Scientific Computing Software Tools

Ever wonder how we got seedless watermelons, disease-resistant wheat, and tomatoes that stay fresh for weeks? It wasn't luck. It was precision science using multiple techniques, each solving a different agricultural challenge. For thousands of years, farmers selected the best plants and bred them together, hoping for improvement. Building the next generation of resilient crops relies on several distinct modification techniques. Here is a look at the actual science driving modern agriculture: 🔹 Cross-breeding combines two sexually compatible species to merge desirable parent traits. This traditional method gave us modern corn from ancient teosinte. 🔹 Polyploidy multiplies chromosome sets to impact fertility and size. This is why seedless grapes exist and why some strawberries are massive compared to their wild ancestors. 🔹 Mutagenesis uses controlled mutagens like targeted radiation to induce random mutations, then selects for beneficial traits. Sounds dramatic, but it's created hundreds of crop varieties we eat daily, including your Ruby Red grapefruit. 🔹 Protoplast Fusion merges cells or cell components from different species, transferring traits that couldn't naturally cross-breed. Think of it as cellular matchmaking beyond sexual compatibility barriers. 🔹 Transgenesis adds genes from completely different species to create varieties with desired traits. Inserting bacterial genes into corn made it pest-resistant (Bt corn), reducing pesticide needs dramatically. 🔹 Genome Editing uses enzyme systems like CRISPR to modify DNA directly within the cell with surgical precision. No foreign genes added, just targeted edits to the plant's own genome. It's how we're creating disease-resistant cacao to save chocolate from extinction. Each technique serves a purpose. Cross-breeding is slow but natural. Genome editing is fast and precise. Transgenesis crosses species barriers. Mutagenesis introduces controlled randomness. The question isn't whether we should modify crops—we've been doing it for 10,000 years. The question is: which tool for which challenge? As we look toward the future of climate-adapted agriculture, which of these technologies do you believe will drive the most critical commercial breakthroughs over the next decade? Drop your reaction below. 👇 #plantscience #cropmodification #agriculturalinnovation #crispr #plantbreeding #biotechnology #foodsecurity #genomeediting #agtech

Physics-Informed Neural Networks in Inversion Problems Inversion problems are present across various fields, such as geophysics, medicine, and materials science, where the primary goal is to estimate hidden parameters or reconstruct information based on observed data. Physics-Informed Neural Networks (PINNs) have emerged as a powerful tool for addressing these problems by embedding physical laws directly into the learning process of the network. The Challenge of Inversion Problems In many inversion problems, the objective is to uncover unknown parameters, such as medium properties or anomaly sources, from indirect measurements. Traditionally, these tasks are tackled using complex numerical methods that demand significant computational resources and often struggle with ambiguities and instability, especially when data is limited. PINNs stand out in inversion problems because they integrate observational data with prior knowledge of the governing physical laws, like partial differential equations (PDEs). By incorporating these physical constraints during training, PINNs can find solutions that respect physical consistency, making the results more robust even under conditions of sparse or noisy data. How It Works in Practice In an inversion problem using PINNs, the network is designed to minimize a composite loss function. This loss term includes both the observational data errors and the residual error of the PDEs that model the physical phenomenon. Through this process, the PINN adjusts the system’s unknown parameters so that the network's predictions align with both the available data and the physical laws, enabling reliable inference of unknown parameters. Applications Geophysics: PINNs are applied to infer subsurface properties, such as density and seismic velocities, from surface data. This approach allows for precise subsurface models without the costly conventional inversion methods. Medicine: In medical imaging, PINNs can be used to reconstruct high-quality images from low-resolution scans, reducing patient radiation exposure while preserving image quality. Advantages Robustness with Sparse Data: PINNs can bypass the need for large datasets by incorporating physical laws directly into the inference process. Noise and Ambiguity Reduction: Since PINNs consider the physical model, they tend to produce more consistent results that are less sensitive to noise in the observed data. Real-Time Applicability: In some cases, PINNs can be implemented to perform real-time inference, which is beneficial for monitoring and controlling dynamic physical systems. The application of PINNs to inversion problems represents a significant advancement, providing a way to solve complex problems more quickly, efficiently, and robustly. These methods are already transforming fields like geophysics and healthcare, and their potential continues to grow as more researchers apply this technology to estimate unknown parameters in physical systems.

