Physics-Informed ML papers, 2023–2026

Key findings

• Physics-Informed ML went from 50 papers in 2023 to 199 in 2026, roughly 4.0x in three years.
• NeurIPS publishes the most Physics-Informed ML work: 167 papers from 2023 to 2026, 35% of the four-venue total.
• The fastest-growing overlap is with Multimodal LLMs: papers tagged with both rose from 0 in 2023 to 11 in 2026.
• Among the 5 topics under Climate and Physical Sciences, it ranks #1 by 2026 paper count; the fastest grower in the group, Weather Forecasting, is up 6.1x versus 4.0x here.

Representative papers

Spread across years and venues so every edition shows up

• NeurIPS 2026Derivative-Informed Training of Neural Operators On-the-Fly via Sketched Tangent Consistency
• ICML 2026Bias-Spectrum Neural Processes for Parametric PDEs: Architecture Priors Meet PDE Constraints
• ICLR 2026CONSIGN: Conformal Segmentation Informed by Spatial Groupings via Decomposition
• CVPR 2026Domain Sensitive Federated Learning with Fisher-Informed Pruning
• NeurIPS 2025Activation-Informed Merging of Large Language Models
• ICML 2025Geometry-Informed Neural Networks
• ICLR 2025Beyond the convexity assumption: Realistic tabular data generation under quantifier-free real linear constraints
• CVPR 2025A Physics-Informed Blur Learning Framework for Imaging Systems
• NeurIPS 2024ChaosBench: A Multi-Channel, Physics-Based Benchmark for Subseasonal-to-Seasonal Climate Prediction
• ICML 2024A Space Group Symmetry Informed Network for O(3) Equivariant Crystal Tensor Prediction
• ICLR 2024AirPhyNet: Harnessing Physics-Guided Neural Networks for Air Quality Prediction
• CVPR 2024A Physics-informed Low-rank Deep Neural Network for Blind and Universal Lens Aberration Correction
• NeurIPS 2023ClimSim: A large multi-scale dataset for hybrid physics-ML climate emulation
• ICML 2023Gradient Descent Finds the Global Optima of Two-Layer Physics-Informed Neural Networks
• ICLR 2023Bi-level Physics-Informed Neural Networks for PDE Constrained Optimization using Broyden's Hypergradients
• CVPR 20233D Human Pose Estimation via Intuitive Physics
• NeurIPS 2026Neural Modal Decomposition: Architectural Priors from Observables
• ICML 2026Fine-to-Coarse Fairness-Informed Multi-View Clustering
• ICLR 2026From Neural Networks to Logical Theories: The Correspondence between Fibring Modal Logics and Fibring Neural Networks
• CVPR 2026LaDy Lagrangian-Dynamic Informed Network for Skeleton-based Action Segmentation via Spatial-Temporal

