Italian ryegrass is a grass weed commonly found in winter wheat fields that are competitive with winter wheat for moisture and nutrients. Ryegrass can cause substantial reductions in yield and grain quality if not properly controlled with the use of herbicides. To control the cost and environmental impact we detect weeds in drone and satellite imagery. Satellite imagery is too coarse to be used for precision spraying, but can aid in planning drone flights and treatments. Drone images on the other hand have sufficiently good resolution for precision spraying. However, ryegrass is hard to distinguish from the crop and annotation requires expert knowledge. We used the Python segmentation models library to test more than 600 different neural network architectures for weed segmentation in drone images and we map accuracy versus the cost of the model prediction for these. Our best system applies herbicides to over 99% of the weeds while only spraying an area 30% larger than the annotated weed area. These models yield large savings if the weed covers a small part of the field.

Hyperspectral Imaging (HSI) is known for its advantages over traditional RGB imaging in remote sensing, agriculture, and medicine. Recently, it has gained attention for enhancing Advanced Driving Assistance Systems (ADAS) perception. Several HSI datasets such as HyKo, HSI-Drive, HSI-Road, and Hyperspectral City have been made available. However, a comprehensive evaluation of semantic segmentation models (SSM) using these datasets is lacking. To address this gap, we evaluated the available annotated HSI datasets on four deep learning-based baseline SSMs: DeepLab v3+, HRNet, PSPNet, and U-Net, along with its two variants: Coordinate Attention (UNet-CA) and Convolutional Block-Attention Module (UNet-CBAM). The original model architectures were adapted to handle the varying spatial and spectral dimensions of the datasets. These baseline SSMs were trained using a class-weighted loss function for individual HSI datasets and evaluated using mean-based metrics such as intersection over union (IoU), recall, precision, F1 score, specificity, and accuracy. Our results indicate that UNet-CBAM, which extracts channel-wise features, outperforms other SSMs and shows potential to leverage spectral information for enhanced semantic segmentation. This study establishes a baseline SSM benchmark on available annotated datasets for future evaluation of HSI-based ADAS perception. However, limitations of current HSI datasets, such as limited dataset size, high class imbalance, and lack of fine-grained annotations, remain significant constraints for developing robust SSMs for ADAS applications.

This study investigates the impact of High-intensity Long-Duration Continuous Auroral Electrojet Activity (HILDCAA) on the relativistic electrons in radiation belt of Earth. Utilizing data from Van Allen Probe mission of NASA, we conducted a comprehensive statistical analysis to understand the impact of HILDCAA events on the radiation belt fluxes. The super epoch analysis was carried out to determine the general response of L-shell, pitch angle, and energy-dependency of relativistic electrons to HILDCAAs. The analysis reveals a significant flux enhancement in the relativistic electron fluxes, predominantly occurring with a delay of 0 to 2 days following the onset of HILDCAA events. The general response indicates that the maximum energy of accelerated electrons reaches up to 6 MeV. Additionally, electrons with perpendicular pitch angles exhibit a significantly greater enhancement in flux and achieve higher maximum acceleration energies compared to those with parallel pitch angles. The observed time-delayed and pitch angle-dependent response related to the onset of HILDCAAs highlights the significant influence of wave-particle interactions, particularly in relation to ultra-low frequency (ULF) waves in this context. This is further supported by ground-based magnetometers and in-situ magnetic field observations from the RBSP probe, which demonstrated enhanced power of ULF waves during HILDCAA events. The study strengthens our current understanding of radiation belt particle acceleration processes and has potential implications for satellite operations and other space-based technologies, both on Earth and in the magnetospheres of other planets.

Following the success of Large Language Models (LLMs), expanding their boundaries to new modalities represents a significant paradigm shift in multimodal understanding. Human perception is inherently multimodal, relying not only on text but also on auditory and visual cues for a complete understanding of the world. In recognition of this fact, audio-visual LLMs have recently emerged. Despite promising developments, the lack of dedicated benchmarks poses challenges for understanding and evaluating models. In this work, we show that audio-visual LLMs struggle to discern subtle relationships between audio and visual signals, leading to hallucinations and highlighting the need for reliable benchmarks. To address this, we introduce AVHBench, the first comprehensive benchmark specifically designed to evaluate the perception and comprehension capabilities of audio-visual LLMs. Our benchmark includes tests for assessing hallucinations, as well as the cross-modal matching and reasoning abilities of these models. Our results reveal that most existing audio-visual LLMs struggle with hallucinations caused by cross-interactions between modalities, due to their limited capacity to perceive complex multimodal signals and their relationships. Additionally, we demonstrate that simple training with our AVHBench improves robustness of audio-visual LLMs against hallucinations. Dataset: https://github.com/kaist-ami/AVHBench

Continental heatwaves can dramatically impact ecosystems and societies, e.g., by leading to excess mortality, wildfires, and harvest failures. With a warming climate, their impacts potentially intensify globally, but the Indian subcontinent appears to be particularly vulnerable to such extreme events. In this study, we use reanalysis and the adjoint of the atmospheric model, PlaSim, to identify drivers of heatwaves occurring April and May over north-central India. Reanalysis results suggest that the existence of high temperatures in the study region is highly sensitive to the low local soil moisture which is observed weeks before a heatwave commences. Soil moisture variability in northern India is influenced by moisture transport from the west during winter--spring. Preceding dry soil moisture conditions can be associated with a `persistent jet' conditions linked to atmospheric dynamical changes in the North Atlantic region. An associated northward shift in the upper tropospheric zonal wind occurs approximately a month prior to the heatwaves, influencing the area and intensity of western disturbances embedded in the jet stream. This weakens the moisture flow from the north of the Arabian Sea, further reducing soil moisture levels and creating conditions conducive to heatwaves. An adjoint sensitivity analysis and forward model perturbation experiments confirm the causal relationships for the proposed heatwave development mechanism over north-central India, identifying the remote influence of North Atlantic sea surface temperature variability on extreme temperatures in India. Our findings highlight the complex interplay of local and remote factors in heatwave development over India.

Time series anomaly prediction plays an essential role in many real-world scenarios, such as environmental prevention and prompt maintenance of cyber-physical systems. However, existing time series anomaly prediction methods mainly require supervised training with plenty of manually labeled data, which are difficult to obtain in practice. Besides, unseen anomalies can occur during inference, which could differ from the labeled training data and make these models fail to predict such new anomalies. In this paper, we study a novel problem of unsupervised time series anomaly prediction. We provide a theoretical analysis and propose Importance-based Generative Contrastive Learning (IGCL) to address the aforementioned problems. IGCL distinguishes between normal and anomaly precursors, which are generated by our anomaly precursor pattern generation module. To address the efficiency issues caused by the potential complex anomaly precursor combinations, we propose a memory bank with importance-based scores to adaptively store representative anomaly precursors and generate more complicated anomaly precursors. Extensive experiments on seven benchmark datasets show our method outperforms state-of-the-art baselines on unsupervised time series anomaly prediction problems.

