Surrogate modeling

Overview

A surrogate model learns a map from inputs to a system's output, so that each new query costs one model evaluation and not a simulation run or a device measurement 2609.16066 §1 2507.05291 §1. Its accuracy is bounded by the reference data it was trained on. The first two variants below train on solver output and keep the solver in the loop for checking final designs 2609.16066 §2.3 2609.16066 §4.3 2507.05291 §4.1. The fourth is different: a learned scorer for runs that have already been simulated, which saves no simulator calls 2412.13802 §5.1.2.

The main design choice is what the surrogate predicts. A classical regressor predicts one objective directly 2609.16066 §4.1.3. A field surrogate predicts a whole field and computes the objectives afterwards 2609.16066 §1 2609.16066 §3.2. A graph surrogate predicts a stress value at every node of a mesh 2507.05291 §3. A material surrogate predicts a stored-energy function, and stress follows from its derivatives 2005.12183 §3.3. A further use replaces the derivatives of a physical device rather than its output 2104.13386 app. S1. That use is not the same idea as a surrogate gradient, which replaces a non-differentiable step inside a network during training of spiking networks. Another use scores a finished run, so that a search can rank it 2412.13802 §4.2.1. A last use is over time rather than over design: a few spatial modes from a transient solver are kept, and a small network predicts how their amplitudes evolve 2609.17101 §4 2609.17101 §5.

Variants

Field surrogate with a branch-trunk neural operator

The surrogate maps the 28 blade-geometry variables and a query point (r, z) on a meridional plane to five basic flow fields: static pressure, static temperature, absolute and relative squared velocity, and mass flux 2609.16066 §2.1 2609.16066 §3.3. Performance metrics such as efficiency, pressure ratio and mass flow are then computed from the predicted fields by averaging, not learned as separate outputs 2609.16066 §3.2 2609.16066 tab. 6. The model is a branch-trunk neural operator (TNO) whose trunk uses Galerkin-type attention blocks instead of softmax attention 2609.16066 §3.1. Card: Panoramic aerodynamic performance prediction for turbomachinery cascades.

Reported results, all author-reported on one compressor (Rotor 37) with a validation split only: mean field error 0.083% against 0.572% for an FNO baseline 2609.16066 tab. 4; one forward pass in about 0.9 ms, against roughly 800 s per CFD sample on the authors' CPU 2609.16066 tab. 3 2609.16066 §2.2. Training the model took 40.5 minutes on the authors' GPU 2609.16066 tab. 3.

Graph surrogate for local stress on a periodic micro-structure

The surrogate maps a periodic micro-structure mesh and its mean stress to the stress at each mesh node, so local stress is available without a full micro-scale solve for each macro point 2507.05291 §1 2507.05291 §3. The mesh becomes a graph whose nodes carry the mean stress, the position and a boundary label, and whose edges carry the length between neighbours. Extra edges join opposite sides of the cell, which is how periodicity enters 2507.05291 §3.1. The network is a message-passing graph neural network with 10 steps and a latent size of 128 2507.05291 §3.2. Training adds a penalty on the discrete stress divergence, weighted by λ, so that predictions approximately satisfy local equilibrium 2507.05291 §3.3. The linear dataset has 10,000 periodic meshes and the hyperelastic one 9,795 2507.05291 §4.1 2507.05291 §5.1. Card: Physics-Informed Graph Neural Networks to Reconstruct Local Fields Considering Finite Strain Hyperelasticity.

Reported results, author-reported on single values with no spread shown: on linear elastic meshes, NMSE is 1.08e-2 with the divergence penalty, against 1.13e-2 for a periodic graph network without it and 2.33e-2 without periodic edges 2507.05291 tab. 1. On hyperelastic meshes, NMSE is 2.18e-2 against 5.36e-2 for the periodic graph network 2507.05291 tab. 2. Its inference is said to be about 500 times faster than finite-element solves on a CPU, but that compares GPU inference with CPU solves and excludes training 2507.05291 §5.3.1.

Differentiable digital model for physical systems

Here the surrogate is a neural network fitted to input-output pairs measured on a physical device. The device runs the forward pass, and the surrogate supplies the derivatives in the backward pass. The loss is computed from the measured outputs, so the surrogate's errors affect the gradient direction and not the loss value 2104.13386 app. S1. This avoids estimating the device's own derivatives by finite differences, which would cost one physical call per input and parameter 2104.13386 app. S1. The digital model in typical runs is trained on about 10^4 uniformly sampled inputs 2104.13386 app. S2. Card: Deep physical neural networks enabled by a backpropagation algorithm for arbitrary physical systems.

Reported results, author-reported on three physical systems. A five-stage optical chain reaches 93% on vowel classification, against about 40% at best for training entirely in simulation 2104.13386 fig. 2 2104.13386 fig. 3. On MNIST, a mechanical plate reaches 87%, an electronic circuit 93% and a hybrid optical network 97% 2104.13386 fig. 4. The paper's in-simulation baseline includes a fitted noise model 2104.13386 app. S2.

