# earth Published articles for earth. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Advanced Global Wildfire Activity Modeling with Hierarchical Graph ODE DevFeed: [Advanced Global Wildfire Activity Modeling with Hierarchical Graph ODE]() Original publisher: [Read original article]() Author: Fan Xu, Wei Gong, Hao Wu, Lilan Peng, Nan Wang, Qingsong Wen, Xian Wu, Kun Wang, Xibin Zhao Published: 2026-11-09T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Artificial Intelligence](), [graph theory](), [Simulation]() Tags: [adaptive](), [behavior](), [consistency](), [continuous](), [cs-lg](), [cube](), [dataset](), [deep-learning](), [earth](), [experiments](), [feature](), [flow](), [forecasting](), [framework](), [fusion](), [global](), [gnn](), [graph](), [hierarchy](), [modeling](), [wildfire]() ### AI overview HiGO models global wildfire activity as continuous, multi-scale dynamics on a hierarchy of Earth-system graphs. Adaptive message passing moves information within and across graph levels, while GNN-parameterized Neural ODE modules learn dynamics at multiple scales. On the SeasFire Cube dataset, the authors report that HiGO outperforms state-of-the-art baselines in long-range forecasting and that its continuous-time predictions are strongly consistent with observations. The excerpt gives no quantitative results or details of the baselines. ### Source excerpt arXiv:2601.01501v2 Announce Type: replace Abstract: Wildfires, as an integral component of the Earth system, are governed by a complex interplay of atmospheric, oceanic, and terrestrial processes spanning a vast range of spatiotemporal scales. Modeling their global activity on large timescales is therefore a critical yet challenging task. While deep learning has recently achieved significant breakthroughs in global weather forecasting, its potential for global wildfire behavior prediction remains underexplored. In this work, we reframe this problem and introduce the Hierarchical Graph ODE (HiGO), a novel framework designed to learn the multi-scale, continuous-time dynamics of wildfires. Specifically, we represent the Earth system as a multi-level graph hierarchy and propose an adaptive filtering message passing mechanism for both intra- and inter-level information flow, enabling more effective feature extraction and fusion. Furthermore, we incorporate GNN-parameterized Neural ODE modules at multiple levels to explicitly learn the continuous dynamics inherent to each scale. Through extensive experiments on the SeasFire Cube dataset, we demonstrate that HiGO significantly outperforms state-of-the-art baselines on long-range wildfire forecasting. Moreover, its continuous-time predictions exhibit strong observational consistency, highlighting its potential for real-world applications. ## PROSWIN: Probabilistic Solar Wind Speed Forecasting Using Deep Distributional Regression From Solar Images DevFeed: [PROSWIN: Probabilistic Solar Wind Speed Forecasting Using Deep Distributional Regression From Solar Images]() Original publisher: [Read original article]() Author: Daniel Collin, Yuri Shprits, Luca Chiarabini, Stefan J. Hofmeister, Nadja Klein, Guillermo Gallego Published: 2026-11-06T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Machine learning](), [Artificial Intelligence](), [Deep neural networks](), [Neural Network](), [Image](), [multimodal-ai]() Tags: [algorithm](), [astro-ph-sr](), [channel](), [combines](), [continuous](), [cs-lg](), [data](), [deep](), [earth](), [error](), [fast](), [forecast](), [forecasting](), [images](), [metric](), [metrics](), [model](), [model-selection](), [physics-data-an](), [physics-space-ph](), [probabilistic](), [regression](), [solar](), [speed]() ### AI overview PROSWIN forecasts hourly solar wind speed at Earth four days ahead and represents uncertainty through probabilistic predictions. It combines solar images and magnetograms with a deep neural network and distributional regression. To account for high-speed stream peaks as well as overall timeline accuracy, the authors introduce a prediction score for model selection. On 14 years of data, reported uncertainty calibration deviates by less than 1% on average; timeline CRPS is 41.0 km/s, HSS peak CRPS 45.3 km/s, and prediction score 42.3 km/s. The 171 Å image channel complements the commonly used 193 Å and 211 Å channels. In comparisons with selected literature models, the authors report accuracy for both timeline and HSS peaks, avoiding a tradeoff between them. These findings come from the supplied abstract. ### Source excerpt arXiv:2609.26683v2 Announce Type: replace-cross Abstract: Accurately predicting fast solar wind conditions is challenging, as uncertainties are large and unquantified by traditional single-value prediction models. In particular, the risks of high-speed solar wind streams (HSSs), which can cause damage to technological infrastructure, cannot be reliably assessed without probabilistic forecasts. We present PROSWIN, a probabilistic machine learning model that forecasts the hourly solar wind speed (SWS) at Earth with a four-day lead time. The approach combines solar images and magnetograms using a deep neural network coupled to a distributional regression algorithm. Because standard error metrics underweight the relevance of HSS peaks, we further introduce the prediction score, a model-selection metric that jointly rewards timeline and HSS peak accuracy. On 14 years of data, our forecast achieves very well-calibrated uncertainties (<1% average deviation). Using the continuous ranked probability score (CRPS), a metric that assesses distributional accuracy, we obtain a timeline CRPS of 41.0 km/s, an HSS peak CRPS of 45.3 km/s, and a prediction score of 42.3 km/s. We find that the 171 {\AA} channel is an important complement to the typically used 193 {\AA} and 211 {\AA} channels and that the prediction score for model selection improves the applicability of the model. Compared to selected models from the literature, ours is the only one that is accurate for both timeline and HSS peak values, rather than trading one off against the other. These results support the advantages of probabilistic over single-value solar wind models. The introduced methods are also transferable to other forecasting problems. ## Utilizing Earth Foundation Models to Enhance the Simulation Performance of Hydrological Models with AlphaEarth Embeddings DevFeed: [Utilizing Earth Foundation Models to Enhance the Simulation Performance of Hydrological Models with AlphaEarth Embeddings]() Original publisher: [Read original article]() Author: Pengfei Qu, Lei Ye, Wenyu Ouyang, Chi Zhang, Yikai Chai, Shuolong Xu, Yongri Piao, Miao Zhang, Huchuan Lu Published: 2026-11-05T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [Embeddings](), [AI Foundation Models](), [Representation learning](), [Predictive AI](), [geospatial](), [Simulation]() Tags: [abstract](), [advanced](), [analysis](), [arxiv](), [based](), [behavior](), [but](), [cluster](), [cross](), [cs-ai](), [cs-lg](), [data](), [dataset](), [deep](), [earth](), [embeddings](), [foundation](), [foundation-models](), [models](), [performance](), [simulation](), [with]() ### AI overview Using the CAMELS dataset, the study evaluates 64-dimensional AlphaEarth Foundations (AEF) embeddings as static basin descriptors for streamflow prediction. AEF embeddings had nearly the same in-sample performance as conventional CAMELS attributes but greater out-of-sample robustness, and generally improved prediction across 100 target basins in ungauged-basin experiments. With AEF inputs, similarity-based donor selection beat random selection for smaller donor pools; that advantage faded as pools grew, and adding donors did not consistently help. Leave-one-cluster-out results showed that benefits varied with basin partitioning, so transferability remained sensitive to which catchment conditions were represented during training. The embeddings were statistically associated with conventional terrain, vegetation, and rooting attributes. ### Source excerpt arXiv:2601.01558v3 Announce Type: replace-cross Abstract: Accurate prediction in ungauged basins (PUB) requires effective transfer across hydrologically diverse regions. While deep learning has