Renewable energy is becoming an increasingly important component of energy systems. However, renewable energy production is heavily dependent on the prevailing weather conditions, which are changing as a result of climate change. It is therefore necessary to build energy systems that are robust to energy shortages caused by weather-dependent changes to energy demand and renewable energy production. To design such systems, we must monitor how changes in the climate are expected to influence future energy production and demand; this is important for policymakers to decide when, where, and by how much renewable energy installed capacities should be increased, for example. In this paper, we study the behaviour of standardised energy indices in future European climate projections, and use this to monitor how characteristics of energy production droughts in Europe are expected to change in the future. We use these results to make suggestions regarding how the energy mix should be adapted in the future to decrease the risk of energy production droughts.

This study investigates the orbital decay and subsequent reentries of 12 Starlink satellites from 16 April to 15 May 2024. By examining Two-Line Element data, we observed a significant increase in orbital decay following the geomagnetic storm on 10 May 2024, consistent with expectations of increased thermospheric density. An unexpected increase in decay rates for 10 satellites was identified around 25 April 2024, while two lower-altitude satellites remained unaffected. Detailed analysis revealed that this enhanced decay rate prior to the storm was influenced by a spike in the O/N2 ratio and an increase in Extreme Ultra Violet (EUV) flux. Moreover, most of the satellites exhibited sharp decay during the early recovery phase of the geomagnetic storm. Based on the positions and local times of changes in decay rates, it is likely that the satellites were affected by various processes during elevated space weather activity, such as enhanced EUV flux, Joule heating, particle precipitation, and the equatorial neutral anomaly. This study highlights the complex role of preconditioning due to enhanced EUV flux and extreme space weather activity in the orbital dynamics of Low-Earth Orbit (LEO) satellites.

While recent vision-language-action models trained on diverse robot datasets exhibit promising generalization capabilities with limited in-domain data, their reliance on compact action heads to predict discretized or continuous actions constrains adaptability to heterogeneous action spaces. We present Dita, a scalable framework that leverages Transformer architectures to directly denoise continuous action sequences through a unified multimodal diffusion process. Departing from prior methods that condition denoising on fused embeddings via shallow networks, Dita employs in-context conditioning -- enabling fine-grained alignment between denoised actions and raw visual tokens from historical observations. This design explicitly models action deltas and environmental nuances. By scaling the diffusion action denoiser alongside the Transformer's scalability, Dita effectively integrates cross-embodiment datasets across diverse camera perspectives, observation scenes, tasks, and action spaces. Such synergy enhances robustness against various variances and facilitates the successful execution of long-horizon tasks. Evaluations across extensive benchmarks demonstrate state-of-the-art or comparative performance in simulation. Notably, Dita achieves robust real-world adaptation to environmental variances and complex long-horizon tasks through 10-shot finetuning, using only third-person camera inputs. The architecture establishes a versatile, lightweight and open-source baseline for generalist robot policy learning. Project Page: https://robodita.github.io/

Recent studies have demonstrated the strong empirical performance of diffusion models on discrete sequences across domains from natural language to biological sequence generation. For example, in the protein inverse folding task, conditional diffusion models have achieved impressive results in generating natural-like sequences that fold back into the original structure. However, practical design tasks often require not only modeling a conditional distribution but also optimizing specific task objectives. For instance, we may prefer protein sequences with high stability. To address this, we consider the scenario where we have pre-trained discrete diffusion models that can generate natural-like sequences, as well as reward models that map sequences to task objectives. We then formulate the reward maximization problem within discrete diffusion models, analogous to reinforcement learning (RL), while minimizing the KL divergence against pretrained diffusion models to preserve naturalness. To solve this RL problem, we propose a novel algorithm, DRAKES, that enables direct backpropagation of rewards through entire trajectories generated by diffusion models, by making the originally non-differentiable trajectories differentiable using the Gumbel-Softmax trick. Our theoretical analysis indicates that our approach can generate sequences that are both natural-like and yield high rewards. While similar tasks have been recently explored in diffusion models for continuous domains, our work addresses unique algorithmic and theoretical challenges specific to discrete diffusion models, which arise from their foundation in continuous-time Markov chains rather than Brownian motion. Finally, we demonstrate the effectiveness of DRAKES in generating DNA and protein sequences that optimize enhancer activity and protein stability, respectively, important tasks for gene therapies and protein-based therapeutics.

Omnidirectional video (ODV) provides an immersive visual experience and is widely utilized in virtual reality and augmented reality. However, restricted capturing devices and transmission bandwidth lead to low-resolution ODVs. Video super-resolution (SR) is proposed to enhance resolution, but practical ODV spatial projection distortions and temporal flickering are not well addressed directly applying existing methods. To achieve better ODV-SR reconstruction, we propose a Spatio-Temporal Distortion Aware Network (STDAN) oriented to ODV characteristics. Specifically, a spatially continuous distortion modulation module is introduced to improve discrete projection distortions. Next, we design an interlaced multi-frame reconstruction mechanism to refine temporal consistency across frames. Furthermore, we incorporate latitude-saliency adaptive weights during training to concentrate on regions with higher texture complexity and human-watching interest. In general, we explore inference-free and real-world viewing matched strategies to provide an application-friendly method on a novel ODV-SR dataset with practical scenarios. Extensive experimental results demonstrate the superior performance of the proposed STDAN over state-of-the-art methods.

Modern climate change presents unprecedented challenges, posing critical crises that threaten sustainable development, human well-being, and planetary health. A significant concern is the potential for global warming to cause irreversible disruptions to the water-carbon cycle, a topic that remains underexplored. This study seeks to address a crucial knowledge gap by examining how increasing wet extremes impact ecosystem productivity. The research agenda focuses on three primary questions: 1) How do the intensity and duration of various wet extremes affect evapotranspiration across different watersheds and terrestrial biomes? 2) How do immediate and lagged responses to wet extremes vary across different biomes, and what insights do these temporal patterns provide about the causal and predictive relationships between wet extreme and evapotranspiration? 3) To what extent do watershed characteristics (such as soil properties, hydrological conditions, and vegetation factors) modulate the relationship between wet extremes and ecosystem productivity? As climate change alters precipitation patterns, understanding these complex ecosystem responses becomes crucial for developing adaptive strategies and improving food and water resource management.