Learned violation scorer for scenario search

Here the model does not stand in for the simulator. A Transformer encoder reads the per-frame states of the vehicles in a finished simulation, seven features per vehicle, and outputs one scalar that the paper calls a violation probability 2412.13802 §4.1 2412.13802 §4.2.1 2412.13802 tab. 2. That scalar is one of four objectives in a multi-objective genetic search over driving scenarios. The others are a road-feature classifier, the minimum distance to another vehicle, and the number of unique violations 2412.13802 §4.2.1. The scorer is trained on 1,000 short scenarios of two minutes each, with binary cross-entropy and Adam 2412.13802 §5.1.2. Card: SimADFuzz: Simulation-Feedback Fuzz Testing for Autonomous Driving Systems.

Reported results, author-reported on one driving system (InterFuser) in CARLA on one map: 35 unique violations in six hours, against 26, 18 and 3 for the three baselines 2412.13802 §5.2.2. The paper lists the single system and map as a threat to external validity 2412.13802 §5.3. The scorer does not reduce cost. Every candidate is still simulated, and the search is bounded by time budgets rather than by a count of simulator calls 2412.13802 §5.1.2 2412.13802 §5.2.

Thermodynamics-constrained network for a material law

Here the surrogate replaces a hand-made constitutive model, the rule that gives stress from strain at a material point. Such models need calibration, and micromechanical simulations are too costly to run inside large finite-element analyses 2005.12183 §1. The network predicts a Helmholtz free-energy function and the increments of internal variables. Stress and dissipation are not network outputs. They are derivatives of the energy network, so the first law holds by construction 2005.12183 §3.3 eq. 13. The training data are generated from a reference thermodynamic model, not from measurements 2005.12183 §4.2. The energy network uses a squared-argument ELU, whose second derivative is nonzero everywhere. Activations whose second derivative vanishes over part of their range fit the stress output poorly 2005.12183 §3.2 tab. 2. The scope is strain-rate-independent, small-strain processes, with isothermal experiments 2005.12183 §2.3 2005.12183 §2.4. Card: Thermodynamics-based Artificial Neural Networks for constitutive modeling.

Reported results, author-reported on synthetic data from 1D and 3D elasto-plastic reference models. In 1D, a standard network with a matched number of hyper-parameters gives the wrong energy and dissipation, and generalizes worse on cyclic and random paths 2005.12183 §5.1.4. Beyond the training range of strain increments, accuracy drops, but the predicted dissipation stays positive 2005.12183 §5.1.3. In 3D, the network agrees with the reference for uniaxial cyclic loading, with less accuracy at the largest increments. The authors report larger errors for it in the mean and deviatoric stress 2005.12183 §5.2.2 2005.12183 §5.2.3. The paper states that code is available on request and gives no repository 2005.12183 §1.

Reduced-order DMD model with a neural amplitude model

Here the surrogate is a time-stepping model of a transient solver. The training snapshots come from a finite-difference solver for the one-dimensional viscous Burgers equation with a discontinuous initial state, at Reynolds numbers 10², 10³ and 10⁴ 2609.17101 §2. Each case gets its own model, and no model is fitted across Reynolds numbers 2609.17101 tab. 1. Randomized dynamic mode decomposition gives a set of spatial modes and their amplitudes. The number of modes is chosen by a search over all ranks, which keeps the rank with the smallest time-averaged error 2609.17101 §4 2609.17101 alg. 1 2609.17101 eq. 29. The randomized SVD inside it draws a random test matrix 2609.17101 alg. 2. A cascade-forward network with 10 hidden units then models the time series of each amplitude, using past estimator outputs and delayed amplitude inputs 2609.17101 §5 2609.17101 fig. 3 2609.17101 eq. 32. The modes and the estimated coefficients together form the data model. The paper does not require a Galerkin projection of the governing equations [arxiv:2609.17101 abstract]. Card: Randomized-DMD digital twin data models.

Reported results, author-reported and in-sample only. The DMD error is 3.2e-7 at Re 10² with 15 modes, 4.4e-7 at Re 10³ with 20 modes and 5.4e-8 at Re 10⁴ with 20 modes 2609.17101 tab. 1. The data-model error, grouped by coefficient, is between 1.3e-4 and 8.1e-4 across the three cases 2609.17101 tab. 2 2609.17101 tab. 3 2609.17101 tab. 4. The paper names no external baseline and reports no run time for the full-order solver. Offline work takes at most about 2 s and the online stage 8 to 17 s, on hardware that is not stated 2609.17101 §6 2609.17101 fig. 13. A speed-up over the solver is therefore not shown.

When to use which

The papers compare against their own baselines and, for the first variant, against classical regressors on one task. The rows below say what the reported evidence supports, and nothing more.