advanced large-sample hydrological modeling, its cross-regional generalization depends in part on the static basin descriptors used to guide model behavior. Earth Foundation Models (EFMs) provide a promising alternative by learning environmental representations from multi-modal satellite data. In this study, we evaluate the 64-dimensional AlphaEarth Foundations (AEF) embeddings as a new form of static basin representation and assess their value for streamflow prediction under both in-sample (IS) and out-of-sample (OOS) settings using the CAMELS dataset. Results show that AEF embeddings offer higher OOS robustness than CAMELS attributes, despite nearly identical IS performance. Mutual-information analysis reveals statistical associations between AEF embeddings and conventional catchment attributes related to terrain, vegetation, and rooting characteristics. We further investigate similarity-based donor-basin selection for PUB using CAMELS attributes and AEF embeddings. Across 100 target basins, AEF static inputs generally improved PUB performance. With AEF inputs, similarity-based donor selection outperformed random selection for smaller donor pools, but its advantage diminished as pool size increased. Adding more donor basins did not consistently improve performance. Leave-one-cluster-out experiments showed AEF input benefits depended on basin partitioning, suggesting that transferability remains sensitive to the catchment conditions represented during training. Overall, this study demonstrates that Earth Foundation Model embeddings provide a transferable representation for PUB prediction and underscores the importance of basin representation quality and donor pool design, motivating future research toward unified hydrological foundation models. ## StarWorld: A Cross-Layer World Model for Starlink Prediction and Streaming DevFeed: [StarWorld: A Cross-Layer World Model for Starlink Prediction and Streaming]() Original publisher: [Read original article]() Author: Bowen Zhang, Pengcheng Luo, Hongjun Xie, Shujin Jia, Genke Yang, Jian Chu Published: 2026-11-05T05:00:00Z Content type: article Language: en Sources: [arXiv: Networking]() Topics: [World models](), [Streaming]() Tags: [adaptive](), [broadband](), [control](), [controller](), [cross](), [cs-ni](), [earth](), [effect](), [environment](), [error](), [experience](), [future](), [futures](), [gap](), [layer](), [model](), [prediction](), [starlink](), [streaming](), [world](), [world-model]() ### AI overview StarWorld models uncertain Starlink access across network prediction and adaptive bitrate streaming by combining learned stochastic network dynamics with executable prediction, control, and playback processes to generate multi-step futures. A calibrated Trust Gate assesses whether those futures are reliable enough to use. On real Starlink measurements, it had the lowest network-future distributional errors among the compared methods at the principal horizons and reduced pairwise controller QoE effect error by 76%. The supplied abstract does not specify the comparison methods or the Trust Gate's calibration details. ### Source excerpt arXiv:2611.05832v1 Announce Type: new Abstract: Starlink and other low Earth orbit (LEO) satellite broadband systems exhibit stochastic access dynamics whose effects propagate from network measurements to throughput prediction and adaptive bitrate (ABR) streaming, making reasoning over multiple plausible futures essential. Existing studies mainly characterize realized traces or individual system layers, leaving cross-layer world modeling of these evolving dynamics largely unexplored. To address this gap, we propose StarWorld, a cross-layer world model for Starlink prediction and streaming. StarWorld couples learned stochastic network dynamics with executable prediction, control, and playback processes to imagine multi-step cross-layer futures, while a calibrated Trust Gate determines whether these futures are sufficiently reliable for use. On real Starlink measurements, StarWorld achieves the lowest network-future distributional errors among the comparison methods at the principal horizons and reduces pairwise controller quality-of-experience (QoE) effect error by 76%. These results support StarWorld as a cross-layer imagination environment for studying prediction and streaming outcomes under uncertain Starlink access conditions. ## Large Language Model-Assisted Adaptive Heuristic Search for Agile Earth Observation Satellite Scheduling DevFeed: [Large Language Model-Assisted Adaptive Heuristic Search for Agile Earth Observation Satellite Scheduling]() Original publisher: [Read original article]() Author: He Wang, Yeye Liu, Ruijie Li, Xuemei Jiang, Junyu Wu, Lining Liu Zhuoqi Man, Tao Li, Yanjie Song Published: 2026-11-04T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [scheduling](), [Large Language Model]() Tags: [adaptive](), [agile](), [cs-ai](), [cs-cl](), [earth](), [earth-observation-satellites](), [energy](), [experiments](), [imaging](), [language](), [language-model](), [large-language-model](), [llm](), [satellite](), [scheduling](), [search]() ### AI overview The method combines an LLM with ant colony search to schedule agile satellite imaging under observation-window, maneuvering, and energy constraints. The LLM creates an instance-specific initial task-selection heuristic; search performance and scheduling information then feed back into the LLM to refine it. Proposed changes are tested through paired solution constructions before entering the search. A unified procedure handles task selection, satellite assignment, and observation timing. Across six scenarios with 50-200 imaging tasks, the method achieved 4.3%-9.6% higher mean observation benefit than the strongest baseline. ### Source excerpt arXiv:2611.04458v1 Announce Type: new Abstract: Agile Earth observation satellites must coordinate imaging tasks within limited observation windows, while accounting for attitude maneuvering and onboard energy constraints. Competition among observation opportunities makes scheduling performance sensitive to the heuristic rules used for task selection. This letter proposes a large language model (LLM)-assisted adaptive heuristic search method for agile Earth observation satellite scheduling. The LLM generates an initial heuristic rule according to the characteristics of the scheduling instance. During the subsequent ant colony search, search performance and scheduling information are fed back to the LLM to refine the heuristic rule. The proposed modifications are evaluated through paired solution constructions before being incorporated into the search process. A unified scheduling procedure determines task selection, satellite assignment, and observation timing under the operational constraints. Experiments on six scheduling scenarios involving 50--200 imaging tasks demonstrate that the proposed method achieves 4.3\%--9.6\% higher mean observation benefit than the strongest baseline. ## OSGR: Observation-Support Graph Reasoning for Earth Observation Question Answering DevFeed: [OSGR: Observation-Support Graph Reasoning for Earth Observation Question Answering]() Original publisher: [Read original article]() Author: Dazhou Yu, Zerong Wang, Riyang Bao, Longxiang Li, Yang Liu, Liang Zhao Published: 2026-11-04T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [llm-reasoning](), [Earth AI](), [Machine learning]() Tags: [agents](), [cs-ai](), [cs-ir](), [earth](), [framework](), [graph](), [grounding](), [kernels](), [llm](), [offline](), [online](), [over](), [queries](), [question](), [ratio](), [reasoning](), [retrieval](), [search](), [space](), [stage](), [target](), [under]() ### AI overview OSGR addresses Earth observation question answering by separating concept-level support-path deduction from the grounding of specific observation items. Its offline Observation-Support Graph identifies valid target-to-result paths; online Relation-Conditioned Grounding uses spatiotemporal compatibility kernels to select grounded item sets. On Earth-Bench and DisasterM3-Reason, the reported GPT-4o accuracy rises from 49.3% to 63.7%, while the execution-step ratio falls from 2.9 to 2.5. ### Source excerpt arXiv:2611.04301v1 Announce Type: new Abstract: Answering geographic queries over Earth observation (EO) archives requires solving a joint retrieval-and-derivation problem under strict spatiotemporal constraints. Existing LLM agents fail to ground observation items or suffer from search-space explosion. We propose OSGR, a two-stage framework that factorizes this search into offline symbolic concept deduction and online Relation-Conditioned Grounding. OSGR leverages an offline Observation-Support Graph to deduce valid target-to-result support paths at the concept level, and applies spatiotemporal compatibility kernels to instantiate grounded observation-item sets. On Earth-Bench and DisasterM3-Reason, OSGR improves GPT-4o accuracy from 49.3\% to 63.7\% while reducing the execution-step ratio from 2.9 to 2.5. ## Fusing Earth observation embeddings and spaceborne terrain with meteorological forcing for flood peak depth prediction from high-water marks DevFeed: [Fusing Earth observation embeddings and spaceborne terrain with meteorological forcing for flood peak depth prediction from high-water marks]() Original publisher: [Read original article]() Author: Wen Zhou, Shaowen Wang, Upmanu Lall, Venkatesh Merwade, Dingqi Ye Published: 2026-11-04T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Embeddings](), [Neural Network](), [AI Foundation Models](), [Estimator]() Tags: [absolute](), [cs-lg](), [deep-learning](), [dem](), [earth](), [embeddings](), [field](), [foundation-model](), [fusion](), [geospatial](), [lead-time](), [model](), [network](), [prediction](), [satellite](), [sensors](), [surveys](), [time](), [vertical](), [water]() ### AI overview A cross-modal model trained on 5,689 USGS high-water-mark surveys predicts flood peak depth at surveyed locations with 0.39 m mean absolute error six hours before the peak. It combines Copernicus GLO-30 terrain, AlphaEarth embeddings, and hourly meteorological forcing limited to the forecast lead time; forcing accounts for 77% of the attributed error reduction, while terrain and embeddings contribute equally to the remainder. At the same marks, depth estimated by differencing measured peak water levels with the DEM has 0.93 m error, suggesting surveyed-depth supervision can implicitly correct DEM error. The result is limited to sampled events and regions: on 11 held-out events the model distinguishes flooded from rain-free conditions but does not resolve depth well (R² = 0.03), leaving transfer to unseen storms unresolved. ### Source excerpt arXiv:2611.05220v1 Announce Type: new Abstract: Flood peak depth on normally dry land governs damage and risk to life, but is observed only sparsely, by post-event high-water mark (HWM) surveys, and depth derived from satellite flood extent is produced after the event and inherits the vertical error of the digital elevation model (DEM) it is computed from. Deep-learning flood models have in turn been trained on hydrodynamic simulations rather than on observed depths. We ask whether flood peak depth can be learned from field observations and predicted before the peak from spaceborne terrain, geospatial foundation-model embeddings and antecedent meteorological forcing. We assembled 5689 USGS HWM surveys from 39 events (2011-2025) in 24 U.S. states and Puerto Rico, assigned each a peak time from nearby sensors, paired each with a rain-free window at the same site, and linked them to Copernicus GLO-30 terrain, AlphaEarth embeddings and 144 h of hourly forcing masked to the lead time. A cross-modal fusion network trained on these observations predicts depth at the surveyed marks with a mean absolute error of 0.39 m 6 h ahead of the peak. Supervision by surveyed depth corrects the DEM error implicitly: differencing the measured peak water level with the same DEM gives 0.93 m at the same marks, 2.4 times larger, although it uses the measured water level and the model does not. Forcing accounts for 77% of the attributed reduction in depth error; terrain and the embeddings are statistically equivalent to each other and contribute the remaining 23%. On eleven held-out events the model still separates flooded from rain-free conditions but no longer resolves depth at the marks (R2 = 0.03), so the skill reported here is confined to the events and regions sampled. Field-surveyed depths can thus supervise flood depth prediction and correct terrain error that extent-based estimation cannot, while transfer to unseen storms remains the open problem. ## Arnis turns any place on Earth into a Minecraft world DevFeed: [Arnis turns any place on Earth into a Minecraft world]() Original publisher: [Read original article]() Author: Alex Perrault Published: 2026-11-03T16:00:00Z Content type: article Language: en Sources: [GitHub Blog]() Topics: [geospatial](), [Code review](), [Pull Request](), [Tool](), [Test-driven development]() Tags: [2022](), [3d](), [agents](), [ai](), [app](), [blocks](), [bug](), [bugs](), [city](), [classroom](), [code](), [code-reviews](), [coding](), [copilot](), [earth](), [github](), [github-copilot](), [github-copilot-app](), [maintainers](), [minecraft](), [open-source](), [open-source-community](), [prototype](), [world]() ### AI overview Arnis turns selected real-world locations into Minecraft worlds using mapped data from OpenStreetMap and other sources. It places mapped elements according to their coordinates and types, so unmapped features cannot appear; detailed landmarks can instead be built from 3D models. Users choose a place and spawn point, start generation, then can add missing map features to OpenStreetMap. The project began as a Python prototype and was later ported to Rust after open GitHub discussions. Its community contributes features and translations. As sole maintainer, Louis Erbkamm uses GitHub Copilot to triage issues, trace and fix bugs, review large pull requests for performance risks, and check sensitive code paths for regression cases. The excerpt says Arnis has also added the Moon and Mars and describes classroom projects using generated neighborhoods. ### Source excerpt Louis Erbkamm built the first prototype in his free time. Now GitHub Copilot helps him maintain it alone. The post Arnis turns any place on Earth into a Minecraft world appeared first on The GitHub Blog. ## Spatial-Temporal Learning-Based Distributed Routing for Dynamic LEO Satellite Networks DevFeed: [Spatial-Temporal Learning-Based Distributed Routing for Dynamic LEO Satellite Networks]() Original publisher: [Read original article]() Author: Po-Heng Chou, Chiapin Wang, Shou-Yu Chen, Hsiang-Ming Wang, Kuan-Fu Chen Published: 2026-11-03T05:00:00Z Content type: article Language: en Sources: [arXiv: Networking]() Topics: [routing-system](), [AI Networking](), [Artificial Intelligence](), [Deep learning](), [Neural Network]() Tags: [ablation](), [abstract](), [arxiv](), [attention](), [based](), [cs-lg](), [cs-ni](), [deep](), [delay](), [delivery](), [distributed](), [distributed-routing](), [earth](), [evolution](), [features](), [learning](), [leo](), [networks](), [routing](), [satellite](), [spatial](), [temporal]() ### AI overview The proposed routing framework lets each satellite choose its next hop from local network-state observations and destination information. It combines a graph attention network to represent changing topology and queue states, an LSTM to capture their evolution over time, and a deep Q-network for routing, without a separate traffic-prediction stage. In simulations of a 45-satellite constellation with normal and bursty traffic, packet delivery was comparable to a GNN-enhanced baseline, but the proposed method had higher delay and queue occupancy. Ablations indicated that spatial and temporal learning each contributed benefits. ### Source excerpt arXiv:2605.02413v2 Announce Type: replace Abstract: This paper proposes a spatial-temporal learning-based routing framework for dynamic low Earth orbit (LEO) satellite networks. The routing problem is formulated as a partially observable Markov decision process (POMDP), where each satellite independently selects its next hop using network-state observations and destination information. A graph attention network (GAT) extracts spatial features from the dynamic satellite topology and queue states, while a long short-term