Parameter-Efficient Fine-Tuning of Foundation Models for CLP Speech Classification

H.S. Shekhawat; Jagabandhu Mishra; S. R. Mahadeva Prasanna; Susmita Bhattacharjee

REVIEW 3 major objections 5 minor 37 references

Parameter-Efficient Fine-Tuning of Foundation Models for CLP Speech Classification

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Fine-tuning Whisper with LoRA or DoRA adapters improves cleft lip and palate speech severity classification over frozen-embedding and handcrafted-feature baselines on two corpora.

desk verdict Useful first benchmark of PEFT adapters for CLP severity, but the evaluation split may leak speaker identity; needs speaker-independent confirmation before the headline gains are credible. read the letter →

arxiv 2507.14898 v1 pith:FTZYAGAX submitted 2025-07-20 eess.AS eess.SP

Susmita Bhattacharjee , Jagabandhu Mishra , H.S. Shekhawat , S. R. Mahadeva Prasanna This is my paper

classification eess.ASeess.SP

keywords cleftlipandpalatespeechhypernasalityseverityclassificationparameter-efficientfine-tuningLoRADoWhisperself-supervisedmodelsclinicaldiagnostics

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper argues that fine-tuning a speech foundation model with small low-rank adapters improves detection and severity grading of cleft lip and palate (CLP) speech, a congenital condition whose nasalization increases with severity. The authors compare frozen embeddings from Wav2Vec2, WavLM, and Whisper, plus handcrafted eGeMAPS and ComParE features with SVM classifiers, and then fine-tune Whisper with LoRA and DoRA adapters. On the English corpus, DoRA raises severity macro-F1 from 0.53 with frozen Whisper and 0.41 with eGeMAPS to 0.67; on the Kannada corpus, LoRA raises it from 0.49 and 0.34 to 0.52. The reported relative severity-classification gains are 26.4% and 63.4% on the English corpus and 6.1% and 52.9% on the Kannada corpus over the best foundation-model and handcrafted-feature baselines. The implication is that parameter-efficient fine-tuning offers a low-resource route to adapting large speech models to disordered child speech.

What carries the argument

The carrying mechanism is low-rank adaptation. LoRA freezes the pre-trained weight matrix $W_0$ and learns a low-rank update $\Delta W = BA$ with rank $r=8$, applied to the query, key, and value projections of Whisper's transformer; DoRA refines this by normalizing the direction of the update and learning a separate magnitude component, giving $W = M \cdot D$. A mean-pooled utterance representation from the last encoder layer feeds a fully connected classifier, and cross-entropy loss trains only the adapters and the classification head while the rest of the encoder stays frozen. This lets a 768-dimensional Whisper encoder adapt to child CLP speech with very little trainable capacity.

What would settle it

Re-run the English and Kannada comparisons with speaker-disjoint splits, placing all utterances of a given speaker in only one partition, and check whether LoRA and DoRA macro-F1 advantages over frozen Whisper and eGeMAPS survive; if the gains shrink toward zero, the reported improvements are largely speaker-identity effects.

Watch

Extended reading notes

Core claim

The central claim is that low-rank adapters applied to Whisper's attention projections recover more CLP-relevant speech information than frozen embeddings or traditional acoustic features, and do so with only a small number of trainable parameters. Concretely, Whisper with DoRA reaches macro-F1 0.67 for four-class severity classification on the English corpus, and Whisper with LoRA reaches 0.52 on the Kannada corpus, versus 0.53 and 0.49 for frozen Whisper embeddings with an SVM, and 0.41 and 0.34 for eGeMAPS features with an SVM. The authors conclude that parameter-efficient fine-tuning improves both CLP detection and severity grading on both datasets, with DoRA giving the best results on English and LoRA giving the best results on Kannada.

Load-bearing premise

The load-bearing premise is that randomly dividing 65 or 60 speakers' utterances into training, development, and evaluation sets never places the same speaker in more than one partition, so the adapters learn CLP severity rather than speaker identity.

