Author

AI Agent Research Hub

Sharing AI, intelligent agents, and cutting-edge scientific computing

Articles

Likes

Views

Comments

Latest from AI Agent Research Hub

26 recent articles

AI Agent Research Hub

Mar 20, 2026 · Artificial Intelligence

Spectral Division of Labor: How HINTS Blends Jacobi and DeepONet for Uniform PDE Convergence

The HINTS framework exploits the complementary spectral biases of classical Jacobi/Gauss‑Seidel relaxations and DeepONet neural operators, alternating them at a fixed ratio to achieve fast, uniform convergence for both positive‑definite and indefinite PDE systems, and integrates seamlessly with multigrid and Krylov solvers.

DeepONetHybrid Iterative MethodsJacobi

0 likes · 27 min read

Spectral Division of Labor: How HINTS Blends Jacobi and DeepONet for Uniform PDE Convergence

AI Agent Research Hub

Mar 18, 2026 · Artificial Intelligence

Variable-Scaling PINN for 2D Navier‑Stokes: How Coordinate Rescaling Improves Stiff PDE Training

This tutorial explains how a simple coordinate scaling (VS‑PINN) reduces stiffness in physics‑informed neural networks, demonstrates its implementation in JAX for the 2D steady incompressible Navier‑Stokes cylinder‑flow benchmark, and shows that after 80 000 Adam iterations the relative errors drop to 2.10 % (u), 5.06 % (v) and 4.45 % (p).

JAXNavier-StokesPINN

0 likes · 24 min read

Variable-Scaling PINN for 2D Navier‑Stokes: How Coordinate Rescaling Improves Stiff PDE Training

AI Agent Research Hub

Mar 16, 2026 · Artificial Intelligence

How NTK Adaptive Weighting and Multi‑Scale Fourier Features Enable PINNs to Solve High‑Frequency PDEs

This tutorial explains why standard physics‑informed neural networks fail on high‑frequency partial differential equations due to spectral bias, and demonstrates how random Fourier feature embeddings, multi‑scale concatenation or spatio‑temporal separation, and Neural Tangent Kernel‑based adaptive loss weighting together overcome the bias and achieve accurate, stable solutions for heat, Poisson, and wave equations using JAX.

Fourier FeaturesJAXMulti-Scale

0 likes · 23 min read

How NTK Adaptive Weighting and Multi‑Scale Fourier Features Enable PINNs to Solve High‑Frequency PDEs

AI Agent Research Hub

Mar 15, 2026 · Artificial Intelligence

The Forgotten Foundational Papers Behind PINNs

This article reviews the 1994 Dissanayake & Phan‑Thien and 1998 Lagaris et al. papers that first introduced feed‑forward neural networks as continuous trial functions for PDEs, contrasting their soft‑penalty and hard‑encoding boundary treatments and showing how they prefigure modern physics‑informed neural networks.

PINNsautomatic differentiationhard boundary encoding

0 likes · 22 min read

The Forgotten Foundational Papers Behind PINNs

AI Agent Research Hub

Mar 14, 2026 · Artificial Intelligence

Adaptive-Weight NTK-PINN Solves High-Frequency Wave Equation Using JAX

This tutorial explains how the Neural Tangent Kernel (NTK) perspective reveals the loss‑balance problem in Physics‑Informed Neural Networks (PINNs), introduces an NTK‑based adaptive‑weight algorithm, provides a full JAX implementation for a 1‑D high‑frequency wave equation, and shows that input normalisation dramatically improves accuracy while only modestly increasing training time.

JAXNTKPINN

0 likes · 27 min read

Adaptive-Weight NTK-PINN Solves High-Frequency Wave Equation Using JAX

AI Agent Research Hub

Mar 13, 2026 · Artificial Intelligence

Deduction vs Induction: 152‑Page Review of Classical vs ML PDE Solvers

This extensive 152‑page review evaluates classical numerical solvers and machine‑learning approaches for partial differential equations using a unified six‑challenge framework, revealing that their fundamental difference lies in epistemology—deductive error bounds versus inductive statistical accuracy—and offering guidance on method choice, hybrid designs, and future research directions.

Computational ChallengesError CertificationHybrid Solvers

0 likes · 26 min read

Deduction vs Induction: 152‑Page Review of Classical vs ML PDE Solvers

AI Agent Research Hub

Mar 10, 2026 · Artificial Intelligence

How Knowledge Distillation Lets Neural Networks Grow Physical Symmetry Without Hard PINN Constraints

The paper introduces Ψ‑NN, a knowledge‑distillation framework that automatically discovers physics‑consistent network structures for PINNs, eliminating the need for manually imposed loss‑function constraints and achieving faster convergence, higher accuracy, and transferable architectures across PDE problems.

Hierarchical ClusteringNetwork Structure Discoveryknowledge distillation

0 likes · 26 min read

How Knowledge Distillation Lets Neural Networks Grow Physical Symmetry Without Hard PINN Constraints

AI Agent Research Hub

Mar 9, 2026 · Artificial Intelligence

How Claude Code AI Agents Generated 100 Research Papers in 10 Days

Within 228 hours, the Fully Automated Research System (FARS) built on Claude Code and other AI agents used 160 NVIDIA GPUs to produce 100 peer‑review‑level papers, achieving an average ICLR score of 5.05—higher than human submissions—while highlighting the expanding role, limits, and safety concerns of AI‑driven scientific automation.

AI agentsAI safetyClaude Code

0 likes · 31 min read

How Claude Code AI Agents Generated 100 Research Papers in 10 Days

AI Agent Research Hub

Mar 2, 2026 · Artificial Intelligence

How AI Agents Can Fully Automate Scientific Research and Boost Productivity

This article surveys the emerging AI‑agent ecosystem that automates the full research lifecycle—from data collection and cleaning to regression, literature synthesis and visualization—highlighting open‑source systems such as OpenScholar, Automated‑AI‑Researcher, AlphaEvolve and PaperBanana, their automation maturity, practical usage guides, known limitations, and essential human‑verification checkpoints.

AI agentsClaude CodeHuman-in-the-loop

0 likes · 26 min read

How AI Agents Can Fully Automate Scientific Research and Boost Productivity

AI Agent Research Hub

Feb 26, 2026 · Artificial Intelligence

Can PINNs Reconstruct Velocity and Pressure Fields from Passive Scalar Visualizations?

This article analyzes the Science paper that uses physics‑informed neural networks (HFM) to infer complete velocity and pressure fields from only passive scalar concentration data such as smoke or dye, detailing the mathematical formulation, network architecture, training strategy, benchmark results, robustness studies, and the method’s limitations and broader impact.

Fluid MechanicsPINNsPassive Scalar

0 likes · 32 min read

Can PINNs Reconstruct Velocity and Pressure Fields from Passive Scalar Visualizations?