106 articles tagged with #scientific-computing. AI-curated summaries with sentiment analysis and key takeaways from 50+ sources.
Researchers demonstrate that combining the FABRIC testbed with LLM-based coding assistants can significantly reduce the effort required to reproduce published scientific experiments. The AI-assisted approach achieved 4-6x reduction in reproduction effort across three case studies, though human intervention remained necessary for complex analytical workflows.
Researchers introduce Clifford Kolmogorov-Arnold Networks (ClKAN), a new neural network architecture designed for function approximation within Clifford Algebra spaces. The approach uses Randomized Quasi-Monte Carlo grid generation to address computational scaling challenges in higher dimensions, with applications in scientific computing and physics simulations.
Researchers introduce Flow Annealing Posterior Sampling (FAPS), a new function-space framework that unifies stochastic-process regression with PDE inverse problems using pretrained flow-matching priors. The method enables probabilistic inference from sparse observations while maintaining computational efficiency and accurate uncertainty quantification, outperforming existing baselines.
Researchers introduce LISDD, a framework for identifying where and why physics-based models fail by localizing errors to specific operating regimes and discovering sparse symbolic corrections. The method outperforms existing global-correction approaches by keeping parameter bias near zero while maintaining statistical rigor through finite-sample testing.
Researchers introduce Process-Reward Tactic Evolution, a training framework that enables LLM agents to reliably execute complex bioinformatics workflows in Galaxy by accumulating reusable tactics from verified workflow rollouts. The approach combines process verification, curriculum learning, and tactic libraries to improve long-horizon task completion, biological correctness, and execution efficiency compared to baseline methods.
Researchers propose Hard-Soft Physics-Informed Neural Networks (HSPINN), a novel framework that improves how AI solves complex mathematical equations by enforcing boundary conditions exactly while treating other constraints as soft penalties with adaptive weighting. This advancement addresses persistent challenges in physics-informed neural networks, achieving faster convergence and higher accuracy across multiple equation types.
SVGym (SciVerseGym) is a new open-source framework that standardizes reinforcement learning workflows for automated crystal discovery by treating materials design as a Markov decision process. The environment decouples agent logic from materials infrastructure, enabling researchers to apply machine learning algorithms to accelerate the discovery of new materials with desired properties.
Researchers propose NOMTO, a framework combining neural operators with symbolic equation discovery to identify governing equations from complex data involving nonlocal operators and memory effects. This advancement extends traditional symbolic discovery methods beyond local derivatives, enabling discovery of more realistic physical and mathematical models.
Researchers propose Time-Frequency Gated Spectral Neural Operators (TF-SNO), a machine learning framework that dynamically adapts its spectral response to model non-stationary partial differential equations where frequency content changes over time. The approach outperforms existing spectral neural operators on six benchmarks by using state-dependent modulation rather than static spectral filters.
Researchers introduce residual-space evolutionary optimization, a framework combining flow-based generative models with evolutionary algorithms to enable data editing without requiring differentiable objectives or gradient-based optimization. The method separates local refinement and broad exploration through self-pollination and cross-pollination mechanisms, validated on image benchmarks and crystal structure data.
Researchers propose physics-informed generative AI architectures that enforce hard physical constraints by construction rather than post-hoc filtering, using semiconductor manufacturing as a test case. The work surveys emerging techniques including physics-informed diffusion models, PDE-constrained variational approaches, and conservation-law-respecting networks to ensure generated designs, data, and processes are physically valid rather than merely plausible.
Researchers introduce Chain of Operators (CHOP), a framework that enables frozen neural operator models to handle out-of-distribution tasks without fine-tuning by constructing chains of explicit mathematical transformations. The approach demonstrates improved generalization across different PDE families while maintaining interpretability.
Astrophysicist Chi-kwan Chan leverages OpenAI's Codex to accelerate black hole simulations, enabling researchers to efficiently model extreme gravitational phenomena and validate Einstein's general relativity predictions. This application demonstrates how AI-assisted coding tools enhance scientific computing workflows in fundamental physics research.
AutoPDE introduces a novel agentic approach to solving partial differential equations by maintaining solver strategies as explicit, inspectable objects rather than implicit code details. The system achieves a 54.5% pass rate on PDE Agent Bench, improving upon existing baselines by 14.2 percentage points through a three-stage process combining PDE analysis, numerical method selection, and adaptive tuning.
Researchers introduce Online Generative Active Sampling (OGAS), an active learning method that improves PDE surrogate models by strategically sampling challenging configurations during training. Using a parallel diffusion model to steer data generation toward difficult regimes, OGAS reduces worst-case prediction errors across multiple PDE types without significant computational overhead.
Researchers present a constrained natural-language interface for finite element simulations that uses LLMs only for front-end parsing tasks while delegating critical solver logic to human-written templates. The system achieves 100% parse validity and demonstrates effective integration of language models with scientific computing by limiting AI to non-critical paths, reducing reliability risks.
Researchers demonstrate that mixtures of neural operators (MoNOs) reduce computational complexity in operator learning by routing inputs through expert models rather than using a single large model. The approach achieves better scaling properties with depth, width, and rank while maintaining approximation quality, with implications for efficient AI system design.
Researchers introduce Deepcontour, a hybrid framework combining deep learning and classical numerical methods to accelerate solutions for large-scale Generalized Eigenvalue Problems. The system achieves up to 5.63x speedup by using a neural operator to predict eigenvalue distributions and automatically optimize integration contours for contour integral solvers.
Researchers introduce Geo-NeW, a neural network method that solves Partial Differential Equations while preserving physical laws and generalizing to unseen geometries. The approach combines learned differential operators with finite element spaces that explicitly encode geometry information, achieving state-of-the-art performance on PDE benchmarks with significant improvements on out-of-distribution test cases.
Researchers evaluated general-purpose AI coding agents on a real neuroscience data-to-discovery pipeline, finding they can automate individual pipeline stages but fail at end-to-end integration. The study reveals critical gaps in AI agents' ability to apply scientific judgment, interpret visual outputs, and manage computational resources—challenges absent from current benchmarks.
Researchers introduce a bidirectional search task linking code snippets with text descriptions and vice versa, addressing the gap between scientific publications and their implementations. They present a large dataset with automatically-generated training data and manually-annotated test sets, along with a modular encoder-based approach that achieves strong in-domain results with promising out-of-domain generalization.
Researchers developed an LLM-based pipeline that automatically translates legacy Fortran scientific code into JAX, a differentiable programming framework. Applied to a 19,000-line land surface model, the approach achieved 24x speedup and 8x faster parameter optimization while enabling gradient-based analysis through automatic differentiation.
Researchers successfully modernized NMAP-RKPM, a 60,000-line Fortran physics simulation engine, from single-threaded MPI to parallel C++ using a structured agentic AI approach. Rather than relying on LLMs alone, the team developed a 'hand-holding' methodology combining manual examples, continuous buildability checks, and scoped sessions that proved highly effective for legacy code transformation.
Researchers introduce OnlyDense, a machine learning framework that reduces computational costs for Lagrangian particle simulation methods like SPH and MPM by representing massive particle systems as functions in Hilbert space rather than discrete particle sets. The method achieves 0.99+ R² accuracy using just 32 basis functions on million-particle simulations, combining classical reduced-order modeling with deep learning.
Researchers introduce Topological Neural Operators (TNOs), a novel framework for machine learning that processes data across multi-dimensional topological structures rather than just points or edges. The approach uses Discrete Exterior Calculus to model interactions while preserving geometric and physical properties, demonstrating improved accuracy on PDE benchmarks including irregular geometry problems.