Editorial extensions

If this is right

CLP severity grading can be improved while training only adapter weights and a classification head, so the large Whisper encoder itself does not need to be stored or retrained per task.
Frozen embeddings from foundation models leave CLP-relevant signal unused; the reported gains show that this residual information is recoverable with modest trainable capacity.
The best adapter differs by corpus, with DoRA ahead on English and LoRA ahead on Kannada, so practical systems should select adapters per language or per corpus rather than assuming one method wins everywhere.
The same recipe could transfer to other speech-disorder grading tasks with small child-speech corpora, since it does not require full-model fine-tuning.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A speaker-disjoint evaluation is the natural next test: because the paper only says the data were randomly divided, the reported gains could partly reflect the model memorizing speakers rather than CLP severity.
The paper attributes the DoRA/LoRA reversal to language differences, but the two corpora also differ in size and class balance, so the reversal may instead reflect those factors; the paper does not disentangle them.
The comparison covers frozen embeddings and handcrafted features but not full fine-tuning of Whisper, so the efficiency-accuracy trade-off of adapters relative to updating all Whisper parameters remains open.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Quick read on arXiv:2507.14898. The useful new bit: it is the first direct comparison of Whisper, Wav2Vec2, and WavLM embeddings on CLP detection and severity grading, and the first application of LoRA/DoRA adapters to Whisper for that task. On two child speech datasets (English NMCPC, Kannada AIISH) it shows Whisper embeddings beat the self-supervised models, and fine-tuning Whisper with LoRA or DoRA nudges macro-F1 up further — from 0.53 to 0.65/0.67 on NMCPC severity and 0.49 to 0.51/0.52 on AIISH. The eGeMAPS-vs-ComParE comparison is also clean and useful. The paper is straightforward and mostly honest in its claims.

The soft spot is the one you'd expect and it is load-bearing. Section III-A says only that utterances were 'randomly divided' into train/dev/eval. Severity is a speaker-level label — each of the 65 NMCPC and 60 AIISH speakers gets one severity rating, and with 22 and 45 utterances per speaker on average, a random utterance-level split will almost certainly put the same speaker in both training and evaluation. A Whisper adapter can then learn speaker identity and copy the label, and the reported F1 gains (26.4% relative over Whisper-SVM on NMCPC, 6.1% on AIISH) would be inflated. The paper never states that the split is speaker-independent. This has to be fixed before the central comparison is interpretable. It may be that the original datasets come with standard speaker-independent partitions; if so, the authors need to say so and re-run.

Two smaller issues: there are no confidence intervals or repeated runs — one partition, one model selection via 5-fold CV, then a single evaluation — so the 0.02 F1 differences between LoRA and DoRA on AIISH are within noise. And the DoRA description in Sec. II-B is garbled (talk of pruning single-rank components is not how DoRA works), though the update equation is roughly right. The trainable classification head on top of PEFT vs. frozen SVM baselines is a mild confound, not fatal.

Bottom line: the paper is a legitimate new application of off-the-shelf PEFT to a small clinical speech task. The empirical claim is plausible but not statistically secured, and the speaker-leakage risk is real. A serious referee should see it, because the fix is straightforward and the recipe could be useful for low-resource speech pathology triage. I'd recommend conditional acceptance after the authors confirm speaker-independent evaluation or re-do the split and report variance.

Referee Report

3 major / 5 minor

Summary. The paper evaluates CLP detection and four-class severity classification from speech using two datasets (NMCPC, English; AIISH, Kannada). It compares traditional handcrafted features (eGeMAPS, ComParE, merged) with an SVM classifier against frozen-embedding representations from Whisper, Wav2Vec2, and WavLM, also paired with SVMs. The best frozen model, Whisper, is then fine-tuned with LoRA and DoRA adapters, adding a classification head and freezing the rest of the encoder. The central claim is that this PEFT approach yields relative macro-F1 improvements over the best foundation-model baseline and the best handcrafted-feature baseline of 26.4% and 63.4% on NMCPC and 6.1% and 52.9% on AIISH for severity classification, with smaller improvements for binary CLP detection (Table I). The paper positions the contribution as the first comprehensive comparison of such foundation-model representations for CLP speech, plus a PEFT adaptation strategy.