memory (LSTM) network captures their temporal evolution. The resulting representation is integrated into a deep Q-network (DQN) for destination-aware next-hop selection without a separate traffic-prediction stage. Simulations in a 45-satellite LEO constellation under normal and bursty traffic demonstrate packet delivery performance comparable to a graph neural network (GNN)-enhanced routing baseline, although the proposed method incurs higher delay and queue occupancy. Ablation studies further demonstrate the complementary benefits of spatial and temporal learning. ## STLT: A Novel System for Energy-Aware Distributed Inference in the Edge-Cloud-Space Continuum DevFeed: [STLT: A Novel System for Energy-Aware Distributed Inference in the Edge-Cloud-Space Continuum]() Original publisher: [Read original article]() Author: Andrija Stanisic, Milos Gravara, Juan Luis Herrera, Stefan Nastic Published: 2026-11-03T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence](), [arXiv: Distributed Computing]() Topics: [Artificial Intelligence](), [AI Architecture](), [On-device AI]() Tags: [cloud](), [cs-ai](), [cs-dc](), [dataset](), [deep-neural-networks](), [detection](), [distributed-inference](), [earth](), [edge](), [energy](), [environment](), [inference](), [latency](), [leo](), [low-earth-orbit](), [models](), [orbit](), [satellite](), [search](), [selection](), [space](), [system]() ### AI overview STLT reduces the energy cost of splitting DNN inference between a low Earth Orbit satellite and a ground station by narrowing split-point choices offline, then selecting among them onboard according to current conditions. The paper says this avoids static choices and the overhead of approaches relying on complex algorithms, historical data, or prediction models. On a real-world satellite telemetry dataset, STLT reduced the candidate search space by 30.7% on average, decision latency by 25.1%, and energy by 27.5% on average. Compared with a state-of-the-art satellite-ground co-inference baseline, it completed up to 20.1% more inference tasks while using 8.8% less energy overall. ### Source excerpt arXiv:2611.01233v1 Announce Type: new Abstract: Low Earth Orbit satellites are pivotal for Earth observation tasks such as wildfire detection, capturing imagery for analysis by Deep Neural Networks (DNNs). On-board analysis reduces downlink volume, but severe energy constraints force a trade-off between computation and data transmission. Splitting the DNN between satellite and ground turns this balance into a design choice, where the split point determines combined computation and downlink energy. Existing approaches typically rely on static split-point decisions that fail to account for dynamic space conditions, or on complex algorithms, historical data, or prediction models that introduce significant overhead. To address these shortcomings, we present STLT, a two-stage system for energy-aware distributed inference between LEO satellites and a ground station. First, an offline analyzer characterizes the DNN architecture to identify a set of favorable split points, consequently reducing the search space. Second, an on-board algorithm performs lightweight energy-aware split selection, dynamically choosing from this pre-computed set based on the current state of the environment. We evaluate STLT on a real-world satellite telemetry dataset, showing that it reduces the split-candidate search space by 30.7% on average, lowers decision latency by 25.1% and energy by 27.5% on average, and delivers up to 20.1% more inference tasks than a state-of-the-art satellite-ground co-inference baseline while consuming 8.8% less energy overall. ## Adaptive Land Units from Earth Embeddings DevFeed: [Adaptive Land Units from Earth Embeddings]() Original publisher: [Read original article]() Author: Albert Alonso, Jamal Jokar Arsanjani, Andres R. Masegosa Published: 2026-11-03T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [geospatial](), [clustering](), [vector embeddings in AI](), [Machine learning](), [Embeddings]() Tags: [adaptive](), [bayesian](), [case-study](), [clustering](), [consumer](), [cost](), [country](), [crop](), [cs-cv](), [cs-lg](), [denmark](), [device](), [earth](), [eess-iv](), [embeddings](), [fewer](), [field](), [gpu](), [grid](), [half](), [images](), [large](), [units]() ### AI overview The study builds adaptive land units by clustering AlphaEarth embeddings with parallel Bayesian hierarchical clustering: units are larger over uniform land and smaller where land is heterogeneous. In Denmark, this unsupervised method reduced 402 million pixels to 1.27 million units, enabling whole-country processing on one consumer-grade GPU. The authors report that regular grids needed about five times as many cells to match land-use labels, and about four times as many to match declared field crops; clustering embeddings achieved the same purity with half as many units as clustering satellite images directly. Land-use maps on these units were within a few accuracy points of pixel-based models at much lower cost. Whole-country estimates with reliable uncertainty took under a minute, while the pixel model did not fit in device memory. These results are reported for Denmark in the supplied abstract. ### Source excerpt arXiv:2611.01050v1 Announce Type: cross Abstract: Modeling land use starts with a choice of spatial units, typically a regular grid of pixels whose resolution is set by the spatial detail the process requires and, often more critically, by the computational budget. With Earth observation data more abundant than ever, we ask whether good land units can instead emerge from the data itself. To do that, we cluster AlphaEarth embeddings with a parallel Bayesian hierarchical clustering into adaptive land units, large where the land is uniform and small where it is highly heterogeneous. Using Denmark as a case study, we turn 402M land pixels into 1.27M newly defined units, about 320 times fewer, without any labels (i.e., unsupervised), making it possible to fit the whole country on a single consumer-grade GPU. Importantly, these units describe the land well, as a regular grid needs about 5 times as many cells to match them against land-use labels, and about 4 times as many against the declared crop of every field. Compared to clustering the satellite images themselves, the embeddings need half the units for the same purity. But more notably, they work as modeling units. Land-use maps learned on them are within a few points of accuracy of pixel models at a small fraction of the cost, and they make whole-country land-use estimates with reliable uncertainty possible in under a minute, where the same model on pixels does not even fit in device memory. ## Forecasting Earth Surface Change in Embedding Space DevFeed: [Forecasting Earth Surface Change in Embedding Space]() Original publisher: [Read original article]() Author: Jingtao Li, Sadiq Jaffer, Anil Madhavapeddy, Pedro Sousa, Anthony Ortiz, Kyle Gao, Zhengpeng Feng, Yanfei Zhong, Kevin White, Srinivasan Keshav Published: 2026-11-03T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [Time Series](), [Embeddings](), [geospatial]() Tags: [change](), [change-detection](), [cs-ai](), [cs-cv](), [earth](), [embedding](), [forecasting](), [foundation-models](), [remote-sensing](), [satellite](), [surface](), [time-series]() ### AI overview E²Forecast predicts future Earth-observation embeddings rather than per-pixel spectral values, aiming to reduce the impact of clouds, shadows, and sensor noise. Its token space handles irregular histories from different sensors and foundation models and supports forecasts at arbitrary query dates. A two-branch transformer predicts a Gaussian mean and per-dimension uncertainty or generates forecast ensembles through rectified flow. On Sentinel-1 and Sentinel-2 benchmarks, embedding forecasts reduced cosine distance from 0.20 to 0.10 and z-score standard deviation from 2.63 to 1.22. Comparing observations with forecasts also enabled unlabeled anomaly detection without retraining, raising F₁ from 41.58 to 48.70 on a global benchmark. The abstract reports consistent superiority across nine remote sensing foundation models. ### Source excerpt arXiv:2611.02211v1 Announce Type: cross Abstract: Repeated satellite imagery permits understanding of how our planet changes over time, letting us predict its future and plan mitigating actions. Most existing studies that rely on a time series of satellite observations forecast the spectral values at each pixel for a future target date. However, clouds, shadows, and sensor noise often corrupt these observations, so that spectral forecasters waste much of their capacity on low-level features or noise rather than real change. With the advent of Earth foundation models, we argue that forecasting in embedding space is a feasible and more effective alternative. To this end, we design a token space that handles irregular histories from different sensors and foundation models and supports forecasts at any query date. Tokens from this space are inputs to our proposed forecaster, E$^2$Forecast, which is a transformer with two branches: a Gaussian branch that directly predicts a mean point forecast and per-dimension standard deviations, and a rectified-flow branch that generates forecast ensembles. On Sentinel-1 and Sentinel-2 benchmarks, forecasting embeddings instead of spectral values halves the cosine distance (from 0.20 to 0.10) and greatly reduces overconfidence, lowering the z-score standard deviation from 2.63 to 1.22. By comparing new observations with its forecasts, E$^2$Forecast can also detect anomalous changes without labels or retraining, raising F$_1$ from 41.58 to 48.70 on a global benchmark. This consistent superiority is validated on nine remote sensing foundation models. Beyond change detection, we believe our approach advances the field by permitting effective reuse of diverse pretrained task heads. Our code is publicly available at https://github.com/ucam-eo/e2forecast. ## Risk-Sensitive Mission Planning for Multi-Target On-Orbit Inspection via Distributional Reinforcement Learning DevFeed: [Risk-Sensitive Mission Planning for Multi-Target On-Orbit Inspection via Distributional Reinforcement Learning]() Original publisher: [Read original article]() Author: Adam Abdin (LGI) Published: 2026-11-02T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [Reinforcement learning](), [Artificial Intelligence]() Tags: [adaptive](), [budget](), [campaign](), [cs-ai](), [distribution](), [earth](), [elements](), [experiments](), [frontier](), [inspect](), [inspection](), [learning](), [low-earth-orbit](), [math-oc](), [mission](), [models](), [orbit](), [planner](), [planning](), [reinforcement-learning](), [risk](), [target]() ### AI overview The paper applies distributional reinforcement learning to choose inspection targets for low-thrust, multi-target servicing missions in low Earth orbit. A QR-DQN agent uses a mean-variance criterion to weigh predicted return against its variability, with target risk changing stochastically and mission time limiting how many targets can be visited. In five-target simulations, increasing risk aversion reduced episodic return variance by about half and lowered mission-loss probability, at the cost of mean return. Learned policies were close in mean return to an efficient frontier computed by enumerating admissible visit orders; risk-averse policies also had lower mission-loss probability than frontier policies. ### Source excerpt arXiv:2610.41647v1 Announce Type: new Abstract: Multi-target on-orbit servicing missions require planning tools that can adapt to evolving operational conditions while accounting for the outcome variability inherent to safety-critical spaceflight. This paper presents an adaptive planning approach for multi-target servicing campaigns in low Earth orbit and applies it to inspection missions flown by a low-thrust servicer. The approach is based on Distributional Reinforcement Learning, which learns the full distribution of mission returns and thereby supports explicit risk-sensitive decision-making at each planning step. Specifically, a Quantile-Regression Deep Q-Network (QR-DQN) agent selects which target to inspect next by jointly considering the expectation and the variance of the predicted return through a mean-variance criterion. The mission environment models circular orbits with a spread in inclination and uses Edelbaum's combined-manoeuvre low-thrust transfer estimates. The observation comprises the orbital elements, the visit history, the observed risk state of each target and the remaining mission-time budget. The risk state of each target changes stochastically during the campaign, making the outcome of an inspection uncertain and creating operationally meaningful distinctions between risk-neutral and risk-averse planning strategies. Because the mission-time budget generally admits only a subset of the candidate targets, the planner must also select which targets to inspect. Numerical experiments on scenarios with five target satellites show that one risk-aversion weight moves the policy across a range of mean return and return variance, over an unchanged environment and reward. Risk-averse agents reduce the variance of the episodic return by approximately one half relative to risk-neutral agents and lower the probability of a mission loss, at a cost in mean return. To validate the learned policies, the efficient frontier, which gives the highest mean return ## Hackaday Links: November 1, 2026 DevFeed: [Hackaday Links: November 1, 2026]() Original publisher: [Read original article]() Author: Tom Nardi Published: 2026-11-02T03:00:09Z Content type: news Language: en Sources: [Hackaday]() Topics: [Internets of interest](), [AdventureX 2026](), [it modernization]() Tags: [2026](), [2026-hackaday-supercon](), [2026-hackaday-superconference](), [apple](), [data](), [deep-dive](), [design](), [display](), [displays](), [duo](), [earth](), [exploring](), [folding](), [hackaday](), [hackaday-columns](), [hackaday-links](), [hacking](), [hubble](), [hubble-space-telescope](), [ifixit](), [internet](), [links](), [nasa](), [slider]() ### AI overview NASA is seeking proposals to keep Hubble in orbit: the telescope has no propulsion of its own and may reenter Earth's atmosphere around 2034. A commercial rescue mission could have years to develop, though NASA has not yet decided to proceed. Separately, some iPhone Duo users report tight hinges, crunching sounds, and broken displays. iFixit theorizes the folding mechanism may fail to support the screen properly, but the cause remains unconfirmed; Apple says only a small number of devices are affected. Radiaverse lets users upload radiation readings from RadiaCode detectors to an interactive globe, similar in concept to WiGLE's map of WiFi access points. ### Source excerpt We're less than a week away from the 2026 Hackaday Supercon, and things are getting a little... intense. So we trust you'll forgive us if this week's post weighs in ...read more ## Quantum-Inspired Convolutional Networks for High-Resolution Pseudo-Label Generation from Multimodal Earth Observation Data DevFeed: [Quantum-Inspired Convolutional Networks for High-Resolution Pseudo-Label Generation from Multimodal Earth Observation Data]() Original publisher: [Read original article]() Author: Abhishek Singh, Giovanni Cuozzo, Alexander Jacob, Claudia Notarnicola Published: 2026-10-30T04:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Machine learning](), [Neural Network](), [Self-supervised learning](), [geospatial]() Tags: [annotation](), [classification](), [cs-cv](), [cs-lg](), [dataset](), [deep-learning](), [earth](), [fusion](), [generate](), [generation](), [learning](), [mapping](), [metrics](), [model](), [multimodal](), [networks](), [quantum](), [remote](), [remote-sensing](), [spatial](), [structure](), [supervision]() ### AI overview The paper presents a quantum-inspired convolutional encoder-decoder that generates high-resolution land-cover pseudo-labels from Sentinel-1 SAR and Sentinel-2 imagery using only coarse MODIS-derived maps. On DFC2020, it first learns multimodal spectral-spatial structure through a self-supervised reconstruction branch, then applies low-resolution label supervision to generate pseudo-labels. The abstract reports that these labels are more spatially coherent and class-consistent than classical baselines when evaluated with test-set classification