LIVE PULSE
4.0 Anthropic CEO Amodei calls for slower AI development and shared safety rules — 11 src2.2 Agility Robotics unveils Digit 5 humanoid for warehouses and factories — 2 src2.0 Apple ships rebuilt Siri with Google Gemini, but not in the EU — 2 src1.8 Siri AI in macOS 27 Golden Gate: FAQ, Germany availability, privacy questions — 2 src1.4 Sam Altman says OpenAI will not go public in 2026, citing AI safety concerns — 5 src1.1 OpenAI contractors review real ChatGPT conversations to rate responses, report says — 2 src1.1 Anthropic data retention policy prompts firms to limit Claude use for sensitive work — 1 src1.1 Study examines issue bias in LLMs used as writing assistants before Swedish 2026 election — 1 src1.1 Study Audits Misalignment in Multi-Modal World Models — 1 src1.1 Retrieval-Grounded Reasoning Approach Proposed for Universal Multimodal Embeddings — 1 src4.0 Anthropic CEO Amodei calls for slower AI development and shared safety rules — 11 src2.2 Agility Robotics unveils Digit 5 humanoid for warehouses and factories — 2 src2.0 Apple ships rebuilt Siri with Google Gemini, but not in the EU — 2 src1.8 Siri AI in macOS 27 Golden Gate: FAQ, Germany availability, privacy questions — 2 src1.4 Sam Altman says OpenAI will not go public in 2026, citing AI safety concerns — 5 src1.1 OpenAI contractors review real ChatGPT conversations to rate responses, report says — 2 src1.1 Anthropic data retention policy prompts firms to limit Claude use for sensitive work — 1 src1.1 Study examines issue bias in LLMs used as writing assistants before Swedish 2026 election — 1 src1.1 Study Audits Misalignment in Multi-Modal World Models — 1 src1.1 Retrieval-Grounded Reasoning Approach Proposed for Universal Multimodal Embeddings — 1 src
HEATPULSEAI MAGAZINES
FLIP · FOLLOW · SAVE
Log inSign up

Physics-informed neural networks

topic8 events
papersTODAY 04:00 UTC

Sparse autoencoders used to probe physics-informed neural network internals

A new arXiv paper introduces PhysSAE, a method that applies sparse autoencoders to inspect what hidden layers in physics-informed neural networks actually represent. The authors aim to determine whether these networks learn localized, physically meaningful features tied to the PDE residuals they are trained on. The work falls within mechanistic interpretability research for scientific machine learning.

interpretability/sparse-autoencoders/pinns/research
papersTODAY 04:00 UTC

Pullback-corrected auxiliary variable optimizer targets multi-term scientific ML losses

A new arXiv paper proposes a pullback-corrected scalar auxiliary variable (PB-SAV) optimizer that adds momentum and adaptive mobility. The method is aimed at scientific machine learning objectives that combine several loss terms, such as the residual, boundary, initial, and data losses used in physics-informed neural networks. The abstract frames the work as addressing optimization challenges specific to these composite objectives.

optimization/scientific-ml/pinns/arxiv
papersTODAY 04:00 UTC

HGTO: Graph-Based Physics-Informed Formulation for Structural Topology Optimization

A new arXiv preprint introduces HGTO, a unified graph-based, physics-informed framework for density-based structural topology optimization. The approach reframes the usual nested loop of material updates, structural analysis, and sensitivity computation, drawing on neural density parameterization and dual-field physics-informed methods that require no labeled data. The abstract positions the work as a data-free alternative to conventional optimization pipelines.

research/physics-informed/optimization/graph-neural-networks
papersTODAY 04:00 UTC

Linearized PINN Uses Pretrained Nonlinear Layers for Solving Differential Equations

Researchers introduce lPINN, a reduced-order neural basis approach for both forward and inverse differential equation problems. In an offline phase, the method derives operator-compatible continuous neural basis functions from an ensemble, then solves equations in a lower-dimensional space. The work aims to combine physics-informed modeling with the efficiency of pretrained linearized representations.

research/physics-informed/neural-networks/differential-equations
papersTODAY 04:00 UTC

Study identifies derivative-fidelity failure mode in physics-informed neural networks

A new arXiv paper argues that physics-informed neural networks can match target function values while still producing inaccurate derivatives. The authors describe this as a distinct failure mode and provide strengthened benchmark evidence for it, based on models trained only on function values. The work suggests that evaluating PINNs by function agreement alone can mask errors in the derivative terms central to solving differential equations.

pinns/benchmarks/research/autodiff
papersTODAY 04:00 UTC

ACR-PINN: Layer-wise Adaptation and Gradient Conflict Resolution for PINNs

Researchers propose ACR-PINN, a physics-informed neural network framework that pairs layer-wise dynamic adaptation of coordinate representations with a method for resolving conflicts among gradients coming from heterogeneous physical constraints. The work frames architecture and optimization as a joint design problem, aiming to improve training when competing constraints pull the model in different directions.

physics-informed/optimization/neural-networks/research
papersSEP 10 04:00 UTC

Gradient-guided Gaussian adaptive sampling proposed for training physics-informed neural networks

A new arXiv paper introduces 3GAS-PINNs, a variant of physics-informed neural networks that uses gradient-guided Gaussian adaptive sampling to place collocation points. The method targets common weaknesses in PINNs on nonlinear partial differential equations, such as slow convergence, gradient imbalance, and poor resolution of demanding regions. By concentrating sampling where it matters most, the approach aims to improve training efficiency and solution accuracy.

pinns/pdes/adaptive-sampling/training
papersSEP 10 04:00 UTC

Edge-of-chaos initialization fails for higher input derivatives in wide networks

New research indicates that while the edge-of-chaos initialization scheme keeps first-order input perturbations stable in very wide randomly initialized networks, higher-order input derivatives become unstable under the same setup. Because techniques such as physics-informed losses, score matching, and derivative regularization rely on those higher derivatives, the results expose a gap in how such networks should be initialized for derivative-based training. The analysis focuses on smooth fully connected networks with scalar inputs.

theory/neural-networks/initialization/training