Significance. If the reported results are statistically secure, the paper would provide useful evidence that lightweight adapters on a weakly supervised speech model can outperform both conventional acoustic features and frozen large-model embeddings for a low-resource clinical speech task. The choice of two typologically different languages (English and Kannada), the systematic comparison across three foundation models, and the use of macro-F1 as the primary metric are strengths. The adaptive component is modest in scale (rank 8 adapters), which makes the approach practically attractive for clinical deployments. However, the current manuscript does not demonstrate that the central comparison survives a speaker-independent evaluation, and it provides no uncertainty quantification. The claimed gains, especially the 6.1% improvement on AIISH severity, could easily be within run-to-run or partitioning noise. The paper would be strengthened substantially by speaker-disjoint splits, repeated runs with confidence intervals, and significance tests.

major comments (3)

[Sec. III-A, Table I] The paper states only that each dataset was 'randomly divided' into training, development, and evaluation sets, without specifying a speaker-independent split. Severity labels are speaker-level attributes: all utterances from one speaker carry the same severity label, and the datasets have only 65 (NMCPC) and 60 (AIISH) speakers. A random utterance-level partition will almost certainly place utterances from the same speaker in both training and evaluation, allowing models to exploit speaker identity rather than generalize to unseen talkers. This is load-bearing because the headline relative improvements in Table I are meant to demonstrate that PEFT adapters improve severity classification over frozen-embedding and handcrafted baselines. The authors must either confirm that the splits are speaker-disjoint or rerun the evaluation with speaker-independent partitioning (e.g., leave-speakers-out cross-validation).
[Sec. IV, Table I] All results are reported as single point estimates with no confidence intervals, standard deviations, or significance tests. Several key comparisons are very close, for example Whisper LoRA vs. Whisper on AIISH severity (0.52 vs. 0.49), Whisper DoRA vs. Whisper LoRA on AIISH detection (0.93 vs. 0.94), and Whisper DoRA vs. LoRA on NMCPC severity (0.67 vs. 0.65). Without repeated runs or a paired test such as McNemar's test or a bootstrap over utterances/speakers, the stated relative improvements are not statistically secured. The authors should report mean and variance over multiple random seeds and, ideally, over multiple speaker-disjoint resamplings, and they should assess which pairwise differences are significant.
[Sec. III, experimental protocol] The description of model selection and evaluation is ambiguous. The text says, 'For all experiments, we pool the training and development sets and apply 5-fold cross-validation. We select the model that achieves the highest macro-averaged F1 score for final evaluation on the evaluation set.' It is not explained how the development set is used for the PEFT models, how many training epochs are run, whether early stopping is applied, or how the 5-fold CV interacts with the fixed train/dev/eval partition. This makes the experimental protocol irreproducible and makes it unclear whether the evaluation-set numbers are selected with any form of peeking. The authors should provide a precise protocol: what is tuned, what is selected, and how the final evaluation set is used.

minor comments (5)

[Sec. II-B, Eq. (2)] In the DoRA description, the definition 'M = ||W0 + BA||' is ambiguous because it is not stated whether M is a scalar, a per-column vector, or a matrix. The original DoRA formulation uses a magnitude vector per output dimension; please clarify the notation used here.
[Sec. III-D] The optimizer is spelled 'AdaM'; it should be 'Adam'. Also, the number of training epochs and any early-stopping criterion are not reported.
[Sec. III-A] Please report the number of utterances per speaker or at least the range, since the risk of speaker leakage depends on how many utterances each speaker contributes. Reporting the random seed or stating that the split is speaker-disjoint would also aid reproducibility.
[Sec. IV-A] The sentence 'These results clearly show that eGeMAPS provides better discrimination between normal and CLP speech than ComParE and its combination with eGeMAPS' is confusing because the merged set is a superset of eGeMAPS; the intended claim is presumably that the reduced, curated eGeMAPS set outperforms the larger merged set.
[References] Reference [8] (Dhananjaya and Yegnanarayana, 'Speaker change detection in casual conversations using excitation source features') appears unrelated to the claim about excitation characteristics in CLP speech; please replace it with an appropriate source on excitation source features in disordered speech.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the PEFT comparison is an empirical benchmark result, not a derivation that reduces to its inputs.

full rationale

The paper's claim is an empirical comparison: frozen Whisper embeddings paired with SVM and eGeMAPS/ComParE baselines are evaluated on held-out partitions, and LoRA/DoRA adapters are trained with cross-entropy on the same partitions, with results reported as F1 scores (Sec. IV-B, Table I). No predicted quantity is defined in terms of a fitted parameter; the baselines and the PEFT models are separate pipelines, and the relative improvements are computed from independent evaluation-set scores. Self-citations to prior CLP work ([4], [15], [24]) appear only as dataset sources and prior context, not as justification of the reported gains. The only notable concern is the dataset split description ('randomly divided into training, development, and evaluation sets,' Sec. III-A), which does not state speaker independence; that is a potential validity threat (speaker leakage), not a circularity, because the measured F1 values are not constructed from the labels or from the cited results. Under the circularity criteria, the derivation chain is self-contained and empirically falsifiable, so the score is 0.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