metrics. The evidence provided is an abstract and does not specify the baselines, metric values, or further experimental limitations. ### Source excerpt arXiv:2610.39773v1 Announce Type: new Abstract: Accurate land-cover mapping from multimodal Earth observation (EO) data remains a challenge in regions where high-resolution annotations are scarce or costly to obtain. This work introduces a quantum convolutional encoder-decoder network for pseudo-label generation using Sentinel-1 SAR and Sentinel-2 multispectral imagery, guided only by coarse MODIS-derived land-cover maps. The proposed framework uses quantum-inspired convolutional blocks to enhance feature extraction and cross-sensor fusion, enabling the model to infer a fine-scale semantic structure from heterogeneous inputs. Experiments were conducted on the DFC2020 dataset, which provides a quadruple of Sentinel-1 SAR, Sentinel-2 multispectral, low-resolution MODIS labels and high-resolution reference land-cover maps. The proposed training framework first learns a robust representation of multimodal inputs through the reconstruction branch before any supervision from low-resolution labels is applied. This self-supervised stage enables the model to capture a fine-scale spectral-spatial structure, which is then used to generate high-resolution pseudo-labels. These pseudo-labels are subsequently evaluated through classification metrics on the test data, demonstrating that the proposed approach produces more spatially coherent and class-consistent compared to classical baselines. The results highlight the potential of quantum-inspired deep learning to improve label generation in low-annotation remote sensing scenarios, offering a scalable path towards high-quality land-cover products for EO applications. ## Rethinking Neural Thickets Without Weight Perturbations DevFeed: [Rethinking Neural Thickets Without Weight Perturbations]() Original publisher: [Read original article]() Author: Nicolas Dufour, Elouan Gard\\\`es, Amil Dravid, Patrick P\\'erez, Loic Landrieu Published: 2026-10-30T04:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence](), [arXiv: Machine Learning]() Topics: [Artificial Intelligence](), [llm-reasoning](), [Neural Network]() Tags: [aggregates](), [analysis](), [arxiv](), [bias](), [block](), [consistency](), [cs-ai](), [cs-cv](), [cs-lg](), [decoder](), [diversity](), [domain](), [earth](), [experts](), [free](), [full](), [gaussian](), [improvement](), [inference](), [language](), [language-models](), [neural](), [weight]() ### AI overview The reported Neural Thickets gains can largely be reproduced without searching for nearby expert weights. Perturbing one decoder block recovers much of the improvement, and adding random bias to the output logits of a fully frozen model matches or exceeds full-weight perturbations on most reasoning tasks. The authors frame this as Perturb-Select-Vote (PSV): generate perturbed predictors, select strong ones using a small validation set, then aggregate their predictions. Their first-order analysis indicates that small weight perturbations act mainly through logit shifts. In vision, logit PSV improves performance under domain shift, including on Earth observation, while full-weight perturbations add little. The abstract presents logit PSV as a training-free alternative across language and vision. ### Source excerpt arXiv:2610.40649v1 Announce Type: cross Abstract: The Neural Thickets hypothesis posits that pretrained language models lie in dense neighborhoods of task-specific experts in weight space, which random Gaussian perturbations can uncover. We reproduce the reported gains, but show that they do not require this geometric explanation. Perturbing a single decoder block already recovers much of the improvement, while keeping the model entirely frozen and adding a random bias to its output logits matches or exceeds full-weight perturbations on most reasoning tasks. This suggests an alternative explanation based on output diversity, validation-based selection, and voting. We formalize this view as Perturb-Select-Vote (PSV), a general procedure that generates perturbed predictors, selects high-performing members on a small validation set, and aggregates their predictions. This interpretation connects Neural Thickets to established inference-time methods, including sampling, self-consistency, and verifier-guided selection. A first-order analysis further shows that small weight perturbations act primarily through their induced logit shifts. We extend this analysis to vision, where logit PSV substantially improves performance under domain shift, including on Earth observation, while full-weight perturbations provide little additional benefit. Together, these results suggest that the gains attributed to nearby weight-space experts can instead be obtained through output diversity, selection, and voting. Logit PSV realizes these benefits with a fully frozen model, providing a simple, training-free alternative to full-weight perturbation ensembles across language and vision. ## Quantized Edge Inference of a Self-Supervised SAR-Optical Encoder on Commodity x86 Hardware DevFeed: [Quantized Edge Inference of a Self-Supervised SAR-Optical Encoder on Commodity x86 Hardware]() Original publisher: [Read original article]() Author: Christian Sanchez-Leon, Dongming Chen Published: 2026-10-28T04:00:00Z Content type: news Language: en Sources: [arXiv: Machine Learning]() Topics: [quantization](), [x86](), [On-device AI](), [Sustainable Computing](), [AI Infrastructure]() Tags: [autoencoder](), [avx2](), [cost](), [cpu](), [cs-cv](), [cs-lg](), [deploy](), [device](), [earth](), [edge](), [edge-inference](), [eess-iv](), [energy](), [gpu](), [hardware](), [inference](), [int8](), [latency](), [linear](), [low-power](), [node](), [optical](), [x86]() ### AI overview On three commodity x86 nodes, INT8 inference of a pretrained SAR-optical masked-autoencoder encoder reduced CPU cost while preserving linear-probe accuracy: micro-mAP changed by only +0.0006, within seed spread. INT8 lowered CPU latency on every node, including a 1.45x gain on the CPU without AVX2, and reduced CPU energy wherever measured; the AVX2+VNNI reference showed a 1.53x latency gain. The best device depended on the CPU: the integrated GPU was 4.38x faster on the weakest node, while the Pareto-optimal choice shifted to the CPU as CPU capability improved. The excerpt describes measurements at FP32, FP16, and INT8, but does not provide full methods or results. ### Source excerpt arXiv:2610.37056v1 Announce Type: cross Abstract: Self-supervised vision transformers are increasingly used as reusable encoders for Earth observation, but their inference cost on the hardware available in the field is rarely characterized. We deploy a pretrained SAR-optical masked-autoencoder encoder on three commodity x86 nodes, from a low-power CPU without AVX2 to an AVX2+VNNI reference, at FP32, FP16 and INT8 on the CPU and, where present, the integrated GPU, and report latency, energy (RAPL and wall plug) and linear-probe accuracy. Quantization is accuracy-neutral: INT8 changes micro-mAP by only +0.0006, within the seed spread, so precision becomes a pure cost decision. INT8 lowers CPU latency on every node, by 1.45x even on the CPU without AVX2, and CPU energy wherever it is measured, and the VNNI reference (1.53x) suggests a further gain beyond AVX2. The Pareto-optimal device shifts from the integrated GPU, which is 4.38x faster than the CPU on the weakest node, to the CPU as the CPU class improves. ## Heterogeneous Space Data Centers: A Key Paradigm for Future Non-Terrestrial and In-Orbit Services DevFeed: [Heterogeneous Space Data Centers: A Key Paradigm for Future Non-Terrestrial and In-Orbit Services]() Original publisher: [Read original article]() Author: Peng Hu, Mustafa Alhassan, Amjad Iqbal Published: 2026-10-26T04:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [data centers](), [AI Integration Strategies]() Tags: [ai](), [ai-infrastructure](), [applications](), [availability](), [budget](), [cases](), [communication](), [computing](), [cost](), [coverage](), [cs-ai](), [cs-sy](), [data-centers](), [deployment](), [design](), [earth](), [eess-sy](), [energy](), [future](), [orbit](), [services](), [space](), [space-data-centers]() ### AI