implementation.
Open AccessA Successive-Elimination Approach to Adaptive Robotic Sensing
We study an adaptive source seeking problem, in which a mobile robot must identify the strongest emitter(s) of a signal in an environment with background emissions. Background signals may be highly heterogeneous and can mislead algorithms that are based on receding horizon control. We propose AdaSearch, a general algorithm for adaptive source seeking in the face of heterogeneous background noise. AdaSearch combines global trajectory planning with principled confidence intervals in order to concentrate measurements in promising regions while guaranteeing sufficient coverage of the entire area. Theoretical analysis shows that AdaSearch confers gains over a uniform sampling strategy when the distribution of background signals is highly variable. Simulation experiments demonstrate that when applied to the problem of radioactive source seeking, AdaSearch outperforms both uniform sampling and a receding time horizon information-maximization approach based on the current literature. We also demonstrate AdaSearch in hardware, providing further evidence of its potential for real-time implementation.