No new entities, particles, forces, or conserved quantities are introduced. The free parameters are standard hyperparameters for the PEFT method and the PCA baseline. The key domain assumptions are that CLP speech has learnable acoustic correlates and that evaluation partitions are speaker-independent.

free parameters (3)

LoRA/DoRA rank = 8
Adapter rank set to 8 for all experiments (Sec. III-D); no sensitivity analysis is reported, so the choice is a hand-picked hyperparameter that affects the central result.
Learning rate = 8e-5
Adam learning rate fixed at 8e-5 (Sec. III-D) without a reported tuning protocol; it is a free hyperparameter for the fine-tuning result.
PCA dimension for ComParE/merged features = 100
ComParE (6,373 dims) and merged (6,461 dims) features are reduced to 100 dimensions by PCA (Sec. III-B); this dimensionality is chosen without stated justification and affects baseline performance.

assumptions (3)

domain assumption CLP speech exhibits consistent acoustic correlates (nasalization, glottal stops, altered formants) that are learnable from short utterances
Stated in the Introduction as the motivation for speech-based CLP classification; if this premise fails, the entire task framing collapses, but it is a standard clinical observation.
domain assumption Foundation model representations capture task-relevant information even though the models are trained on adult, unimpaired speech
The paper's core hypothesis (Sec. I, II-C) is that Whisper/Wav2Vec2/WavLM embeddings carry features useful for CLP; this is not proven independently and is the premise being tested.
domain assumption The evaluation partition is representative and does not share speakers across train and eval
Sec. III-A describes random division of utterances without stating speaker independence; the validity of all reported F1 scores depends on this assumption.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Parameter-Efficient Fine-Tuning of Foundation Models for CLP Speech Classification." pith.science (2026). https://pith.science/paper/FTZYAGAX

@misc{pith2026250714898,
  author       = {Pith},
  title        = {Pith review of: Parameter-Efficient Fine-Tuning of Foundation Models for CLP Speech Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FTZYAGAX}},
  note         = {Machine review of arXiv:2507.14898}
}

read the original abstract

We propose the use of parameter-efficient fine-tuning (PEFT) of foundation models for cleft lip and palate (CLP) detection and severity classification. In CLP, nasalization increases with severity due to the abnormal passage between the oral and nasal tracts; this causes oral stops to be replaced by glottal stops and alters formant trajectories and vowel space. Since foundation models are trained for grapheme prediction or long-term quantized representation prediction, they may better discriminate CLP severity when fine-tuned on domain-specific data. We conduct experiments on two datasets: English (NMCPC) and Kannada (AIISH). We perform a comparative analysis using embeddings from self-supervised models Wav2Vec2 and WavLM, and the weakly supervised Whisper, each paired with SVM classifiers, and compare them with traditional handcrafted features eGeMAPS and ComParE. Finally, we fine-tune the best-performing Whisper model using PEFT techniques: Low-Rank Adapter (LoRA) and Decomposed Low-Rank Adapter (DoRA). Our results demonstrate that the proposed approach achieves relative improvements of 26.4% and 63.4% in macro-average F1 score over the best foundation model and handcrafted feature baselines on the NMCPC dataset, and improvements of 6.1% and 52.9% on the AIISH dataset, respectively.

Figures

Figures reproduced from arXiv: 2507.14898 by the authors.

**Figure 4.** illustrates the overall experimental framework. We evaluate model performance using both accuracy and macroaveraged F1 score, giving special importance to the macroaveraged F1 score due to class imbalance. For all experiments, we pool the training and development sets and apply 5-fold cross-validation. We select the model that achieves the highest macro-averaged F1 score for final evaluation on the evaluation set.… view at source ↗

**Figure 3.** Proposed framework: Audio input processed by a [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 5.** Figure 5: Performance comparison (F1-scores) of the best [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

37 extracted references · 31 canonical work pages

[1]

Predicting hypernasality using spectrogram via deep convolutional neural network (dcnn)

Murni Mohd Amir, Ani Liza Asnawi, Nur Aishah Zainal, and Ah- mad Zamani Jusoh. Predicting hypernasality using spectrogram via deep convolutional neural network (dcnn). 2024 IEEE International Conference on Computing (ICOCO) , pages 398–403, 2024

work page 2024
[2]

