partial differential equations
PulseAugur coverage of partial differential equations — every cluster mentioning partial differential equations across labs, papers, and developer communities, ranked by signal.
- instance of alphaXiv 90%
- used by Neural Operators 90%
- instance of partial differential equation 90%
- instance of Allen–Cahn equation 90%
- used by alphaXiv 70%
- instance of Gotit.pub 70%
- instance of ScienceCast 70%
- uses Neural Operators 70%
- used by ScienceCast 70%
- used by Gotit.pub 70%
- instance of Neural Operators 70%
- instance of Influence Flower 70%
11 day(s) with sentiment data
-
New hybrid FEM-PINN method tackles time-dependent PDEs
Researchers have developed a novel hybrid numerical method that merges the time finite element method with deep neural networks to solve time-dependent partial differential equations. This approach differs from conventi…
-
LLMs generate neural network weights and architectures for physics problems
Researchers have developed Physics2Weight (PtW), a novel framework that leverages large language models (LLMs) to directly generate neural network architectures and weights for solving partial differential equations. Th…
-
Twindex offers exact proofs for model parameter indistinguishability
Twindex is a new tool that provides exact, replayable proofs to demonstrate when different parameter sets for a model are indistinguishable based on given data. It generates a JSON certificate and a checker that verifie…
-
New WaSP method trains AI to solve PDEs without expensive data
Researchers have developed the Wasserstein Data Synthesis Pipeline (WaSP), a novel method for training neural operators to solve partial differential equations (PDEs) without requiring expensive input-solution pairs. Wa…
-
Conformal prediction faces challenges with time-dependent PDEs
Researchers have identified a critical issue with conformal prediction methods when applied to time-dependent partial differential equations (PDEs). Standard conformal prediction, which offers distribution-free coverage…
-
New WAMRViT model uses dynamic quadtree tokenization for adaptive PDE forecasting
Researchers have developed WAMRViT, a novel Vision Transformer architecture designed for adaptive mesh forecasting of partial differential equations (PDEs). This model utilizes dynamic quadtree tokenization, allowing it…
-
New self-supervised method creates universal latent space for neural operators
Researchers have developed a novel self-supervised pre-training method for neural operators, decoupling representation learning from operator learning. This approach, inspired by advances in natural language processing …
-
LLMs and deep generative models advance PDE solving capabilities
Researchers have developed new methods for solving complex partial differential equations (PDEs) using large language models (LLMs) and deep generative models. One approach, OpInf-LLM, uses operator inference with LLMs …
-
New theory guarantees stability and consistency for physics-informed neural networks
A research paper published on arXiv, titled "Non-Asymptotic Stability and Consistency Guarantees for Physics-Informed Neural Networks via Coercive Operator Analysis," introduces a theoretical framework for analyzing the…
-
New meshless models offer efficient PDE reduction
Researchers have introduced Meshless Latent Dynamics Models (LDMs) as a novel approach for model order reduction (MOR) of time-dependent partial differential equations (PDEs). Unlike traditional methods that rely on fix…
-
Jacobian Fine-Tuning Stabilizes Neural PDE Emulators
A new research paper explores Jacobian fine-tuning as a method to stabilize neural emulators for chaotic partial differential equations (PDEs). The study decomposes error sources, separating physical error growth from m…
-
FC-PINO enhances precision for physics-informed neural operators
Researchers have introduced FC-PINO, a novel architecture for Physics-Informed Neural Operators (PINOs) designed to enhance precision in solving partial differential equations (PDEs). Traditional PINO methods struggle w…
-
New framework enables norm-optimal model reduction in RKHSs
Researchers have developed a new framework for model order reduction (MOR) within reproducing kernel Hilbert spaces (RKHSs). This framework extends existing necessary interpolatory conditions for norm-optimal reduced-or…
-
Graph Neural Networks Enhanced for Hyperbolic Transport Modeling with CFL Conditioning
Researchers have developed a new method for graph neural networks (GNNs) to better model hyperbolic partial differential equations (PDEs), which often involve complex phenomena like shocks and discontinuities. The propo…
-
New generative model framework uses nonlinear Schrödinger systems
Researchers have introduced a novel generative modeling framework that leverages forward-backward partial differential equations (PDEs), drawing connections to mean-field game theory and nonlinear Schrödinger systems. T…
-
FoundPDE: New foundation model for symbolic discovery of PDEs
Researchers have developed FoundPDE, a generative foundation model designed for the symbolic discovery of partial differential equations (PDEs) from data. This transformer-based model utilizes Fourier and spatial-deriva…
-
New neural network method enhances PDE solving capabilities
Researchers have developed SFRBE--RaNN, a novel adaptive randomized neural network method designed to more effectively solve partial differential equations (PDEs). This two-stage approach redistributes random features b…
-
New Test-Time Forecasting framework enables foundation models to predict unseen dynamics
Researchers have introduced a new framework called Test-Time Forecasting (TTF) that enables foundation models to predict the dynamics of unseen dynamical systems. This framework utilizes looped transformers and gradient…
-
New method optimizes neural surrogate training for PDEs using multi-fidelity data
A new research paper introduces a method for training neural surrogates for partial differential equations (PDEs) more efficiently. The approach leverages the tunable cost of PDE simulations by training on a mixture of …
-
New AI model ReMAIN reconstructs complex plasma dynamics from sparse data
Researchers have developed a new deep learning model called ReMAIN (Recurrent Multiscale Affine-modulated Inference Network) to reconstruct complex plasma dynamics from limited sensor data. This model improves upon prev…