overview The paper proposes heterogeneous space data centers (SDCs) combining large and small nodes with different computing capabilities and power budgets to serve ground, aerial, and in-orbit users. It outlines node sizing and placement, space-to-space and space-to-ground links, and examines dawn-dusk low Earth orbit configurations for solar availability, visibility, and latency-aware coverage. Evaluations using baseline workloads consider latency, service reliability, and energy needs across SDC roles and deployment scales, identifying trade-offs involving node size, orbit, and service portfolio. The supplied text does not provide numerical results. ### Source excerpt arXiv:2610.32260v1 Announce Type: cross Abstract: Lower space-launch cost and advances in AI are driving space-based data centers (SDCs) that provide AI-centric computing and communication services to in-orbit and non-terrestrial assets. However, architectural principles, service models, and energy provisioning for such SDCs, particularly in low-Earth orbit (LEO), remain largely unexplored. We introduce heterogeneous SDCs composed of large and small nodes, differentiated by computing capability, size, and power budget, to support applications spanning ground, aerial, and in-orbit users. We define use cases, node sizing and placement strategies, and interconnection topologies integrating space-to-space and space-to-ground links. Focusing on dawn-dusk LEO configurations, we analyze solar power availability, visibility windows, and latency-aware coverage. Using baseline workloads for typical scenarios, we evaluate latency and service reliability. We quantify energy requirements of different SDC roles and deployment scales, highlighting trade-offs among node size, orbit selection, and service portfolio, and provide design guidance and future research directions for AI infrastructure in space. ## ARENA: Active Region attENtion Aggregation for interpretable solar flare prediction DevFeed: [ARENA: Active Region attENtion Aggregation for interpretable solar flare prediction]() Original publisher: [Read original article]() Author: Luis Bauer, Annalena Kofler, Andr\\'es Asensio Ramos, Aparna Venkataramanasastry, Valent\\'in Mart\\'inez Pillet, Bernhard Sch\\"olkopf Published: 2026-10-26T04:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Neural Network](), [Transformer]() Tags: [abstract](), [aggregation](), [annotations](), [approach](), [architecture](), [arena](), [arxiv](), [astro-ph-im](), [astro-ph-sr](), [attention](), [based](), [can](), [code](), [communications](), [cs-cv](), [cs-lg](), [driving](), [earth](), [flare](), [forecasting](), [https](), [learning](), [region](), [solar]() ### AI overview ARENA predicts solar flares from a variable number of active-region cutouts and uses attention to identify the region most relevant to each prediction. It is trained with multiple-instance learning using only observation-level flare/no-flare labels, without labels for the flare's origin; the learned attention can therefore localize a likely source region without explicit origin annotations. The abstract reports a Heidke skill score of 0.45, comparable to other flare-prediction approaches. ### Source excerpt arXiv:2610.32758v1 Announce Type: cross Abstract: Predicting solar flares is a challenging space-weather forecasting problem with potentially significant impacts on Earth's upper atmosphere, including disruptions to radio communications. We propose Active Region attENtion Aggregation (ARENA), a neural network architecture that predicts solar flares from a variable number of active-region cutouts while learning to identify the region most relevant to the prediction. ARENA is trained using multiple-instance learning under weak supervision, requiring only a flare/no-flare label for the observation and no labels specifying the flare origin. The learned attention mechanism therefore provides a means of localizing the active region driving the prediction without explicit origin annotations. ARENA achieves a heidke skill score of 0.45 which is comparable to other flare prediction approaches. These results demonstrate that active-region-based aggregation can provide an efficient and interpretable approach to solar flare prediction. The code is available at https://github.com/LuisBauer22/ARENA. ## SIGNN++: Robust Graph Neural Star Identification on Earth DevFeed: [SIGNN++: Robust Graph Neural Star Identification on Earth]() Original publisher: [Read original article]() Author: Jeongbin Sohn, Jina Lee, Max Yen, Dowan Gwon, Jeahn Han, Floyd Hepburn-Dickins, Brian Coltin, Ryan Soussan, Pyojin Kim Published: 2026-10-23T04:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Machine learning](), [Neural Network](), [fourier analysis](), [Graphs]() Tags: [alignment](), [angular](), [astro-ph-im](), [code](), [components](), [continuous](), [cs-lg](), [data](), [datasets](), [earth](), [edge](), [environments](), [features](), [fourier-transform](), [framework](), [gnss](), [graph](), [implementation](), [learning](), [neural](), [robust](), [star]() ### AI overview SIGNN++ extends the SIGNN graph neural network method for star identification in GNSS-denied nighttime navigation, targeting reliability under real terrestrial observation disturbances. It adds a more realistic training noise model, Fourier transforms of angular-distance edge features, and RANSAC-based Perspective Star Alignment for identification and attitude refinement. The abstract reports higher star-identification accuracy and solved rates than SIGNN and publicly available baselines, with comparable attitude accuracy and practical computation time for continuous operation. It does not provide numerical results or further evaluation details. ### Source excerpt arXiv:2610.30186v1 Announce Type: cross Abstract: We present SIGNN++, a learning-based star identification (SID) framework for celestial navigation in GNSS-denied nighttime environments. Although researchers have developed learning-based SID methods to improve robustness to observation noise, prior evaluations have focused mainly on synthetic or non-terrestrial environments, leaving their reliability under real terrestrial conditions unclear. To address these limitations, SIGNN++ extends SIGNN, a graph neural network-based SID method, with three components: a more realistic noise model for training data, Fourier Transform (FT) of angular-distance edge features, and RANSAC-based Perspective Star Alignment for identification and attitude refinement. The results show that SIGNN++ is more robust to terrestrial observation disturbances and achieves higher star identification accuracy and solved rates than the original SIGNN and publicly available SID baselines, while maintaining comparable attitude accuracy and practical computation time for continuous operation. The implementation code of SIGNN++ and datasets will be available at https://signnpp.github.io/signnpp/. ## I had plans for that second DevFeed: [I had plans for that second]() Original publisher: [Read original article]() Author: Hazel Burton Published: 2026-10-22T18:04:32Z Content type: news Language: en Sources: [Cisco Talos Blog]() Topics: [PSR-15: HTTP Server Request Handlers](), [known exploitable vulnerabilities](), [Cybercrime]() Tags: [1201](), [2](), [2026](), [about](), [account](), [added](), [adversary](), [agents](), [aging](), [ai](), [ai-agents](), [ai-attack](), [ai-hacking](), [ais](), [critical](), [earth](), [i](), [php](), [threat-source-newsletter](), [up](), [vulnerabilities]() ### AI overview The newsletter links Earth's changing rotation to risks for timekeeping systems: leap seconds have caused technology disruptions, and a future negative second could pose an untested challenge. It says the leap second will be replaced by a leap hour starting next May, and recommends incident response exercises for scenarios that have never occurred. Its security alert describes two unauthenticated PHP remote code execution flaws, "php2shell," affecting builds since 5.4. Talos warns that exposed PHP applications are at risk; PHP 8.1 and older have no patch, so it strongly advises updating to PHP 8.2 or later. Cisco Talos customers are