Wav2vec 2.0: A framework for self-supervised learning of speech representations

Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, and Michael Auli. Wav2vec 2.0: A framework for self-supervised learning of speech representations. In Proceedings of the 34th International Conference on Neural Information Processing Systems, NIPS’20, Red Hook, NY , USA,

work page
[3]

Bayerl, Elmar Noth, Korbinian Riedhammer, and Tobias Bocklet

Ilja Baumann, Dominik Wagner, Franziska Braun, Sebastian P. Bayerl, Elmar Noth, Korbinian Riedhammer, and Tobias Bocklet. Influence of utterance and speaker characteristics on the classification of children with cleft lip and palate. INTERSPEECH 2023, 2022

work page 2023
[4]

Sensitivity analysis of maskcy- clegan based voice conversion for enhancing cleft lip and palate speech recognition

Susmita Bhattacharjee and Rohit Sinha. Sensitivity analysis of maskcy- clegan based voice conversion for enhancing cleft lip and palate speech recognition. In 2022 International Conference on Signal Processing and Communications (SPCOM), pages 1–5, July 2022

work page 2022
[5]

Wavlm: Large-scale self-supervised pre-training for full stack speech processing

Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Micheal Zeng, and Furu Wei. Wavlm: Large-scale self-supervised pre-training for full stack speech processing. IEEE Journal of Selected Topics in Signal Processing, 16:1505–1518...

work page 2021
[6]

Automatic classification of severity of articulation disorders in dysarthric speech

H Christensen, S Cunningham, C Fox, and P Green. Automatic classification of severity of articulation disorders in dysarthric speech. In Proceedings of Interspeech , 2012

work page 2012
[7]

Support-vector networks

Corinna Cortes and Vladimir Vapnik. Support-vector networks. Machine Learning, 20(3):273–297, 1995

1995
[8]

Speaker change detection in casual conversations using excitation source features

N Dhananjaya and B Yegnanarayana. Speaker change detection in casual conversations using excitation source features. Speech communication , 50(2):153–161, 2008

work page 2008

Show all 37 references

[9]

Akhilesh Kumar Dubey, S. R. Mahadeva Prasanna, and Samarendra Dandapat. Detection and assessment of hypernasality in repaired cleft palate speech using vocal tract and residual features. The Journal of the Acoustical Society of America , 146 6:4211, 2019

2019
[10]

Akhilesh Kumar Dubey, Deepak Kumar Singh, and B. B. Tiwari. Hypernasality severity analysis using spectral and residual features. 2021 10th International Conference on Internet of Everything, Microwave Engineering, Communication and Networks (IEMECON) , pages 1–6, 2021

2021
[11]

Scherer, Bj ¨orn Schuller, Johan Sundberg, Elisabeth Andr ´e, Carlos Busso, Laurence Devillers, Julien Epps, Petri Laukka, Shrikanth S

Florian Eyben, Klaus R. Scherer, Bj ¨orn Schuller, Johan Sundberg, Elisabeth Andr ´e, Carlos Busso, Laurence Devillers, Julien Epps, Petri Laukka, Shrikanth S. Narayanan, and Khiet P. Truong. The geneva minimalistic acoustic parameter set (gemaps) for voice research and affect...

2016
[12]

Parismita Gogoi, Vishwanath Pratap Singh, Seema Khadirnaikar, Soma Siddhartha, Sishir Kalita, Jagabandhu Mishra, Md Sahidullah, Priyankoo Sarmah, and S. R. M. Prasanna. Leveraging AM and FM rhythm spectrograms for dementia classification and assessment, 2025

2025
[13]

Parameter-efficient fine-tuning for large models: A comprehensive sur- vey

Zeyu Han, Chao Gao, Jinyang Liu, Jeff Zhang, and Sai Qian Zhang. Parameter-efficient fine-tuning for large models: A comprehensive sur- vey. Trans. Mach. Learn. Res. , 2024, 2024

2024
[14]

Edward Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, and Weizhu Chen

J. Edward Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, and Weizhu Chen. Lora: Low-rank adaptation of large language models. ArXiv, abs/2106.09685, 2021

2021 arXiv
[15]

Single frequency filter bank based long-term average spectra for hypernasality detection and assessment in cleft lip and palate speech

Mohammad Hashim Javid, Krishna Gurugubelli, and Anil Kumar Vup- pala. Single frequency filter bank based long-term average spectra for hypernasality detection and assessment in cleft lip and palate speech. In ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speec...