This research addresses the need for high-definition (HD) maps for autonomous vehicles (AVs), focusing on road lane information derived from aerial imagery. While Earth observation data offers valuable resources for map creation, specialized models for road lane extraction are still underdeveloped in remote sensing. In this study, we perform an extensive comparison of twelve foundational deep learning-based semantic segmentation models for road lane marking extraction from high-definition remote sensing images, assessing their performance under transfer learning with partially labeled datasets. These models were fine-tuned on the partially labeled Waterloo Urban Scene dataset, and pre-trained on the SkyScapes dataset, simulating a likely scenario of real-life model deployment under partial labeling. We observed and assessed the fine-tuning performance and overall performance. Models showed significant performance improvements after fine-tuning, with mean IoU scores ranging from 33.56% to 76.11%, and recall ranging from 66.0% to 98.96%. Transformer-based models outperformed convolutional neural networks, emphasizing the importance of model pre-training and fine-tuning in enhancing HD map development for AV navigation.

The assumption of Gaussian or Gaussian mixture data has been extensively exploited in a long series of precise performance analyses of machine learning (ML) methods, on large datasets having comparably numerous samples and features. To relax this restrictive assumption, subsequent efforts have been devoted to establish "Gaussian equivalent principles" by studying scenarios of Gaussian universality where the asymptotic performance of ML methods on non-Gaussian data remains unchanged when replaced with Gaussian data having the same mean and covariance. Beyond the realm of Gaussian universality, there are few exact results on how the data distribution affects the learning performance. In this article, we provide a precise high-dimensional characterization of empirical risk minimization, for classification under a general mixture data setting of linear factor models that extends Gaussian mixtures. The Gaussian universality is shown to break down under this setting, in the sense that the asymptotic learning performance depends on the data distribution beyond the class means and covariances. To clarify the limitations of Gaussian universality in the classification of mixture data and to understand the impact of its breakdown, we specify conditions for Gaussian universality and discuss their implications for the choice of loss function.

We consider the problem of adapting a contrastively pretrained vision-language model like CLIP (Radford et al., 2021) for few-shot classification. The literature addresses this problem by learning a linear classifier of the frozen visual features, optimizing word embeddings, or learning external feature adapters. We introduce an alternative way for few-shot CLIP adaptation without adding ''external'' parameters to optimize. We find that simply fine-tuning the embedding projection matrix of the vision encoder leads to better performance than all baselines. Furthermore, we show that regularizing training with the distance between the fine-tuned and pretrained matrices adds reliability for adapting CLIP, making the results stable across different learning rates in the ''validation-free'' setting. This simple approach, coined ProLIP, yields state-of-the-art performance on 11 few-shot classification benchmarks, few-shot cross-dataset transfer, domain generalization, and base-to-new class generalization. We also show that ProLIP significantly outperforms prompt tuning when extended to another task of test-time adaptation, while being one order of magnitude faster to train. Code will be made available at: https://github.com/astra-vision/ProLIP .

Induced seismicity has emerged as a source of a significant earthquake hazard associated with recent development of unconventional energy resources. Therefore, it is imperative to develop stochastic models that can accurately describe the observed seismicity rate and forecast its evolution. In this study, a mechanism suggested by linear response theory is incorporated into a stochastic earthquake model to account for changes in the seismicity rate. It is derived that the induced rate can be modelled as a convolution of the forcing, related to fluid injection operations, and a specific response kernel. The model is incorporated into a Bayesian framework to compute the probabilities for the occurrence of the largest expected events during future time intervals. The applicability of the model is illustrated by analyzing the injection and seismicity data at the Geysers geothermal field in California. The suggested approach provides further insight into the probabilistic assessment of earthquake hazard associated with fluid injection operations. It also can be used for probing the rheological properties of the subsurface by analysing the inherent characteristic time-scales associated with the subsurface response to external forcing.

Precision agriculture relies heavily on effective weed management to ensure robust crop yields. This study presents RoWeeder, an innovative framework for unsupervised weed mapping that combines crop-row detection with a noise-resilient deep learning model. By leveraging crop-row information to create a pseudo-ground truth, our method trains a lightweight deep learning model capable of distinguishing between crops and weeds, even in the presence of noisy data. Evaluated on the WeedMap dataset, RoWeeder achieves an F1 score of 75.3, outperforming several baselines. Comprehensive ablation studies further validated the model's performance. By integrating RoWeeder with drone technology, farmers can conduct real-time aerial surveys, enabling precise weed management across large fields. The code is available at: \url{https://github.com/pasqualedem/RoWeeder}.

Despite the increasing effectiveness of language models, their reasoning capabilities remain underdeveloped. In particular, causal reasoning through counterfactual question answering is lacking. This work aims to bridge this gap. We first derive novel metrics that balance accuracy in factual and counterfactual questions, capturing a more complete view of the reasoning abilities of language models than traditional factual-only based metrics. Second, we propose several fine-tuning approaches that aim to elicit better reasoning mechanisms, in the sense of the proposed metrics. Finally, we evaluate the performance of the fine-tuned language models in a variety of realistic scenarios. In particular, we investigate to what extent our fine-tuning approaches systemically achieve better generalization with respect to the base models in several problems that require, among others, inductive and deductive reasoning capabilities.

Computer simulations of minimal population-dynamics models have long been used to explore questions in ecosystems coexistence and species biodiversity, via simple agent-based models of three interacting species, referred to as $R$, $P$, and $S$, and where individual agents compete with one another in predator/prey contests that are determined by the cyclic dominance rules of the Rock-Paper-Scissors game. Recent publications have explored the dynamics of five-species models, based on the Rock-Paper-Scissors-Lizard-Spock (RPSLS) game. A 2022 paper by Zhong et al. reported simulation studies of species coexistence in spatial RPSLS systems in which one or more directed edges are ablated from the five-vertex tournament digraph defining the RPSLS game: Zhong et al. showed that the ablation of a single inter-species interaction can lead to a collapse in biodiversity. In this paper I present first results from simulation studies of evolutionary spatial cyclic games where there are seven species, but where each species is still in its own local five-species RPSLS-like interaction network: the dominance networks I use for this are a subset of the $n$-node $k$-regular circulant digraphs $D(n,\Omega)$ for odd-numbered $n$ and $|\Omega|=2$. My results indicate that Zhong et al.'s results are due to the specific fully-connected tournament dominance network used in their RPSLS model: when other, equally realistic, RPSLS-style dominance networks are used instead, no such sudden collapse in biodiversity occurs. The Python source-code used for the work reported here is freely available on GitHub.

Aquaculture is expected to account for two-thirds of global fish consumption by 2030, highlighting the need for sustainable and efficient practices. Feeding is crucial to aquaculture success, influenced by factors like fish size, environment, and health. This study addresses a gap in feeding control for sea cages by developing a real-time monitoring system, using AI models and computer vision to analyze feeding behavior with European sea bass as pilot species. Key metrics like fish speed and a new feeding behavior index (FBI) were used to assess responses to different feeding scenarios. The results revealed distinct behavior patterns based on feeding quantity, with imbalances in activity when fish are overfed or underfed. The results can be used for predicting the level of satiation of the fish and controlling feeding duration.

