We address Task A of the 1st DAFx Parameter Estimation Challenge, which aims to retrieve the physical parameters of a plate model from an impulse response. To do so, we use the Simulation-Based Inference (SBI) framework, in which we train a neural network to estimate a density over plate parameters given an impulse response, using a dataset generated by the simulator. Inference for a new impulse response then requires only a forward pass through the network, without involving the simulator. For each test observation, we fine-tune a specific network: additional simulation rounds are performed by sampling parameters from the current estimated distribution, simulating the corresponding impulse responses, and fine-tuning to produce the specialized network.
Figures & tables
Parameter
Symbol
Range
Unit
Areal mass density
μ
[2.43,106.15]
kgm−2
Bending-to-mass ratio
D/μ
[0.28,201.2]
m4s−2
Tension-to-mass ratio
T0/μ
[9.4×10−5,411.5]
m2s−2
Plate length ( y )
Ly
[1.1,4.0]
m
Readout position ( x )
xo
[0.51,1.0]
frac. Lx
Readout position ( y )
yo
[0.51,1.0]
frac. Ly
Table 1: Ranges of the six estimated parameters, used to normalize the NMSE ( 2 ). The invariant ranges ( μ , D/μ , T0/μ ) are those induced by the raw-parameter ranges [ 3 ] ; Lx=1\text{,}\mathrm{m}$$ is fixed.
Figure 1: Summary of the neural architecture.
Figure 2: Architecture of one CNN branch.
Figure 3: Sequential SBI refinement.
Per parameter
Global
μ
D/μ
T0/μ
Ly
xo
yo
PSO baseline
50.56
26.16
9.72
8.46
80.43
95.18
83.39
Offline
2.17
2.18
0.10
0.23
5.50
3.23
1.75
Refined
3.80
2.49
0.03
0.14
11.45
4.93
3.74
Selected
0.47
0.68
0.02
0.08
1.66
0.15
0.25
Oracle
0.34
0.44
0.03
0.08
1.25
0.10
0.15
Table 2: Average per-parameter NMSE ( ×10−3 , lower is better) on the 50 -IR test set. PSO is the challenge’s particle-swarm baseline.
Figure 4: Distribution of the per-parameter NMSE across the 50 -IR test set.
Figure 5: TARP expected coverage of the offline posterior over 500 held-out plates.
We present a simulation-trained, non-iterative estimator for Task A of the 1st DAFx Parameter Estimation Challenge. Each unnormalized plate-reverb impulse response is summarized by amplitude, spectral, and decay descriptors, and an ensemble of tree regressors estimates the six target parameters in one pass. Across two independent synthetic validation sets, the normalized models outperform the training-set mean and an earlier raw-regression baseline. On a shared set, the final ensemble also outperforms a single run of the official default PSO at substantially lower inference cost. Since the official labels are hidden, parameter accuracy is measured on simulator-matched data, and the released responses support only audio-side consistency checks. The estimator returns point estimates without uncertainty.
Minhui Lu, Joshua D. Reiss
Centre for Digital Music Queen Mary University of London London, United Kingdom
Task B of the 1st DAFx Parameter Estimation Challenge requires estimating the frequencies, decay rates, gains, and number of modes in a dense plate-reverb impulse response. Weak and overlapping modes make sparse peak detection prone to severe undercounting. We train an ExtraTrees regressor on simulator-generated data to predict mode counts in four frequency bands. These counts define dense frequency grids, after which a differentiable all-pole resonator model refines decay and gain while keeping frequency fixed. On two separate synthetic validation sets, the system reduces a local challenge-style error by about 66% relative to the official default peak-picking baseline. The improvement is mainly associated with lower mode-count mismatch, while decay and gain remain the largest error sources. These findings support separating modal-density estimation from continuous parameter fitting.
Minhui Lu, Joshua D. Reiss
Centre for Digital Music Queen Mary University of London London, United Kingdom
Simulation-based inference (SBI) with machine learning is an increasingly important tool for solving inverse problems in science and engineering, including parameter inference and the inversion of detector effects. We provide an overview of the Bayesian and frequentist statistical frameworks, describe how machine-learning-based SBI methods, such as neural posterior estimation and neural likelihood estimation, can be used for parameter estimation within these frameworks, and show that the same methods can also be applied to Empirical Bayes or unfolding tasks. We also discuss how to validate inference results and the limitations of SBI with machine learning.
Maximilian Dax, Theo Heimel, Gilles Louppe
ELLIS Institute Tübingen, Max Planck Institute for Intelligent Systems, Tübingen AI Center, Tübingen, Germany · CP3, Université catholique de Louvain, Louvain-la-Neuve, Belgium · Montefiore Institute, University of Liège, Belgium