2020
[16]

Kingma and Jimmy Ba

Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization. CoRR, abs/1412.6980, 2014

2014 arXiv
[17]

Ann W. Kummer. Cleft Palate and Craniofacial Anomalies: Effects on Speech and Resonance . Delmar Cengage Learning, 2007

2007
[18]

Scaling down to scale up: A guide to parameter-efficient fine-tuning

Vladislav Lialin, Vijeta Deshpande, and Anna Rumshisky. Scaling down to scale up: A guide to parameter-efficient fine-tuning. ArXiv, abs/2303.15647, 2023

2023 arXiv
[19]

Dora: Weight-decomposed low-rank adaptation

Shih-Yang Liu, Chien-Yi Wang, Hongxu Yin, Pavlo Molchanov, Yu- Chiang Frank Wang, Kwang-Ting Cheng, and Min-Hung Chen. Dora: Weight-decomposed low-rank adaptation. ArXiv, abs/2402.09353, 2024

2024 arXiv
[20]

Methodology for perceptual assessment of speech in patients with cleft palate: A critical review of the literature

Anette Lohmander and Maria Olsson. Methodology for perceptual assessment of speech in patients with cleft palate: A critical review of the literature. The Cleft Palate-Craniofacial Journal , 41:64 – 70, 2004

2004
[21]

Mathad, Nancy J Scherer, Kathy Chapman, Julie M

Vikram C. Mathad, Nancy J Scherer, Kathy Chapman, Julie M. Liss, and Visar Berisha. A deep learning algorithm for objective assessment of hypernasality in children with cleft palate. IEEE Transactions on Biomedical Engineering, 68:2986–2996, 2020

2020
[22]

John Moody and Christian J. Darken. Fast learning in networks of locally-tuned processing units. Neural Computation , 1(2):281–294, 1989

1989
[23]

Enhanced cleft lip and palate classification using siglip 2: A comparative study with vision transformers and siamese networks

Oraphan Nantha, Benjaporn Sathanarugsawait, and Prasong Praneet- polgrang. Enhanced cleft lip and palate classification using siglip 2: A comparative study with vision transformers and siamese networks. Applied Sciences, 2025

2025
[24]

Hypernasality severity analysis in cleft lip and palate speech using vowel space area

K Nikitha, Sishir Kalita, CM Vikram, M Pushpavathi, and SR Mahadeva Prasanna. Hypernasality severity analysis in cleft lip and palate speech using vowel space area. In Interspeech, pages 1829–1833, 2017

2017
[25]

Cleft Palate Speech

& Karnell Peterson-Falzone, Hardin-Jones. Cleft Palate Speech . Springer, 2010

2010
[26]

Robust speech recognition via large-scale weak supervision

Alec Radford, Jong Wook Kim, Tao Xu, Greg Brockman, Christine McLeavey, and Ilya Sutskever. Robust speech recognition via large-scale weak supervision. In International Conference on Machine Learning , 2022

2022
[27]

Speaker change detection using excitation source and vocal tract system information

Mousmita Sarma, Sree Nilendra Gadre, Biswajit Dev Sarma, and SR Ma- hadeva Prasanna. Speaker change detection using excitation source and vocal tract system information. In 2015 Twenty First National Conference on Communications (NCC) , pages 1–6. IEEE, 2015

2015
[28]

Scherer, Fabien Ringeval, Mohamed Chetouani, Felix Weninger, Florian Eyben, Erik Marchi, Marcello Mortillaro, Hugues Salamin, Anna Polychroniou, Fabio Valente, and Samuel Kim

Bj ¨orn Schuller, Stefan Steidl, Anton Batliner, Alessandro Vincia- relli, Klaus R. Scherer, Fabien Ringeval, Mohamed Chetouani, Felix Weninger, Florian Eyben, Erik Marchi, Marcello Mortillaro, Hugues Salamin, Anna Polychroniou, Fabio Valente, and Samuel Kim. The interspeech 2...

