pith. sign in

arxiv: 1906.09958 · v1 · pith:7WLZKOGLnew · submitted 2019-06-21 · 📡 eess.SP

Classification model for microphone type recognition

Miroslava Jordovic Pavlovic , Aleksandar Kupusinac , Marica Popovic This is my paper

Pith reviewed 2026-05-25 19:05 UTC · model grok-4.3

classification 📡 eess.SP
keywords microphone classificationphotoacoustic experimentsmultilayer perceptronneural networksimulated datasignal processingpattern recognition
0
0 comments X

The pith

A multilayer perceptron network trained on simulated data classifies microphone types for photoacoustic experiments while meeting accuracy, reliability, and real-time requirements.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper develops a classification model to identify microphone types used in photoacoustic experiments. It applies a multilayer perceptron network to a large set of simulated experimental values to create the model. The approach aims to provide a tool that operates accurately, reliably, and in real time, which are essential for practical photoacoustic work. A reader might care because proper microphone selection affects the quality of photoacoustic measurements, and an automated classifier could streamline the process.

Core claim

The classification model is obtained by applying a multilayer perceptron network on a large dataset of simulated experimental values. The model satisfies the basic requirements of a photoacoustic experiment: accuracy, reliability and real time operations.

What carries the argument

multilayer perceptron network applied to a large dataset of simulated experimental values for microphone type classification

Load-bearing premise

Simulated experimental values accurately capture the real-world variations and noise characteristics needed for reliable microphone type recognition in actual photoacoustic setups.

What would settle it

A direct comparison showing that the model's performance on real photoacoustic signals is substantially lower than on simulated data would indicate the approach does not transfer well.

Figures

Figures reproduced from arXiv: 1906.09958 by Aleksandar Kupusinac, Marica Popovic, Miroslava Jordovic Pavlovic.

Figure 1
Figure 1. Figure 1: Curves a) amplitude and b) phase of the distorted photoacoustic signal built upon the few records of the dataset used for network training for all three microphones In our previous research, we managed to isolate and correct deficiency of the measurement system, apropos the microphone, by neural network application [9]. In order to simplify the research, it was assumed that the microphone is the main part … view at source ↗
Figure 2
Figure 2. Figure 2: Two hidden layer ANN structure [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
read the original abstract

This paper presents a classification model for microphone type recognition in photoacoustic experiment. The classification model is obtained by applying a multilayer perceptron network on a large dataset of simulated experimental values. The model satisfies the basic requirements of a photoacoustic experiment: accuracy, reliability and real time operations.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 0 minor

Summary. The manuscript presents a classification model for microphone type recognition in photoacoustic experiments. It obtains the model by training a multilayer perceptron network on a large dataset of simulated experimental values and asserts that this model satisfies the requirements of accuracy, reliability, and real-time operations.

Significance. If the central claim were supported by evidence, the work could offer a practical ML-based approach to automate microphone identification in photoacoustic setups. However, the complete absence of any quantitative results, validation details, or generalization tests makes it impossible to assess whether the contribution has any technical significance.

major comments (3)
  1. [Abstract] Abstract: The claim that the model 'satisfies the basic requirements of a photoacoustic experiment: accuracy, reliability and real time operations' is unsupported by any metrics, validation procedure, baseline comparisons, confusion matrices, timing benchmarks, or error analysis.
  2. [Abstract] Abstract: No information is supplied on the simulation procedure, the microphone types being classified, the MLP architecture, training hyperparameters, or any test of whether the simulated data reproduces the signal statistics and noise distributions of real photoacoustic hardware.
  3. [Abstract] Abstract: The generalization step from simulated data to actual experiments is asserted without any held-out real-data evaluation, leaving the central claim that the model works in 'photoacoustic experiment' unverified.

Simulated Author's Rebuttal

3 responses · 1 unresolved

We thank the referee for the comments. The points raised correctly note the absence of supporting details and evidence in the manuscript. We respond to each major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The claim that the model 'satisfies the basic requirements of a photoacoustic experiment: accuracy, reliability and real time operations' is unsupported by any metrics, validation procedure, baseline comparisons, confusion matrices, timing benchmarks, or error analysis.

    Authors: We agree that the manuscript provides no quantitative metrics or validation details to support the claim. The assertion is based on unreported internal tests. We will revise to include performance metrics, timing benchmarks, and basic validation results. revision: yes

  2. Referee: [Abstract] Abstract: No information is supplied on the simulation procedure, the microphone types being classified, the MLP architecture, training hyperparameters, or any test of whether the simulated data reproduces the signal statistics and noise distributions of real photoacoustic hardware.

    Authors: The manuscript is concise and omits these details. We will revise to describe the simulation procedure, microphone types, MLP architecture, training hyperparameters, and any checks on simulated signal fidelity. revision: yes

  3. Referee: [Abstract] Abstract: The generalization step from simulated data to actual experiments is asserted without any held-out real-data evaluation, leaving the central claim that the model works in 'photoacoustic experiment' unverified.

    Authors: The model was trained and evaluated only on simulated data. We will revise to state explicitly that no real experimental data was used for validation and that generalization to hardware remains untested. revision: partial

standing simulated objections not resolved
  • The lack of real experimental data to evaluate generalization from simulation to actual photoacoustic hardware.