New review out in Trends in Plant Science from our group, tackling one of the most persistent problems in plant breeding: moving complex traits into elite varieties is hard, slow, and expensive, and breeders have long lacked the tools to plan it rigorously. Led by Seema Yaadav, the paper frames trait introgression not as a fixed backcross protocol but as a constrained multi-objective optimisation problem, balancing success probability, elite genome recovery, time, and cost. It maps the full toolkit to tackle this: predictive cross metrics, pan-genomes and introgressiomics libraries, speed breeding, doubled haploids, and recombination engineering. The vision is 'designed diversity'. Pipelines where the genotype is planned computationally before a single cross is made, then refined iteratively as data accumulates. The building blocks are largely here. The challenge now is integrating them into breeder-facing tools that work in practice. Great to see Seema lead this one. Congratulations also to co-authors Meredith McNeil (CSIRO), Peter Dodds (CSIRO), and Ben Hayes (UQ) and thanks to the Grains Research and Development Corporation for supporting our R&D in this space. 📄 https://lnkd.in/gz6tCkAT

The recent market shocks have left a tremendous effect on investors’s mindmap. The volatility and the jump in the asset prices movements are extremely high. On a behavioural finance level, there is surely panic in the market leaving less headroom to ponder about the situations for normal retail investors. Thus, the implementation of mathematical models becomes a necessity not only to predict pricing value but considering volatility, jumps and high shocks. Although, the reason is different but at the end considering the dip in Japan stock market was lower than the Covid-19 pandemic. Using stochastic process mathematical models like Heston model could be used to predict both the asset price and its volatility, allowing for a mean-reverting volatility process while Hull White model for incorporating jumps in the asset prices. This way we get the volatility, jump and asset price. Also, if we consider multivariate volatility (time varying correlations with standardized returns) with correlation b/w the multiple assets, a great recommendation to opt for the extended GARCH model with dynamic conditional coorelation (DCC). Once you could predict the dynamic correlation with varying time portfolio optimization becomes more efficient with time-varying covariance matrix. No wonder, why maths with finance using tech makes such predictions better and high accuracy rates. #quantitativefinance #quant #finance #riskmanagement #japan

🌧️ Rainfall data analysis as a fundamental input for advanced hydrological modelling . Rainfall data is the governing variable in hydrological studies, as it directly affects the estimation of surface runoff, the hydrological response of basins, and the accuracy of mathematical model outputs used in flood risk assessment and water infrastructure design. 📊 The hydrological importance of rainfall analysis Accurate analysis of rainfall data aims to: Describe the statistical characteristics of rainfall (frequency, intensity, variability) Represent the temporal and spatial distribution of precipitation Identify design storms Reduce uncertainty in hydrological models. 🧠 Advanced statistical analysis of rainfall The choice of statistical method depends on the nature of the data and the length of the time series. The most prominent methods are: 🔹 Frequency Analysis Application of probability distributions such as: Gumbel Extreme Value Type I Log-Pearson Type III Generalised Extreme Value (GEV) Goodness of Fit test using: Kolmogorov–Smirnov Chi-Square Anderson–Darling. 🔹 Intensity-Duration-Frequency (IDF) Curves Derivation of mathematical relationships between intensity (I), duration (D), and frequency (T) Form the basis for the design of stormwater drainage networks and urban infrastructure. ⏱️ Temporal Analysis Time series analysis to detect: Long-term trends (Trend Analysis) Climate changes and their impact on precipitation patterns Use of tests: Mann–Kendall Sen’s Slope Estimator. 🌍 Spatial Rainfall Analysis Due to the heterogeneity of precipitation, rainfall is spatially represented using: Thiessen Polygons Inverse Distance Weighting (IDW) Kriging (Geostatistical Methods) Integration with geographic information systems (GIS) is an essential step in improving rainfall representation at the catchment level. 💧 Linking rainfall and hydrological models Rainfall analysis results are used directly in: Rational Method (for small basins with rapid response) SCS Curve Number Method for estimating loss and surface runoff Rainfall–Runoff Models such as: HEC-HMS WMS SWMM ⚠️ Technical challenges Incomplete or irregular rainfall records High spatial variability of storms The impact of climate change on the stability of statistical assumptions (Stationarity). Any hydrological model, regardless of its computational accuracy, remains dependent on the quality of the rainfall data analysis input into it. Rainfall analysis is not a preliminary step, but rather the essence of the entire hydrological process.