The surge of pelagic \emph{Sargassum} in the Intra-America Seas, particularly the Caribbean Sea, since the early 2010s has raised significant ecological concerns. This study emphasizes the need for a mechanistic understanding of \emph{Sargassum} dynamics to elucidate the ecological impacts and uncertainties associated with blooms. By introducing a novel transport model, physical components such as ocean currents and winds are integrated with biological aspects affecting the \emph{Sargassum} life cycle, including reproduction, grounded in an enhanced Maxey--Riley theory for floating particles. Nonlinear elastic forces among the particles are included to simulate interactions within and among \emph{Sargassum} rafts. This promotes aggregation, consistent with observations, within oceanic eddies, which facilitate their transport. This cannot be achieved by the so-called leeway approach to transport, which forms the basis of current \emph{Sargassum} modeling. Using satellite-derived data, the model is validated, outperforming the leeway model. Publicly accessible codes are provided to support further research and ecosystem management efforts. This comprehensive approach is expected to improve predictive capabilities and management strategies regarding \emph{Sargassum} dynamics in affected regions, thus contributing to a deeper understanding of marine ecosystem dynamics and resilience.

Accurate semantic segmentation of remote sensing imagery is critical for various Earth observation applications, such as land cover mapping, urban planning, and environmental monitoring. However, individual data sources often present limitations for this task. Very High Resolution (VHR) aerial imagery provides rich spatial details but cannot capture temporal information about land cover changes. Conversely, Satellite Image Time Series (SITS) capture temporal dynamics, such as seasonal variations in vegetation, but with limited spatial resolution, making it difficult to distinguish fine-scale objects. This paper proposes a late fusion deep learning model (LF-DLM) for semantic segmentation that leverages the complementary strengths of both VHR aerial imagery and SITS. The proposed model consists of two independent deep learning branches. One branch integrates detailed textures from aerial imagery captured by UNetFormer with a Multi-Axis Vision Transformer (MaxViT) backbone. The other branch captures complex spatio-temporal dynamics from the Sentinel-2 satellite image time series using a U-Net with Temporal Attention Encoder (U-TAE). This approach leads to state-of-the-art results on the FLAIR dataset, a large-scale benchmark for land cover segmentation using multi-source optical imagery. The findings highlight the importance of multi-modality fusion in improving the accuracy and robustness of semantic segmentation in remote sensing applications.

We consider a sequential multi-task problem, where each task is modeled as the stochastic multi-armed bandit with K arms. We assume the bandit tasks are adjacently similar in the sense that the difference between the mean rewards of the arms for any two consecutive tasks is bounded by a parameter. We propose two algorithms (one assumes the parameter is known while the other does not) based on UCB to transfer reward samples from preceding tasks to improve the overall regret across all tasks. Our analysis shows that transferring samples reduces the regret as compared to the case of no transfer. We provide empirical results for our algorithms, which show performance improvement over the standard UCB algorithm without transfer and a naive transfer algorithm.

The increasing number of unmanned aerial vehicles (UAVs) in urban environments requires a strategy to minimize their environmental impact, both in terms of energy efficiency and noise reduction. In order to reduce these concerns, novel strategies for developing prediction models and optimization of flight planning, for instance through deep reinforcement learning (DRL), are needed. Our goal is to develop DRL algorithms capable of enabling the autonomous navigation of UAVs in urban environments, taking into account the presence of buildings and other UAVs, optimizing the trajectories in order to reduce both energetic consumption and noise. This is achieved using fluid-flow simulations which represent the environment in which UAVs navigate and training the UAV as an agent interacting with an urban environment. In this work, we consider a domain domain represented by a two-dimensional flow field with obstacles, ideally representing buildings, extracted from a three-dimensional high-fidelity numerical simulation. The presented methodology, using PPO+LSTM cells, was validated by reproducing a simple but fundamental problem in navigation, namely the Zermelo's problem, which deals with a vessel navigating in a turbulent flow, travelling from a starting point to a target location, optimizing the trajectory. The current method shows a significant improvement with respect to both a simple PPO and a TD3 algorithm, with a success rate (SR) of the PPO+LSTM trained policy of 98.7%, and a crash rate (CR) of 0.1%, outperforming both PPO (SR = 75.6%, CR=18.6%) and TD3 (SR=77.4% and CR=14.5%). This is the first step towards DRL strategies which will guide UAVs in a three-dimensional flow field using real-time signals, making the navigation efficient in terms of flight time and avoiding damages to the vehicle.

End-to-end autonomous driving offers a streamlined alternative to the traditional modular pipeline, integrating perception, prediction, and planning within a single framework. While Deep Reinforcement Learning (DRL) has recently gained traction in this domain, existing approaches often overlook the critical connection between feature extraction of DRL and perception. In this paper, we bridge this gap by mapping the DRL feature extraction network directly to the perception phase, enabling clearer interpretation through semantic segmentation. By leveraging Bird's-Eye-View (BEV) representations, we propose a novel DRL-based end-to-end driving framework that utilizes multi-sensor inputs to construct a unified three-dimensional understanding of the environment. This BEV-based system extracts and translates critical environmental features into high-level abstract states for DRL, facilitating more informed control. Extensive experimental evaluations demonstrate that our approach not only enhances interpretability but also significantly outperforms state-of-the-art methods in autonomous driving control tasks, reducing the collision rate by 20%.

In recent years, analysis of remote sensing data has benefited immensely from borrowing techniques from the broader field of computer vision, such as the use of shared models pre-trained on large and diverse datasets. However, satellite imagery has unique features that are not accounted for in traditional computer vision, such as the existence of multiple revisits of the same location. Here, we explore the best way to use revisits in the framework of fine-tuning pre-trained remote sensing models. We focus on an applied research question of relevance to climate change mitigation -- power substation segmentation -- that is representative of applied uses of pre-trained models more generally. Through extensive tests of different multi-temporal input schemes across diverse model architectures, we find that fusing representations from multiple revisits in the model latent space is superior to other methods of using revisits, including as a form of data augmentation. We also find that a SWIN Transformer-based architecture performs better than U-nets and ViT-based models. We verify the generality of our results on a separate building density estimation task.