covered by Snort signatures 67359 and 67360. ### Source excerpt The Earth is speeding up, the leap second is being abolished, oh, and there's critical vulnerabilities in PHP. ## Ai2 releases OlmoEarth v1.3 state-of-the-art AI-ready satellite embeddings for the whole planet DevFeed: [Ai2 releases OlmoEarth v1.3 state-of-the-art AI-ready satellite embeddings for the whole planet]() Original publisher: [Read original article]() Author: Kyt Dotson Published: 2026-10-22T13:00:30Z Content type: release Language: en Sources: [SiliconANGLE]() Topics: [Embeddings](), [Artificial Intelligence](), [AI-Ready Data](), [vector embeddings in AI]() Tags: [1-3](), [ai](), [ai2](), [allen-institute-for-ai](), [analysis](), [analyst](), [artificial-intelligence](), [browser](), [carbon](), [change](), [changes](), [earth](), [embedding](), [embedding-model](), [embeddings](), [fingerprint](), [model](), [news](), [olmoearth](), [radar-satellites](), [releases](), [resolutions](), [satellite](), [satellite-ai-model](), [the-latest](), [top-story-6]() ### AI overview Ai2's OlmoEarth v1.3 release pairs precomputed annual satellite embeddings for the planet at 10-, 64- and 128-meter resolutions with a browser-based explorer for making maps without signing in or using a GPU. Each embedding compresses a year of radar and visible/infrared observations into a numerical fingerprint that can help identify land-cover patterns and compare change over time. The excerpt contrasts this quick first-pass analysis with the prior workflow of fine-tuning a model for each task, which required substantial GPU resources and imagery storage; fine-tuning remains useful when deeper modeling is needed. Ai2 says v1.3 took 53% of v1.2's training compute time and 18% of v1's, while its carbon emissions were 50% and 22%, respectively. The article illustrates change analysis with reported mining-area comparisons in Peru, but provides no independent validation details. ### Source excerpt Ai2, the Allen Institute for Artificial Intelligence, today released a satellite map embedding model called OlmoEarth v1.3 and precomputed annual embeddings for the whole planet at various resolutions with a browser-based map-making tool. The model compresses a year of satellite observations into a numerical fingerprint for each 10-by-10-meter patch of ground at the smallest resolution, [...] The post Ai2 releases OlmoEarth v1.3 state-of-the-art AI-ready satellite embeddings for the whole planet appeared first on SiliconANGLE. ## Learning global river system hydrodynamics with a topology-informed foundation model DevFeed: [Learning global river system hydrodynamics with a topology-informed foundation model]() Original publisher: [Read original article]() Author: Hancheng Ren, Gang Zhao, Shuo Wang, Louise Slater, Dai Yamazaki, Shu Liu, Jingfang Fan, Xueying Li, Shibo Cui, Ziming Yu, Shengyu Kang, Depeng Zuo, Dingzhi Peng, Zongxue Xu, Bo Pang Published: 2026-10-21T04:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Machine learning](), [AI Foundation Models](), [Inference]() Tags: [autoregressive](), [connectivity](), [cs-lg](), [earth](), [energy](), [foundation-model](), [global](), [machine-learning](), [network](), [physics](), [physics-geo-ph](), [topology](), [training]() ### AI overview GraphRiverCast (GRC) predicts daily multivariate hydrodynamics at river-reach level across the global network, using sparse gauge observations and requiring no prior river conditions. It uses river connectivity and energetics to reconstruct states from topology and forcing. The abstract reports that it outperforms strong physics-based and machine-learning models, retains skill far beyond its training horizon, and generalises to ungauged reaches and finer-resolution networks. Compared with standard autoregressive models, attribution experiments indicate that GRC relies more on topology and forcing than temporal recurrence. These findings are reported in the abstract; it provides no numerical performance results or detailed evaluation conditions. ### Source excerpt arXiv:2602.22293v3 Announce Type: replace Abstract: Accurate river prediction is essential for water, food, and energy security, yet extending it reliably from sparse in-situ gauges to the full river network remains an open challenge. Although machine learning has transformed Earth-system modelling, the same sparse and uneven observations have largely confined its advances for rivers to discrete basins rather than connected river systems. Here we introduce GraphRiverCast (GRC), a topology-informed river foundation model for reach-level prediction across the global river network. Exploiting the connectivity and energetics of rivers, it is trained on sparse gauges and predicts daily multivariate hydrodynamics with no prior river conditions. GRC outperforms strong physics-based and machine-learning models. Attribution experiments show that, compared with standard autoregressive models that rely more on temporal recurrence, GRC reconstructs river states from topology and forcing, learning the dynamics themselves and inheriting the self-correcting character of river systems. Its skill therefore barely decays with lead time, even far beyond its training horizon. It also generalises to ungauged reaches and finer-resolution networks, bringing reliable hydrodynamic prediction to poorly monitored regions. More broadly, harnessing the structure and energetics of a system offers a strategy for machine learning in other observation-limited Earth systems. ## ORCA: Onboard Reflective Collaborative Agents for Reliable Task Offloading in LEO Satellite Networks DevFeed: [ORCA: Onboard Reflective Collaborative Agents for Reliable Task Offloading in LEO Satellite Networks]() Original publisher: [Read original article]() Author: Yunxin Geng, Shengxiang Hao, Yichen Wu, Haotian Liu, Weiwei Jiang, Lu Cao, Han Meng, Yuxuan Pan, Zhiqing Wei Published: 2026-10-21T04:00:00Z Content type: article Language: en Sources: [arXiv: Networking]() Topics: [Multi Agent Systems](), [Simulation]() Tags: [abstract](), [agent](), [agents](), [architecture](), [arxiv](), [autonomous](), [behavior](), [can](), [capabilities](), [computing](), [critical](), [cs-ni](), [decentralized](), [earth](), [eess-sp](), [efficiency](), [learn](), [leo](), [links](), [networks](), [offloading](), [satellite]() ### AI overview ORCA structures autonomous task offloading in LEO satellite networks as a four-agent Perceive -> Reason -> Act -> Learn loop, addressing partial observations, changing inter-satellite links, heterogeneous computing, and uncertain neighbor behavior. In simulations of a heterogeneous 60-satellite Walker-delta constellation, it achieved competitive system efficiency in cooperative settings and greater task-success robustness and offloading reliability when faced with previously unseen malicious behavior. The abstract attributes this reliability to outcome-grounded reflection and reports that it works without centralized runtime coordination; it does not provide detailed implementation or quantitative results. ### Source excerpt arXiv:2610.28404v1 Announce Type: cross Abstract: Low Earth orbit (LEO) satellites increasingly perform time-critical onboard processing, yet heterogeneous computing capabilities, rapidly changing inter-satellite links (ISLs) and uncertain neighbor behavior make decentralized task offloading difficult under partial observations. This paper presents ORCA, an architecture of Onboard Reflective Collaborative Agents for autonomous onboard task offloading. ORCA organizes offloading decisions into a four-Agent \emph{Perceive}$\rightarrow$\emph{Reason}$\rightarrow$ \emph{Act}$\rightarrow$\emph{Learn} loop. It aims to improve efficiency and reliability in dynamic, heterogeneous, and unreliable LEO satellite networks. The simulation results of a heterogeneous 60-satellite Walker-delta constellation demonstrates that ORCA delivers competitive system efficiency in cooperative settings while achieving superior task-success robustness and offloading reliability under previously unseen malicious node behavior, indicating that outcome-grounded reflection can improve the reliability of decentralized onboard offloading without centralized runtime coordination. [Next page]()