Interested in diffusion models and geophysics? Or just curious how AI can tackle scientific inverse problems at scale? The PhysicsNeMo team at NVIDIA has been exploring diffusion models for physics-ML, and we’ve released a new example on Full-Waveform Inversion (FWI)—the seismic technique that reconstructs subsurface velocity models by fitting recorded waveforms. 🎯 The library provides an end-to-end training recipe for elastic FWI with variable density, including data prep, training, and both zero-shot and physics-informed sampling on an extended E-FWI pipeline. 🧠 We use a built-in U-Net diffusion backbone from the PhysicsNeMo SDK and train it with the EDM framework, pairing seismic inputs with velocity-model outputs in a configuration tailored to geophysics. 🧩 The generation pipeline supports DPS-style physics-informed posterior sampling at generation time, adding a guidance term so samples better satisfy the elastic wave equation and observed data. 🎛️ Both conditional and unconditional setups are supported, so you can explore conditioning strategies—including patterns like classifier-free guidance—without being locked into one regime. 📈 Generation scales to large ensembles for uncertainty-aware decisions, with distributed sampling for UQ, sensitivity exploration, and even guiding well-log/core sampling placement using ensemble variance. 🌍 Who is this for? Oil & Gas exploration and CO2 sequestration teams, seismology researchers, and anyone working on AI for inverse problems. 🔧 Beyond FWI: the same diffusion-plus-physics recipe extends to non-destructive testing and medical/acoustic tomography–style problems. 🔗 Keep it simple: all how-to details live in the example docs—jump in when you’re ready. Interested? Give it a try! • Example docs (Diffusion-FWI): https://lnkd.in/gGBUifzw • PhysicsNeMo docs hub: https://lnkd.in/gkxWHuGa • PhysicsNeMo GitHub repo: https://lnkd.in/gX8fmt4x #Geophysics #FWI #SeismicImaging #GenerativeAI #DiffusionModels #PhysicsAI #AI4Science #Seismology #NVIDIA #PhysicsNeMo

Seismic Modeling within the Petrel Environment🔰 What is Seismic Modeling? Seismic modeling is the process of simulating the propagation of sound waves through a 3D geological model. Simply put, we take a hypothetical model of the Earth's layers (distribution of velocity and density) and introduce an "artificial seismic wave" to observe how it behaves. The result? We generate a Synthetic Seismogram that shows what the seismic data should look like if our geological model is correct. The Petrel Workflow: From Reservoir to Synthetic Seismogram Forward Seismic Modeling in Petrel is executed by integrating geological modeling steps with geophysical tools: 1. Building the Static Geological Model Input: The work relies on a 3D Structural Grid built from interpreting horizons and faults within Petrel. Property Modeling: Grid cells are populated with rock properties (such as porosity, permeability, and fluid saturation) using statistical methods (like Gaussian simulation). 2. Calculating Acoustic Properties Petrophysical Transformation: Here, we move from geological to seismic properties. We use petrophysical equations (such as Wyllie's equations or Raymer-Hunt-Gardner relations) to estimate the P-Wave Velocity and Density for every cell in the model. Acoustic Impedance Calculation: Acoustic Impedance (the fundamental attribute measured by seismic signals) is calculated by multiplying velocity by density. 3. Generating the Synthetic Seismic Volume Applying the Wavelet: To generate a synthetic seismogram that matches the real data, an appropriate Wavelet must be applied. This wavelet is extracted from the actual field seismic data to ensure a realistic simulation. Actual Seismic Modeling: Petrel uses the model's acoustic properties (Acoustic Impedance) and the wavelet to calculate Reflection Coefficients at layer boundaries, which are then combined via Convolution to generate a synthetic 3D seismic volume. 4. Comparison and Refinement Visual Matching: The synthetic seismic volume is displayed alongside the real (imported) seismic volume in the Petrel viewer. Adjustment: If there is a mismatch, the geological model, the petrophysical properties, or the velocity model used for depth-to-time conversion must be reviewed and adjusted. #Petrel #Seismic_Modeling #Geoscience #EandP #SeismicModeling #Petrel #Geophysics