Jamming devices disrupt signals from the global navigation satellite system (GNSS) and pose a significant threat, as they compromise the robustness of accurate positioning. The detection of anomalies within frequency snapshots is crucial to counteract these interferences effectively. A critical preliminary countermeasure involves the reliable classification of interferences and the characterization and localization of jamming devices. This paper introduces an extensive dataset comprising snapshots obtained from a low-frequency antenna that capture various generated interferences within a large-scale environment, including controlled multipath effects. Our objective is to assess the resilience of machine learning (ML) models against environmental changes, such as multipath effects, variations in interference attributes, such as interference class, bandwidth, and signal power, the accuracy jamming device localization, and the constraints imposed by snapshot input lengths. Furthermore, we evaluate the performance of a diverse set of 129 distinct vision encoder models across all tasks. By analyzing the aleatoric and epistemic uncertainties, we demonstrate the adaptability of our model in generalizing across diverse facets, thus establishing its suitability for real-world applications. Dataset: https://gitlab.cc-asp.fraunhofer.de/darcy_gnss/controlled_low_frequency

Earth system models inform water policy and interventions, but knowledge gaps in hydrologic representations limit the credibility of projections and impacts assessments. The literature does not provide conclusive evidence that incorporating higher resolutions, comprehensive process models, and latest parameterization schemes, will result in improvements. We compare hydroclimate representations and runoff projections across two generations of Coupled Modeling Intercomparison Project (CMIP) models, specifically, CMIP5 and CMIP6, with gridded runoff from Global Runoff Reconstruction (GRUN) and ECMWF Reanalysis V5 (ERA5) as benchmarks. Our results show that systematic embedding of the best available process models and parameterizations, together with finer resolutions, improve runoff projections with uncertainty characterizations in 30 of the largest rivers worldwide in a mechanistically explainable manner. The more skillful CMIP6 models suggest that, following the mid-range SSP370 emissions scenario, 40% of the rivers will exhibit decreased runoff by 2100, impacting 260 million people.

Accurate field boundary delineation is a critical challenge in digital agriculture, impacting everything from crop monitoring to resource management. Existing methods often struggle with noise and fail to generalize across varied landscapes, particularly when dealing with cloud cover in optical remote sensing. In response, this study presents a new approach that leverages time series data from Sentinel-2 (S2) and Sentinel-1 (S1) imagery to improve performance under diverse cloud conditions, without the need for manual cloud filtering. We introduce a 3D Vision Transformer architecture specifically designed for satellite image time series, incorporating a memory-efficient attention mechanism. Two models are proposed: PTAViT3D, which handles either S2 or S1 data independently, and PTAViT3D-CA, which fuses both datasets to enhance accuracy. Both models are evaluated under sparse and dense cloud coverage by exploiting spatio-temporal correlations. Our results demonstrate that the models can effectively delineate field boundaries, even with partial (S2 or S2 and S1 data fusion) or dense cloud cover (S1), with the S1-based model providing performance comparable to S2 imagery in terms of spatial resolution. A key strength of this approach lies in its capacity to directly process cloud-contaminated imagery by leveraging spatio-temporal correlations in a memory-efficient manner. This methodology, used in the ePaddocks product to map Australia's national field boundaries, offers a robust, scalable solution adaptable to varying agricultural environments, delivering precision and reliability where existing methods falter. Our code is available at https://github.com/feevos/tfcl.

Scent marks play a crucial role in both territorial and sexual communication in many species. We investigated how free-ranging dogs respond to scent marks from individuals of different identities in terms of sex and group, across varying strategic locations within their territory. Both male and female dogs showed heightened interest in scent marks compared to control, exhibiting stronger territorial responses,. with males being more territorial than females. Overmarking behaviour was predominantly observed in males, particularly in response to male scent marks and those from neighbouring groups. Behavioural cluster analysis revealed distinct responses to different scent marks, with neighbouring group male scents eliciting the most distinct reactions. Our findings highlight the multifaceted role of scent marks in free-ranging dog communication, mediating both territorial defence and intrasexual competition. The differential responses based on the identity and gender of the scent-marker emphasize the complexity of olfactory signalling in this species. This study contributes to understanding the social behaviour of dogs in their natural habitat, and opens up possibilities for future explorations in the role of olfactory cues in the social dynamics of the species.

Sedimentary basins beneath many Antarctic ice streams host substantial volumes of groundwater, which can be exchanged with a "shallow" subglacial hydrological system of till and channelised water. This exchange contributes substantially to basal water budgets, which in turn modulate the flow of ice streams. The geometry of these sedimentary basins is known to be complex, and the groundwater therein has been observed to vary in salinity due to historic seawater intrusion. However, little is known about the hydraulic properties of subglacial sedimentary basins, and the factors controlling groundwater exfiltration and infiltration. We develop a mathematical model for two-dimensional groundwater flow beneath a marine-terminating ice stream on geological timescales, taking into account the effect of seawater intrusion. We find that seawater may become "trapped" in subglacial sedimentary basins, through cycles of grounding line advance and retreat or through "pockets" arising from basin geometry. In addition, we estimate the sedimentary basin permeability which reproduces field observations of groundwater salinity profiles from beneath Whillans Ice Stream in West Antarctica. Exchange of groundwater with the shallow hydrological system is primarily controlled by basin geometry, with groundwater being exfiltrated where the basin becomes shallower and re-infiltrating where it becomes deeper. However, seawater intrusion also has non-negligible effects on this exchange.

Aviation's non-CO$_2$ effects, especially the impact of aviation-induced contrails, drive atmospheric changes and can influence climate dynamics. Although contrails are believed to contribute to global warming through their net warming effect, uncertainties persist due to the challenges in accurately measuring their radiative impacts. This study aims to address this knowledge gap by investigating the relationship between aviation-induced contrails, as observed in Meteosat Second Generation (MSG) satellite imagery, and their impact on radiative forcing (RF) over a two-week study. Results show that while daytime contrails generally have a cooling effect, the higher number of nighttime contrails results in a net warming effect over the entire day. Net RF values for detected contrails range approximately from -8 TW to 2.5 TW during the day and from 0 to 6 TW at night. Our findings also show a 41.03% increase in contrail coverage from January 24-30, 2023, to the same week in 2024, accompanied by a 128.7% rise in contrail radiative forcing (CRF), indicating greater warming from the added contrails. These findings highlight the necessity of considering temporal factors, such as the timing and duration of contrail formation, when assessing their overall warming impact. They also indicate a potential increase in contrail-induced warming from 2023 to 2024, attributable to the rise in contrail coverage. Further investigation into these trends is crucial for the development of effective mitigation strategies.

Satellite image time series (SITS) data provides continuous observations over time, allowing for the tracking of vegetation changes and growth patterns throughout the seasons and years. Numerous deep learning (DL) approaches using SITS for crop classification have emerged recently, with the latest approaches adopting Transformer for SITS classification. However, the quadratic complexity of self-attention in Transformer poses challenges for classifying long time series. While the cutting-edge Mamba architecture has demonstrated strength in various domains, including remote sensing image interpretation, its capacity to learn temporal representations in SITS data remains unexplored. Moreover, the existing SITS classification methods often depend solely on crop labels as supervision signals, which fails to fully exploit the temporal information. In this paper, we proposed a Satellite Image Time Series Mamba (SITSMamba) method for crop classification based on remote sensing time series data. The proposed SITSMamba contains a spatial encoder based on Convolutional Neural Networks (CNN) and a Mamba-based temporal encoder. To exploit richer temporal information from SITS, we design two branches of decoder used for different tasks. The first branch is a crop Classification Branch (CBranch), which includes a ConvBlock to decode the feature to a crop map. The second branch is a SITS Reconstruction Branch that uses a Linear layer to transform the encoded feature to predict the original input values. Furthermore, we design a Positional Weight (PW) applied to the RBranch to help the model learn rich latent knowledge from SITS. We also design two weighting factors to control the balance of the two branches during training. The code of SITSMamba is available at: https://github.com/XiaoleiQinn/SITSMamba.