2013
[29]

Govind, Samudravijaya K., and Suryakanth V

Kilaru Sireesha, Akhilesh Kumar Dubey, D. Govind, Samudravijaya K., and Suryakanth V . Gangashetty. Variational mode decomposition based features for detection of hypernasality in cleft palate speech. Biomedical Signal Processing and Control , 97:106689, 2024

2024
[30]

Cleft palate: The nature and remediation of communication problems

Jackie Stengelhofen. Cleft palate: The nature and remediation of communication problems. Churchill Livingstone, 1993

1993
[31]

Protima Nomo Sudro, Rohan Kumar Das, Rohit Sinha, and S. R. Mahadeva Prasanna. Significance of data augmentation for improving cleft lip and palate speech recognition. In 2021 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC),...

2021
[32]

VikramC., Ayush Tripathi, Sishir Kalita, and S

M. VikramC., Ayush Tripathi, Sishir Kalita, and S. R. Mahadeva Prasanna. Estimation of hypernasality scores from cleft lip and palate speech. In Interspeech, 2018

2018
[33]

Global, regional and national burden of orofacial clefts from 1990 to 2019: an analysis of the global burden of disease study 2019

Dawei Wang, Boyu Zhang, Qi Zhang, and Yiping Wu. Global, regional and national burden of orofacial clefts from 1990 to 2019: an analysis of the global burden of disease study 2019. Annals of Medicine , 55, 05 2023

1990
[34]

Whitehill and Cynthia H F Chau

Tara L. Whitehill and Cynthia H F Chau. Single-word intelligibility in speakers with repaired cleft palate. Clinical Linguistics & Phonetics , 18:341 – 355, 2004

2004
[35]

D. J. Zajac and L. D. Vallino. Evaluation and Management of Cleft Lip and Palate: A Developmental Perspective . Plural Publishing, 2017

2017
[36]

Zajac, Caitrin Plante, Amanda Lloyd, and Katarina L

David J. Zajac, Caitrin Plante, Amanda Lloyd, and Katarina L. Haley. Reliability and validity of a computer-mediated, single-word intelligi- bility test: Preliminary findings for children with repaired cleft lip and palate. The Cleft Palate-Craniofacial Journal , 48:538 – 549, 2011. 6

2011
[2020]

Curran Associates Inc

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

Predicting hypernasality using spectrogram via deep convolutional neural network (dcnn)

Murni Mohd Amir, Ani Liza Asnawi, Nur Aishah Zainal, and Ah- mad Zamani Jusoh. Predicting hypernasality using spectrogram via deep convolutional neural network (dcnn). 2024 IEEE International Conference on Computing (ICOCO) , pages 398–403, 2024

work page 2024

[2] [2]

Wav2vec 2.0: A framework for self-supervised learning of speech representations

Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, and Michael Auli. Wav2vec 2.0: A framework for self-supervised learning of speech representations. In Proceedings of the 34th International Conference on Neural Information Processing Systems, NIPS’20, Red Hook, NY , USA,

work page

[3] [3]

Bayerl, Elmar Noth, Korbinian Riedhammer, and Tobias Bocklet

Ilja Baumann, Dominik Wagner, Franziska Braun, Sebastian P. Bayerl, Elmar Noth, Korbinian Riedhammer, and Tobias Bocklet. Influence of utterance and speaker characteristics on the classification of children with cleft lip and palate. INTERSPEECH 2023, 2022

work page 2023

[4] [4]

Sensitivity analysis of maskcy- clegan based voice conversion for enhancing cleft lip and palate speech recognition

Susmita Bhattacharjee and Rohit Sinha. Sensitivity analysis of maskcy- clegan based voice conversion for enhancing cleft lip and palate speech recognition. In 2022 International Conference on Signal Processing and Communications (SPCOM), pages 1–5, July 2022

work page 2022

[5] [5]

Wavlm: Large-scale self-supervised pre-training for full stack speech processing

Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Micheal Zeng, and Furu Wei. Wavlm: Large-scale self-supervised pre-training for full stack speech processing. IEEE Journal of Selected Topics in Signal Processing, 16:1505–1518...

work page 2021

[6] [6]

Automatic classification of severity of articulation disorders in dysarthric speech

H Christensen, S Cunningham, C Fox, and P Green. Automatic classification of severity of articulation disorders in dysarthric speech. In Proceedings of Interspeech , 2012

work page 2012

[7] [7]

Support-vector networks

Corinna Cortes and Vladimir Vapnik. Support-vector networks. Machine Learning, 20(3):273–297, 1995

1995

[8] [8]

Speaker change detection in casual conversations using excitation source features

N Dhananjaya and B Yegnanarayana. Speaker change detection in casual conversations using excitation source features. Speech communication , 50(2):153–161, 2008

work page 2008