Circularity Check

0 steps flagged

No circularity; standard MLP training on simulated data with no self-referential reductions or derivations

full rationale

The paper presents a classification model obtained by applying a multilayer perceptron network on a large dataset of simulated experimental values, claiming it satisfies accuracy, reliability and real-time requirements. No equations, derivations, fitted parameters renamed as predictions, self-citations, or ansatzes are described that would reduce any result to its inputs by construction. The process is a conventional supervised learning pipeline on external simulated data; the generalization claim to real experiments is an empirical assertion rather than a definitional or self-referential step. This is the most common honest non-finding for applied ML papers without mathematical derivations.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review; no free parameters, axioms, or invented entities are described in the provided text.

pith-pipeline@v0.9.0 · 5559 in / 899 out tokens · 20635 ms · 2026-05-25T19:05:21.550284+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

13 extracted references · 13 canonical work pages

  1. [1]

    Abu-Mostafa, Malik Magdon-Ismail , Hsuan-Tien Lin

    Yaser S. Abu-Mostafa, Malik Magdon-Ismail , Hsuan-Tien Lin. Learning From Data. AMLBook 2012

    work page 2012

  2. [2]

    The thermoelastic bending and thermal diffusion processes influence on photoacoustic signal generation using open photoacoustic cell technique,

    A. Somer and A. Novatski, “The thermoelastic bending and thermal diffusion processes influence on photoacoustic signal generation using open photoacoustic cell technique,” J. Appl. Phys., vol. 114, no. 6, p. 63503, 2013

    work page 2013

  3. [3]

    Photoacoustic Source Detection and Reflection Artif act Removal Enabled by Deep Learning,

    D. Allman, A. Reiter, and M. A. L. Bell, “Photoacoustic Source Detection and Reflection Artif act Removal Enabled by Deep Learning,” IEEE Trans. Med. Imaging, vol. 37, no. 6, pp. 1464–1477, 2018

    work page 2018

  4. [4]

    Laser Fluence Recognition Using Computationally Intelligent Pulsed Photoacoustics Within the Trace Gases Analysis,

    M. Lukić, Ćojbašić, M. D. Rabasović, D. D. Markushev, and D. M. Todorović, “Laser Fluence Recognition Using Computationally Intelligent Pulsed Photoacoustics Within the Trace Gases Analysis,” Int. J. Thermophys., vol. 38, no. 11, 2017

    work page 2017

  5. [5]

    Computationally intelligent pulsed photoacoustics,

    M. Lukić, Ž. Ćojbašić, M. D. Rabasović, and D. D. Markushev, “Computationally intelligent pulsed photoacoustics,” Meas. Sci. Technol., vol. 25, no. 12, p. 125203, 2014

    work page 2014

  6. [6]

    Photothermal Depth Profiling Of Optical Gradient M aterials By Neural Network,

    M. N. Popovic, D. Furundzic, and S. P. Galovic, “Photothermal Depth Profiling Of Optical Gradient M aterials By Neural Network,” Publ. Astron. Obs. Belgrade, vol. 89, no. May 2015, pp. 147–150, 2010

    work page 2015

  7. [7]

    Electro-Acustic Influence of the Measuring System on the Photo-Acustic Signal Amplitude and Phase in Frequency Domain,

    S. M. Aleksic, D. K. Markushev, D. S. Pantic, M. D. Rabasovic, D. D. Markushev, and D. M. Todorović, “Electro-Acustic Influence of the Measuring System on the Photo-Acustic Signal Amplitude and Phase in Frequency Domain,” Univ. Facta, vol. 8, no. 1, pp. 1–13, 2010

    work page 2010

  8. [8]

    Helmholtz Resonances in Photoacoustic Experiment with Laser-Sintered Polyamide Including Thermal Memory of Samples,

    M. N. Popovic et al., “Helmholtz Resonances in Photoacoustic Experiment with Laser-Sintered Polyamide Including Thermal Memory of Samples,” Int. J. Thermophys., vol. 37, no. 12, pp. 1–9, 2016

    work page 2016

  9. [9]

    Artificial NeuralNetworks Application In Solid State Photoacoustics based on microphone response recognition in the frequency domain

    M. Jordović-Pavlović, M. Stanković, M. Popović, Ž. Ć ojbaš ić and D. Galović, Slobodanka; Markushev, “Artificial NeuralNetworks Application In Solid State Photoacoustics based on microphone response recognition in the frequency domain” , Journal of Computational Electronics, Springer, unpublished

    work page

  10. [10]

    Deep learning,

    Y. G. H. Yann LeCun; Bengio, “Deep learning,” Nature, vol. 13, no. 1, p. 35, 2015

    work page 2015

  11. [11]

    Understanding the difficulty of training deep feedforward neural networks,

    X. Glorot and Y. Bengio, “Understanding the difficulty of training deep feedforward neural networks,” in Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statist ics, 2010, vol. 9, pp. 249–256

    work page 2010

  12. [12]

    Adam: {A} Method for Stochastic Optimization,

    D. P. Kingma and J. Ba, “Adam: {A} Method for Stochastic Optimization,” CoRR, vol. abs/1412.6, 2014

    work page 2014

  13. [13]

    https://www.tensorflow.org, last visited on March, 30th, 2019

    work page 2019