In the past three years, there has been significant interest in hyperspectral imagery (HSI) classification using vision Transformers for analysis of remotely sensed data. Previous research predominantly focused on the empirical integration of convolutional neural networks (CNNs) to augment the network's capability to extract local feature information. Yet, the theoretical justification for vision Transformers out-performing CNN architectures in HSI classification remains a question. To address this issue, a unified hierarchical spectral vision Transformer architecture, specifically tailored for HSI classification, is investigated. In this streamlined yet effective vision Transformer architecture, multiple mixer modules are strategically integrated separately. These include the CNN-mixer, which executes convolution operations; the spatial self-attention (SSA)-mixer and channel self-attention (CSA)-mixer, both of which are adaptations of classical self-attention blocks; and hybrid models such as the SSA+CNN-mixer and CSA+CNN-mixer, which merge convolution with self-attention operations. This integration facilitates the development of a broad spectrum of vision Transformer-based models tailored for HSI classification. In terms of the training process, a comprehensive analysis is performed, contrasting classical CNN models and vision Transformer-based counterparts, with particular attention to disturbance robustness and the distribution of the largest eigenvalue of the Hessian. From the evaluations conducted on various mixer models rooted in the unified architecture, it is concluded that the unique strength of vision Transformers can be attributed to their overarching architecture, rather than being exclusively reliant on individual multi-head self-attention (MSA) components.

The Job Shop Scheduling Problem (JSP) is central to operations research, primarily optimizing energy efficiency due to its profound environmental and economic implications. Efficient scheduling enhances production metrics and mitigates energy consumption, thus effectively balancing productivity and sustainability objectives. Given the intricate and diverse nature of JSP instances, along with the array of algorithms developed to tackle these challenges, an intelligent algorithm selection tool becomes paramount. This paper introduces a framework designed to identify key problem features that characterize its complexity and guide the selection of suitable algorithms. Leveraging machine learning techniques, particularly XGBoost, the framework recommends optimal solvers such as GUROBI, CPLEX, and GECODE for efficient JSP scheduling. GUROBI excels with smaller instances, while GECODE demonstrates robust scalability for complex scenarios. The proposed algorithm selector achieves an accuracy of 84.51\% in recommending the best algorithm for solving new JSP instances, highlighting its efficacy in algorithm selection. By refining feature extraction methodologies, the framework aims to broaden its applicability across diverse JSP scenarios, thereby advancing efficiency and sustainability in manufacturing logistics.

Remote sensing (RS) change analysis is vital for monitoring Earth's dynamic processes by detecting alterations in images over time. Traditional change detection excels at identifying pixel-level changes but lacks the ability to contextualize these alterations. While recent advancements in change captioning offer natural language descriptions of changes, they do not support interactive, user-specific queries. To address these limitations, we introduce ChangeChat, the first bitemporal vision-language model (VLM) designed specifically for RS change analysis. ChangeChat utilizes multimodal instruction tuning, allowing it to handle complex queries such as change captioning, category-specific quantification, and change localization. To enhance the model's performance, we developed the ChangeChat-87k dataset, which was generated using a combination of rule-based methods and GPT-assisted techniques. Experiments show that ChangeChat offers a comprehensive, interactive solution for RS change analysis, achieving performance comparable to or even better than state-of-the-art (SOTA) methods on specific tasks, and significantly surpassing the latest general-domain model, GPT-4. Code and pre-trained weights are available at https://github.com/hanlinwu/ChangeChat.

This research introduces an advanced method for diagnosing diseases in sweet orange leaves by utilising advanced artificial intelligence models like YOLOv8 . Due to their significance as a vital agricultural product, sweet oranges encounter significant threats from a variety of diseases that harmfully affect both their yield and quality. Conventional methods for disease detection primarily depend on manual inspection which is ineffective and frequently leads to errors, resulting in delayed treatment and increased financial losses. In response to this challenge, the research utilized YOLOv8 , harnessing their proficiencies in detecting objects and analyzing images. YOLOv8 is recognized for its rapid and precise performance, while VIT is acknowledged for its detailed feature extraction abilities. Impressively, during both the training and validation stages, YOLOv8 exhibited a perfect accuracy of 80.4%, while VIT achieved an accuracy of 99.12%, showcasing their potential to transform disease detection in agriculture. The study comprehensively examined the practical challenges related to the implementation of AI technologies in agriculture, encompassing the computational demands and user accessibility, and offering viable solutions for broader usage. Moreover, it underscores the environmental considerations, particularly the potential for reduced pesticide usage, thereby promoting sustainable farming and environmental conservation. These findings provide encouraging insights into the application of AI in agriculture, suggesting a transition towards more effective, sustainable, and technologically advanced farming methods. This research not only highlights the efficacy of YOLOv8 within a specific agricultural domain but also lays the foundation for further studies that encompass a broader application in crop management and sustainable agricultural practices.

The analysis of urban seismic signals offers valuable insights into urban environments and society. Yet, accurate detection and localization of seismic sources on a city-wide scale with conventional seismographic network is unavailable due to the prohibitive costs of ultra-dense seismic arrays required for imaging high-frequency anthropogenic sources. Here, we leverage existing fiber-optic networks as a distributed acoustic sensing system to accurately locate urban seismic sources and estimate how their intensity varies over time. By repurposing a 50-kilometer telecommunication fiber into an ultra-dense seismic array, we generate spatiotemporal maps of seismic source power (SSP) across San Jose, California. Our approach overcomes the proximity limitations of urban seismic sensing, enabling accurate localization of remote seismic sources generated by urban activities, such as traffic, construction, and school operations. We also show strong correlations between SSP values and environmental noise levels, as well as various persistent urban features, including land use patterns and demographics.

