Riverside Technology, inc.’s Post

Very pleased to report that our article "Assessing the Feasibility of an NWP Satellite Data Assimilation System Entirely Based on AI Techniques" has been published in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing https://lnkd.in/gXcx6Z4k. "Data assimilation (DA) is facing major challenges in terms of its ability of handling the ever-increasing volume of valid, useful, and potentially impactful environmental data and the problem is expected to be exacerbated in the near future if a solution to dramatically increase efficiency is not found. A new approach to perform large-volume data fusion and assimilation, based entirely on artificial intelligence (AI) modern techniques including machine learning and computer vision techniques, is proposed in this study. This approach to DA is applied and demonstrated to real environmental data measured by NOAA-20 and MetOp-C microwave satellite-sounders to reproduce traditional numerical weather prediction DA performances from the U.S. National Oceanic and Atmospheric Administration (NOAA). We assess the impact of our AI-based analysis on forecasts by; 1) performing statistical assessments versus the European Centre for Medium-Range Weather Forecasts analyses, 2) assimilating the AI-based analyzed fields as pseudo-sounding observations in the NOAA global data assimilation system (GDAS), and 3) running forecast experiments using FV3GFS initialized with those observations. To identify the impact of our AI-based assimilations, we compare the forecast skill of several experiments where GDAS is driven with conventional and satellite radiometric observations and with conventional and AI-based pseudo-observations. The results presented are encouraging but are considered only a first initial step toward demonstrating an entirely AI-based environmental data fusion/assimilation system capable to efficiently handle large-volume data and take of the information content available." #dataassimilation #gapfilling #remotesensing #machinelearning #artificialintelligence #noaa

🚁 🌍 AI/ML to Accelerate GRG Development: 1️⃣ AI/ML for Quick Infrastructure Detection: Preligens' AI/ML algorithms are capable of detecting and vectorizing various infrastructures within a predefined area, including buildings, intersections, roads, barriers, and tents with high accuracy and speed. 2️⃣ Direct Integration with ArcGIS: Deep learning models integrated directly with ArcGIS Pro generate GRG grid and operational maps, simplifying complex terrains for seamless navigation during critical missions. Benefits in Action: 🛰 Faster processing of satellite imagery: Analysts can generate a map within minutes, ensuring critical information isn't missed, especially in challenging and frequently changing desert environments. ⏱ Time efficiency: Analysts focus on strategic decisions, not laborious tasks, saving precious hours when every moment counts. 🤝 Interoperability: Smooth coordination at all levels enhances joint operations and coalition efforts. #AI #GIS #GRG #GEOINT 

Our Geospatial Science and Human Security Division works to integrate human dynamics, geoinformatics, #datascience, autonomous systems, & resilient communications to provide decision-makers with timely information on critical situations. Geospatial AI, or #GeoAI, is helping to transform how we observe, analyze, and visualize landscapes and human dynamics from across the world to provide novel insights on human populations, the built environment, and critical infrastructures. Recently, Lexie Yang and Philipe Dias were invited to participate as panelists in the Harvard CGA’s conference “Evaluating the Science of Geospatial AI”. Scientists from across academia, national labs, and industries came together to examine the current conditions, opportunities, and connections between #AI and #GIS. Thanks to their knowledge and participation, Yang and Dias are helping to further the research on GeoAI.

Would #AI revolutionize the process of #scientific discovery in fields that traditionally relied on simulation-based numerical methods? How might the increasing reliance on AI for scientific computations and modeling influence the architecture of #supercomputers? https://lnkd.in/eiAu4u23

Save up to 80% of your time with AI-assisted seismic interpretation workflows!   It's amazing to see how technology has evolved over the past few years. With AI-ML automated workflows, tasks like horizon and fault interpretation, and volume attribute generation are done in a fraction of the time, without requiring geoscientists with expertise in data or computer science. This allows geoscientists to focus on interpretation of the results such as structure and attribute maps, and integration with other data types for accelerated decision making, thereby taking advantage of rig slots and supply chain availability.  https://lnkd.in/ghSFaef5

📢More News from the ESA EDHPC Conference! 🚀 Today our Vito colleague Bart Beusen presented a paper that delves deep into the realm of AI-driven image compression. Here's a brief overview of the key takeaways: 🔍 Extension to Hyperspectral: The research has successfully extended the CORSA method, originally designed for Sentinel-2 data, to hyperspectral images. This means the algorithm can now handle images with more than just 4 bands, going up to 20 to 40 bands. This is a significant leap in the field of remote sensing image compression! 🤖 Model Quantization: The research investigates the requirements for running the algorithm onboard a satellite. Considerations include model downsizing, use of separable convolutions, and model quantization to int8. 📊 Comparison to CCSDS 123.0-B-2 Standard: When pitted against the standard compression algorithm for on-board near-lossless compression, the new method clearly outshines the CCSDS 123.0-B-2 standard in terms of MSE, PSNR, and SSIM. This is a testament to its performance and potential for future applications. Kudos to the author and the entire team behind this research! 🎉 #EDHPC #ESA #AI #ImageCompression #Hyperspectral

Bagel Network: Revolutionizing AI Data Collaboration Bagel Network, a decentralized machine learning data network, aims to tackle the challenges of data accessibility and collaboration in AI research. By establishing a bidirectional marketplace, it enables easy sharing and licensing of ML datasets among AI agents and engineers. This initiative promises to lower the cost of data acquisition, facilitate collaboration, and allow AI agents to monetize their data contributions. With a recent $3.1 million pre-seed funding round, Bagel Network holds the potential to revolutionize the distribution of machine learning data, making AI development more affordable and accessible to all stakeholders. #BagelNetwork #AI #datacollaboration #machinelearning #decentralized #marketplace #dataacquisition #collaboration #funding #revolutionizing #accessibility

Last day at GeoConvention Partnership 2024. Taking in some insightful discussions about utilizing AI to aid and assist in subsurface modeling, interpretation and field advancements. Overriding takeaway is that the more data the better, however the caveat is that the data must be good, clean and accurate. RECON Petrotechnologies #openholelogging #waterwelllogging #oilfieldservices