Geophysics and Open Source! In the dynamic field of #geophysics, one of the main challenges is access to the #software tools required for robust #data #processing. Often supplied with proprietary instruments or associated with significant commercial costs, these softwares often hinder researchers and professionals alike. Fortunately, several Open-Source alternatives provide powerful tools for geophysical data processing while fostering a culture of collaboration and knowledge sharing within the community. Here are some amazing #open-source packages for geophysical data processing: 1. ResIPy: designed for processing #geoelectrical data, specifically #direct #current (DC) and #induced #polarization (IP) data. It offers a Python API and a standalone graphical user interface (GUI) for high-level filtering, error modeling, inversion/forward modeling, and post-processing. More information at https://lnkd.in/dG47y4dA. 2. EMagPy: focused on processing frequency domain electromagnetic measurements (#FDEM) obtained with electromagnetic induction (#EMI) devices. It provides a Python API for Jupyter notebooks and a standalone GUI. More details can be found at https://lnkd.in/dmeRD3xY. 3. PyGimli: an open-source library for #modeling and #inversion in geophysics. It offers tools for structured and unstructured mesh management, finite element and finite volume solvers, various geophysical forward operators, and Gauss-Newton based frameworks for constrained, joint, and fully-coupled inversions. Explore PyGimli at https://lnkd.in/daM_hqS2. 4. SimPEG: a #Python package dedicated to #simulation and #parameter estimation in geophysics. It focuses on finite volume simulation and provides modular components for spatial discretization, optimization routines, and geophysical problems. SimPEG supports 1D, 2D, and 3D problems and is designed for large-scale inversions. More information at https://lnkd.in/dFmt2FvX. 5. ObsPy: a Python framework specifically designed for #seismological data analysis. It offers a comprehensive suite of tools for #processing, #visualizing, and #manipulating #seismic data. ObsPy is widely used in earthquake monitoring, crustal studies, seismic imaging, and more. You can access it at https://docs.obspy.org/. 6. RefraPy: a Python software package with a graphical interface for analyzing #seismic #refraction data developed by Victor Guedes. It consists of two modules, Refrapick and Refrainv, providing functionalities for data analysis and inversion. More details at https://lnkd.in/dGaduWkF. 7. GeoVES: a free #Excel-based tool for 1D inversion of vertical resistivity soundings (#VES) developed by Andreas de Jong. Useful for students learning inversions of 1D #resistivity data, particularly for #model #fitting. You can download the tool from https://bit.ly/2OpcMvm and refer to this YouTube video https://lnkd.in/deAAvGgT for guidance on how to use it.

New paper in Journal of Geophysical Research: Machine Learning and Computation: “PyQuake3D: A Python Tool for 3-D Earthquake Sequence Simulations of Seismic and Aseismic Slip.” PyQuake3D is a Python-based, open-source package for 3-D fault-slip simulations using quasi-dynamic BEM with hierarchical matrices. It scales from laptops to clusters (MPI) with optional GPU acceleration, and we validate it against benchmarks with applications from laboratory experiments, heterogeneous-friction faults, and large plate boundary faults. Paper (DOI): https://lnkd.in/dJ6GAD4S Code: https://lnkd.in/ddYNStxY Docs: https://lnkd.in/dwD7nJAM #OpenSource #Python #Seismology #EarthquakePhysics #ComputationalGeophysics #HPC #GPU

Your next increment will be decided by a Normal Distribution! HR teams typically ask leaders to "fit the bell curve" when it comes to their team's annual performance ratings. That curve is nothing but a normal distribution - one of the most practical applications of probablity distributions. Think of probability distributions as a way to map the possible outcomes (from an experiment or a situation) and their probabilities. Distributions are super useful when building Linear models in Data Science. Knowing which model to fit is a function of the distribution of the data - better fit = better model performance. Some of the most important distributions are: 1. Normal Distribution - Helps estimate population distribution using sample mean and standard deviation (saving a lot of time and money in practical experiements) - One of the most naturally ocurring distributions in life (adult heights, weights, exam scores etc) - Linear Regression residuals are assumed to have a normal distribution (more on this when we get to ML fundamentals) - Practical applications include Customer Satisfaction scores, stock market returns, product defect rates in manufacturing and many others 2. Binomial Distribution - Used for situations with two possible outcomes (success or failure) - Applications include modelling success rate of sales calls, pass/fail rate of quality control checks in manufacturing etc 3. Poisson Distribution - Used for counting the number of events that happen in a fixed interval of time or space - Specifically designed for modelling count data - Practical applications include # of Daily website visitors, # hourly customer support calls, customer arrivals in a store etc 4. Exponential Distribution - Used for estimating time between events such as customer arrivals in a store, customer wait time on calls, time between sales on an e-commerce platform 5. Log-Normal Distribution - Similar properties to the normal distribution, except that mean and standard deviation are calculated post taking log - Distribution is positively skewed (See image and compare shape with normal) - Practical applications include asset prices, income distribution in an economy etc. Link to Github in comments. --------------------------------------------------------------------------------- Follow me Karan Berry for daily insights into Statistics and Data Science fundamentals.