Various Earth anomalies have destroyed the stable, balanced state, resulting in fatalities and serious destruction of property. With the advantages of large-scale and precise observation, high-resolution remote sensing images have been widely used for anomaly monitoring and localization. Powered by the deep representation, the existing methods have achieved remarkable advances, primarily in classification and change detection techniques. However, labeled samples are difficult to acquire due to the low probability of anomaly occurrence, and the trained models are limited to fixed anomaly categories, which hinders the application for anomalies with few samples or unknown anomalies. In this paper, to tackle this problem, we propose the anomaly change detection (AnomalyCD) technique, which accepts time-series observations and learns to identify anomalous changes by learning from the historical normal change pattern. Compared to the existing techniques, AnomalyCD processes an unfixed number of time steps and can localize the various anomalies in a unified manner, without human supervision. To benchmark AnomalyCD, we constructed a high-resolution dataset with time-series images dedicated to various Earth anomalies (the AnomalyCDD dataset). AnomalyCDD contains high-resolution (from 0.15 to 2.39 m/pixel), time-series (from 3 to 7 time steps), and large-scale images (1927.93 km2 in total) collected globally Furthermore, we developed a zero-shot baseline model (AnomalyCDM), which implements the AnomalyCD technique by extracting a general representation from the segment anything model (SAM) and conducting temporal comparison to distinguish the anomalous changes from normal changes. AnomalyCDM is designed as a two-stage workflow to enhance the efficiency, and has the ability to process the unseen images directly, without retraining for each scene.

Accurate prediction of global sea surface temperature at sub-seasonal to seasonal (S2S) timescale is critical for drought and flood forecasting, as well as for improving disaster preparedness in human society. Government departments or academic studies normally use physics-based numerical models to predict S2S sea surface temperature and corresponding climate indices, such as El Ni\~no-Southern Oscillation. However, these models are hampered by computational inefficiencies, limited retention of ocean-atmosphere initial conditions, and significant uncertainty and biases. Here, we introduce a novel three-dimensional deep learning neural network to model the nonlinear and complex coupled atmosphere-ocean weather systems. This model incorporates climatic and temporal features and employs a self-attention mechanism to enhance the prediction of global S2S sea surface temperature pattern. Compared to the physics-based models, it shows significant computational efficiency and predictive capability, improving one to three months sea surface temperature predictive skill by 13.7% to 77.1% in seven ocean regions with dominant influence on S2S variability over land. This achievement underscores the significant potential of deep learning for largely improving forecasting skills at the S2S scale over land.

Monitoring carbon dioxide (CO2) injected and stored in subsurface reservoirs is critical for avoiding failure scenarios and enables real-time optimization of CO2 injection rates. Sequential Bayesian data assimilation (DA) is a statistical method for combining information over time from multiple sources to estimate a hidden state, such as the spread of the subsurface CO2 plume. An example of scalable and efficient sequential Bayesian DA is the ensemble Kalman filter (EnKF). We improve upon existing DA literature in the seismic-CO2 monitoring domain by applying this scalable DA algorithm to a high-dimensional CO2 reservoir using two-phase flow dynamics and time-lapse full waveform seismic data with a realistic surface-seismic survey design. We show more accurate estimates of the CO2 saturation field using the EnKF compared to using either the seismic data or the fluid physics alone. Furthermore, we test a range of values for the EnKF hyperparameters and give guidance on their selection for seismic CO2 reservoir monitoring.

Preparing to rapidly respond to emerging infectious diseases is becoming ever more critical. "SpillOver: Viral Risk Ranking" is an open-source tool developed to evaluate novel wildlife-origin viruses for their risk of spillover from animals to humans and their risk of spreading in human populations. However, several of the factors used in the risk assessment are dependent on evidence of previous zoonotic spillover and/or sustained transmission in humans. Therefore, we performed a reanalysis of the "Ranking Comparison" after removing eight factors that require post-spillover knowledge and compared the adjusted risk rankings to the originals. The top 10 viruses as ranked by their adjusted scores also had very high original scores. However, the predictive power of the tool for whether a virus was a human virus or not as classified in the Spillover database deteriorated when these eight factors were removed. The area under the receiver operating characteristic curves (AUROC) for the original score, 0.94, decreased to 0.73 for the adjusted scores. Furthermore, we compared the mean and standard deviation of the human and non-human viruses at the factor level. Most of the excluded spillover-dependent factors had dissimilar means between the human and non-human virus groups compared to the non-spillover dependent factors, which frequently demonstrated similar means between the two groups with some exceptions. We concluded that the original formulation of the tool depended heavily on spillover-dependent factors to "predict" the risk of zoonotic spillover for a novel virus. Future iterations of the tool should take into consideration other non-spillover dependent factors and omit those that are spillover-dependent to ensure the tool is fit for purpose.

This study investigates the trophic cascades and habitat suitability in Udanti Sitnadi Tiger Reserve (USTR), highlighting the roles of apex predators, subordinate predators, and prey species in maintaining ecosystem balance. Using the Trophic Species Distribution Model (SDM), we explored prey-predator interactions and habitat suitability, revealing that tigers, due to prey depletion, increasingly rely on cattle, while leopards adapt by preying on smaller species. The study emphasizes the need for prey augmentation and habitat restoration to support apex predators. Additionally, climate change projections for 2021-2040 and 2081-2100 under CMIP6 scenarios SSP245 and SSP585 indicate significant regional habitat shifts, necessitating adaptive management strategies. Kuladighat is projected to face habitat contraction, while Sitanadi may experience habitat expansion. Effective conservation efforts such as habitat restoration, prey augmentation and predator recovery are the most important steps needed to maintain the purpose of a Tiger reserve and conservation potential of Udanti-Sonabeda Tiger Conservation Unit (TCU). To achieve these dynamics, focusing on community participation, anti-poaching measures, and scientific recommendations are the most crucial components to focus on. This comprehensive analysis underscores the critical role of targeted conservation activities in prey-depleted landscapes to ensure the long-term survival of tigers and the overall health of forest ecosystems, enhancing biodiversity and mitigating human-wildlife conflicts in USTR.

Crop phenology determines crop growth stages and is valuable information for decision makers to plant and adapt agricultural management strategies to enhance food security. In the era of big Earth observation data ubiquity, attempts have been made to accurately predict crop phenology based on Remote Sensing (RS) data. However, most studies either focused on large scale interpretations of phenology or developed methods which are not adequate to help crop modeler communities on leveraging the value of RS data evaluated using more accurate and confident methods. Here, we estimate phenological developments for eight major crops and 13 phenological stages across Germany at 30m scale using a novel framework which fuses Landsat and Sentinel 2 (Harmonized Landsat and Sentinel data base; HLS) and radar of Sentinel 1 with a Machine Learning (ML) model. We proposed a thorough feature fusion analysis to find the best combinations of RS data on detecting phenological developments based on the national phenology network of Germany (German Meteorological Service; DWD) between 2017 and 2021. The nation-wide predicted crop phenology at 30 m resolution showed a very high precision of R2 > 0.9 and a very low Mean Absolute Error (MAE) < 2 (days). These results indicate that our fusing strategy of optical and radar datasets is highly performant with an accuracy highly relevant for practical applications, too. The subsequent uncertainty analysis indicated that fusing optical and radar data increases the reliability of the RS predicted crop growth stages. These improvements are expected to be useful for crop model calibrations and evaluations, facilitate informed agricultural decisions, and contribute to sustainable food production to address the increasing global food demand.