Implementation sketch

The loop below covers the two simulator-bound variants: sample the design space, run the simulator, fit the model, search on the model, and verify the result with the simulator 2609.16066 §2.3 2609.16066 §4.3. The graph variant fits the same loop with a graph input and a per-node target 2507.05291 §3. The details of the architecture are not stated in full for the field variant, so this is a skeleton, not a reimplementation. The physics-aware variant does not follow this loop. It trains on device measurements and uses the model only for gradients 2104.13386 app. S1. The learned-scorer variant adds one scoring call after each simulation and removes no simulation 2412.13802 §5.1.2. The material-law variant follows neither loop. It trains on data from a reference material model, and its stress comes from differentiating the energy network [arxiv:2005.12183 §3.3] 2005.12183 §4.2. The reduced-order variant follows neither loop. It decomposes solver snapshots in time and fits a network to the amplitude series, with no design search and no per-candidate simulation 2609.17101 §4 2609.17101 §5.

# Surrogate loop for a simulator-bound design task  [§2.3] [§3.3] [§4.3]
# Input:  design-variable box; CFD simulator
# Output: surrogate model and a CFD-verified design

def surrogate_loop(design_box, simulator):
    # Sample designs spread across the box (Latin hypercube in the paper)
    designs = latin_hypercube(design_box, n=2900)     # [§2.3]
    # Run the simulator once per design (about 800 s each in the paper)
    fields = [simulator(d) for d in designs]          # [§2.2]
    # Reduce each simulation to the training target
    # (circumferential mass-flux average in the paper)  [§3.2] [§3.3]
    targets = [average_fields(f) for f in fields]
    # Fit the surrogate on a train split, check on a validation split
    model = fit(designs[:2500], targets[:2500])       # [§3.3]
    # Search on the surrogate, not the simulator (GA or NSGA-II)  [§4.3]
    candidate = search(lambda d: metrics(model(d)))
    # Verify the chosen design with the simulator before using it  [§4.3]
    return candidate, simulator(candidate)

Contradictions

No paper disagrees with another in the reported numbers, but the metrics are not comparable. The field error in percent 2609.16066 tab. 4, the NMSE of the graph variant 2507.05291 tab. 1, the classification accuracy of the physical variant 2104.13386 fig. 2 and the unique-violation count of the scorer variant 2412.13802 §5.2.2 and the DMD and data-model errors of the reduced-order variant 2609.17101 eq. 27 2609.17101 tab. 1 measure different things on different tasks. Read them within each paper only.

Inside the cited papers, several points disagree or do not match:

• Inside the field-surrogate paper, the text and the printed equation describe the branch-trunk combination differently, so check that step before implementing 2609.16066 §3.1 2609.16066 eq. 1.
• Inside the graph paper, the linear divergence value is 4.56e-4 in the text and 3.23e-4 in the table, with no explanation given 2507.05291 §4.3 2507.05291 tab. 1.
• Inside the physical-variant paper, the in-simulation baseline is described as "barely distinguishable from random guessing" in the figure caption, and as reaching about 40% at its best in the text 2104.13386 fig. 3 2104.13386 fig. 2.
• Inside the scorer paper, the six-hour margin over DriveFuzz is 35 − 18 = 17 from the table, but the text and conclusion say 27 more 2412.13802 §5.2.2 2412.13802 §7.
• Inside the reduced-order paper, the snapshot count is 301 in the text 2609.17101 §2 but 300 in the caption of its Table 1 2609.17101 tab. 1. The text claims the method has the best correlation, yet the table lists 1.0000 for every case, so the metric does not separate them 2609.17101 §6 2609.17101 tab. 1.
• Inside the scorer paper, the output is described as a violation probability, but the text does not say whether it is squashed before it is used as a fitness term 2412.13802 §4.2.1.

Papers

• Panoramic aerodynamic performance prediction for turbomachinery cascades: a transformer-style neural operator predicts compressor flow fields from blade geometry, and metrics are computed from the fields.
• Physics-Informed Graph Neural Networks to Reconstruct Local Fields Considering Finite Strain Hyperelasticity: a graph network predicts the stress at every node of a periodic micro-structure, with a divergence penalty for equilibrium.
• Deep physical neural networks enabled by a backpropagation algorithm for arbitrary physical systems: a digital model fitted to measured device data supplies the gradients for training physical networks.
• SimADFuzz: Simulation-Feedback Fuzz Testing for Autonomous Driving Systems: a genetic fuzzer whose Transformer scorer of simulated runs is one of four selection objectives.
• Thermodynamics-based Artificial Neural Networks for constitutive modeling: a network predicts a free-energy function, and stress and dissipation follow by differentiation, so the first law holds by construction.
• Randomized-DMD digital twin data models: a randomized DMD gives spatial modes, and a cascade-forward network predicts their amplitudes over time, for one Burgers setting at a time.