All Physics-Informed ML papers

481 papers, page 2 of 10

51. ICLR 2026ExoPredicator: Learning Abstract Models of Dynamic Worlds for Robot Planning
52. ICLR 2026Fast Convergence of Natural Gradient Descent for Over-parameterized Physics-Informed Neural Networks
53. ICLR 2026Fast training of accurate physics-informed neural networks without gradient descent
54. ICLR 2026From Neural Networks to Logical Theories: The Correspondence between Fibring Modal Logics and Fibring Neural Networks
55. ICLR 2026GenSR: Symbolic regression based on equation generative space
56. ICLR 2026GeoFAR: Geography-Informed Frequency-Aware Super-Resolution for Climate Data
57. ICLR 2026Geometric Autoencoder Priors for Bayesian Inversion: Learn First Observe Later
58. ICLR 2026Gradient-Based Program Synthesis with Neurally Interpreted Languages
59. ICLR 2026GUIDE: Gated Uncertainty-Informed Disentangled Experts for Long-tailed Recognition
60. ICLR 2026Harmonized Cone for Feasible and Non-conflict Directions in Training Physics-Informed Neural Networks
61. ICLR 2026Improving Extreme Wind Prediction with Frequency-Informed Learning
62. ICLR 2026Iterative Training of Physics-Informed Neural Networks with Fourier-enhanced Features
63. ICLR 2026Joint Distribution–Informed Shapley Values for Sparse Counterfactual Explanations
64. ICLR 2026Light Differentiable Logic Gate Networks
65. ICLR 2026Locally Subspace-Informed Neural Operators for Efficient Multiscale PDE Solving
66. ICLR 2026LRIM: a Physics-Based Benchmark for Provably Evaluating Long-Range Capabilities in Graph Learning
67. ICLR 2026Meta-Learning Theory-Informed Inductive Biases using Deep Kernel Gaussian Processes
68. ICLR 2026Overtone: Cyclic Patch Modulation for Cleaner, Faster Physics Emulators
69. ICLR 2026PatchDNA: A Flexible and Biologically-Informed Alternative to Tokenization for DNA
70. ICLR 2026PepTri: Tri-Guided All-Atom Diffusion for Peptide Design via Physics, Evolution, and Mutual Information
71. ICLR 2026Physics vs Distributions: Pareto Optimal Flow Matching with Physics Constraints
72. ICLR 2026Physics-Constrained Fine-Tuning of Flow-Matching Models for Generation and Inverse Problems
73. ICLR 2026Physics-Informed Audio-Geometry-Grid Representation Learning for Universal Sound Source Localization
74. ICLR 2026Physics-Informed Inference Time Scaling for Solving High-Dimensional Partial Differential Equations
75. ICLR 2026Physics-informed learning under mixing: How physical knowledge speeds up learning
76. ICLR 2026PINFDiT: Energy-Based Physics-Informed Diffusion Transformers for General-purpose Time Series Tasks
77. ICLR 2026PolySHAP: Extending KernelSHAP with Interaction-Informed Polynomial Regression
78. ICLR 2026PoseX: AI Defeats Physics-based Methods on Protein Ligand Cross-Docking
79. ICLR 2026ProofFlow: A Dependency Graph Approach to Faithful Proof Autoformalization
80. ICLR 2026QLCoder: A Query Synthesizer For Static Analysis of Security Vulnerabilities
81. ICLR 2026Rapid Training of Hamiltonian Graph Networks Using Random Features
82. ICLR 2026Representing local protein environments with machine learning force fields
83. ICLR 2026Sample-efficient evidence estimation of score based priors for model selection
84. ICLR 2026Sharing State Between Prompts and Programs
85. ICLR 2026SONATA: Synergistic Coreset Informed Adaptive Temporal Tensor Factorization
86. ICLR 2026STDDN: A Physics-Guided Deep Learning Framework for Crowd Simulation
87. ICLR 2026Structural Inference: Interpreting Small Language Models with Susceptibilities
88. ICLR 2026Towards Persistent Noise-Tolerant Active Learning of Regular Languages with Class Query
89. ICLR 2026Transformers are Inherently Succinct
90. ICLR 2026Two failure modes of deep transformers and how to avoid them: a unified theory of signal propagation at initialisation
91. ICLR 2026Uncertainty-Aware Diagnostics for Physics-Informed Machine Learning
92. ICLR 2026UrbanGraph: Physics-Informed Spatio-Temporal Dynamic Heterogeneous Graphs for Urban Microclimate Prediction
93. ICLR 2026Weight-Space Linear Recurrent Neural Networks
94. ICLR 2026WINA: Weight Informed Neuron Activation for Accelerating Large Language Model Inference
95. ICML 2026A Deep Learning Model of Mental Rotation Informed by Interactive VR Experiments
96. ICML 2026A Geometric Lens on Physics-Aligned Data Compression
97. ICML 2026A model of errors in transformers
98. ICML 2026APIC: Orthogonalized Neuro-Symbolic Modeling for Nonlinear Dissipative Dynamics
99. ICML 2026AutoMat: Physics-Guided Agentic Reasoning for Solving Ill-Posed Inverse Microscopy Problems
100. ICML 2026Bias-Spectrum Neural Processes for Parametric PDEs: Architecture Priors Meet PDE Constraints

FAQ

How many Physics-Informed ML papers were published at NeurIPS, ICLR, ICML and CVPR from 2023 to 2026?

481 in total: 50 in 2023, 85 in 2024, 147 in 2025, 199 in 2026.

How many Physics-Informed ML papers are at NeurIPS 2026?

NeurIPS 2026 has 54 Physics-Informed ML papers, versus 65 the year before.

How many Physics-Informed ML papers are at ICLR 2026?

ICLR 2026 has 64 Physics-Informed ML papers, versus 38 the year before.

How many Physics-Informed ML papers are at ICML 2026?

ICML 2026 has 51 Physics-Informed ML papers, versus 25 the year before.

How many Physics-Informed ML papers are at CVPR 2026?

CVPR 2026 has 30 Physics-Informed ML papers, versus 19 the year before.

Which conference publishes the most Physics-Informed ML papers?

NeurIPS, with 167 papers from 2023 to 2026.