Understanding energetic electron precipitation is crucial for accurate space weather modeling and forecasting, impacting the Earth's upper atmosphere and human infrastructure. This study presents a low-cost, low-mass, and low-power solution for high-fidelity analysis of electron precipitation events by measuring the resulting bremsstrahlung X-ray emissions. Specifically, we report on results from the flight of a radiation detector payload based on a silicon pixel read-out Timepix detector technology, and its successful utilization onboard a `burster' weather balloon. We launched this payload during the May 2024 superstorm, capturing high-resolution measurements of both background galactic cosmic ray radiation as well as storm-time energetic electron precipitation. We further developed particle and radiation detection algorithms to separate bremsstrahlung X-rays from other particle species in the pixel-resolved trajectories as seen in the Timepix detector. The measurements revealed a distinctive four-peak structure in X-ray flux, corresponding to periodic four-minute-long bursts of energetic electron precipitation between 21:20 and 21:40 UT. This precipitation was also observed by a riometer station close to the balloon launch path, further validating balloon measurements and the developed X-ray identification algorithm. The clear periodic structure of the measured precipitation is likely caused by modulation of the electron losses from the radiation belt by harmonic Pc5 ULF waves, observed contemporaneously on the ground. The study underscores the potential of compact, low-cost payloads for advancing our understanding of space weather. Specifically, we envision a potential use of such Timepix-based detectors in space science, for example on sounding rockets or nano-, micro-, and small satellite platforms.

Effective water quality monitoring in coastal regions is crucial due to the progressive deterioration caused by pollution and human activities. To address this, this study develops time-series models to predict chlorophyll-a (Chl-a), suspended solids (SS), and turbidity using Sentinel-2 satellite data and Google Earth Engine (GEE) in the coastal regions of Hong Kong. Leveraging Long Short-Term Memory (LSTM) Recurrent Neural Networks, the study incorporates extensive temporal datasets to enhance prediction accuracy. The models utilize spectral data from Sentinel-2, focusing on optically active components, and demonstrate that selected variables closely align with the spectral characteristics of Chl-a and SS. The results indicate improved predictive performance over previous methods, highlighting the potential for remote sensing technology in continuous and comprehensive water quality assessment.

Crop field detection is a critical component of precision agriculture, essential for optimizing resource allocation and enhancing agricultural productivity. This study introduces KonvLiNA, a novel framework that integrates Convolutional Kolmogorov-Arnold Networks (cKAN) with Nystr\"om attention mechanisms for effective crop field detection. Leveraging KAN adaptive activation functions and the efficiency of Nystr\"om attention in handling largescale data, KonvLiNA significantly enhances feature extraction, enabling the model to capture intricate patterns in complex agricultural environments. Experimental results on rice crop dataset demonstrate KonvLiNA superiority over state-of-the-art methods, achieving a 0.415 AP and 0.459 AR with the Swin-L backbone, outperforming traditional YOLOv8 by significant margins. Additionally, evaluation on the COCO dataset showcases competitive performance across small, medium, and large objects, highlighting KonvLiNA efficacy in diverse agricultural settings. This work highlights the potential of hybrid KAN and attention mechanisms for advancing precision agriculture through improved crop field detection and management.

In this paper, we detail the high-performance implementation of our spaceborne radar simulator for satellite oceanography. Our software simulates the sea surface and the signal to imitate, as far as possible, the measurement process, starting from its lowest level mechanisms. In this perspective, raw data are computed as the sum of many illuminated scatterers, whose time-evolving properties are related to the surface roughness, topography, and kinematics. To achieve efficient performance, we intensively use GPU computing. Moreover, we propose a fast simulation mode based on the assumption that the instantaneous Doppler spectrum within a range gate varies on a timescale significantly larger than the PRI. The sea surface can then be updated at a frequency much smaller than the PRF, drastically reducing the computational cost. When the surface is updated, Doppler spectra are computed for all range gates. Signals segments are then obtained through 1D inverse Fourier transforms and pondered to ensure a smooth time evolution between surface updates. We validate this fast simulation mode with a radar altimeter simulation case of the Sentinel-3 SRAL instrument, showing that simulated raw data can be focused and retrieved using state-of-the-art algorithms. Finally, we show that, using a modest hardware configuration, our simulator can generate enough data in one day to compute the SWH and SSH spectra of a scene. This demonstrate that we achieved an important state-of-the-art speed-up.

Barchans are crescent-shape dunes ubiquitous on Earth and other celestial bodies, which are organized in barchan fields where they interact with each other. Over the last decades, satellite images have been largely employed to detect barchans on Earth and on the surface of Mars, with AI (Artificial Intelligence) becoming an important tool for monitoring those bedforms. However, automatic detection reported in previous works is limited to isolated dunes and does not identify successfully groups of interacting barchans. In this paper, we inquire into the automatic detection and tracking of barchans by carrying out experiments and exploring the acquired images using AI. After training a neural network with images from controlled experiments where complex interactions took place between dunes, we did the same for satellite images from Earth and Mars. We show, for the first time, that a neural network trained properly can identify and track barchans interacting with each other in different environments, using different image types (contrasts, colors, points of view, resolutions, etc.), with confidence scores (accuracy) above 70%. Our results represent a step further for automatically monitoring barchans, with important applications for human activities on Earth, Mars and other celestial bodies.

Zero-shot coordination (ZSC) remains a major challenge in the cooperative AI field, which aims to learn an agent to cooperate with an unseen partner in training environments or even novel environments. In recent years, a popular ZSC solution paradigm has been deep reinforcement learning (DRL) combined with advanced self-play or population-based methods to enhance the neural policy's ability to handle unseen partners. Despite some success, these approaches usually rely on black-box neural networks as the policy function. However, neural networks typically lack interpretability and logic, making the learned policies difficult for partners (e.g., humans) to understand and limiting their generalization ability. These shortcomings hinder the application of reinforcement learning methods in diverse cooperative scenarios.We suggest to represent the agent's policy with an interpretable program. Unlike neural networks, programs contain stable logic, but they are non-differentiable and difficult to optimize.To automatically learn such programs, we introduce Knowledge-driven Programmatic reinforcement learning for zero-shot Coordination (KnowPC). We first define a foundational Domain-Specific Language (DSL), including program structures, conditional primitives, and action primitives. A significant challenge is the vast program search space, making it difficult to find high-performing programs efficiently. To address this, KnowPC integrates an extractor and an reasoner. The extractor discovers environmental transition knowledge from multi-agent interaction trajectories, while the reasoner deduces the preconditions of each action primitive based on the transition knowledge.