Refining Transcripts With TV Subtitles by Prompt-Based Weakly Supervised Training of ASR

Hugo Van hamme; Xinnian Zhao

REVIEW 2 major objections 5 minor 20 references

Refining Transcripts With TV Subtitles by Prompt-Based Weakly Supervised Training of ASR

T0 review · 2 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read TV subtitles, used as prompts, let Whisper refine its own transcripts with no verbatim labels.

desk verdict Subtitle-prompted self-training is a genuinely fresh idea worth refereeing, but the headline WER is tainted by test-set-selected hyperparameters. read the letter →

arxiv 2509.04491 v1 pith:BXPERG4B submitted 2025-09-01 cs.CL cs.AI

Xinnian Zhao , Hugo Van Hamme This is my paper · ORCID

classification cs.CLcs.AI

keywords weaklysupervisedASRsubtitlepromptingWhisperfine-tuningpseudo-transcriptsGinicoefficientweightedattentionFlemishbroadcastrarewords

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

The paper tries to establish that misaligned TV subtitles, which are too noisy to serve as training targets, can instead be used as contextual prompts while pseudo-transcripts generated by Whisper serve as targets, letting the model iteratively refine its own output. On 760 hours of Flemish broadcast audio, this recipe lowers word error rate from 18.75% to 10.34% on a verbatim-annotated test set, with no manual transcriptions used in training. The authors also introduce an inference-time weighted-attention mechanism that uses Gini coefficients of cross-attention to emphasize subtitle tokens relevant to the speech. If the method is right, it turns an abundant, cheaply available resource — subtitles — into a lever for improving ASR in low-resource domains.

What carries the argument

Subtitle prompting (SP) training: the decoder input is structured as <|sop|>subtitles<|sot|><|nl|><|transcribe|><|notimestamps|>, with pseudo-transcripts (generated by Whisper-large-v3) as targets, enabling weakly supervised fine-tuning. The inference-time weighted attention (WA) mechanism computes a Gini coefficient for each prompt token from the first cross-attention layer's attention distribution over speech frames; concentrated attention yields high Gini values, and these weights scale the prompt's key and value matrices before self-attention, emphasizing speech-relevant subtitle tokens and down-weighting distractors.

What would settle it

Train the same model on the same data with the subtitle prompt format but replace subtitle text with random or scrambled word sequences; if WER does not degrade relative to no-prompt fine-tuning, the claimed informative role of subtitles as prompts is not supported. Conversely, if verbatim transcripts used as prompts do not outperform subtitle prompts beyond the noise level of the subtitles themselves, the contextual value attributed to subtitles would be cast in doubt.

Watch

Extended reading notes

Core claim

The paper claims that a Whisper ASR model can be fine-tuned without verbatim transcripts by prepending subtitle text to the decoder input as a prompt (format <|sop|>subtitles<|sot|><|nl|><|transcribe|><|notimestamps|>), using the model's own pseudo-transcripts as targets, and iterating. Directly prompting a pre-trained Whisper with subtitles produces truncated or empty output, but fine-tuning with this prompt format teaches the model to treat subtitles as guidance rather than text to repeat or truncate against. Adding a weighted attention scheme at inference — scaling prompt key/value vectors by Gini coefficients computed from first-layer cross-attention weights — further improves WER. Acros

Load-bearing premise

Fine-tuning the decoder with the subtitle-prompt format teaches the model to treat subtitle text as guidance rather than text to copy or truncate, even though the paper gives no analysis of how this separation is learned.

Editorial extensions

If this is right

Subtitle-prompted fine-tuning turns abundant, loosely aligned subtitles into a usable signal for ASR improvement without verbatim labels.
Iterative retraining with updated pseudo-transcripts converges: WER decreases from 11.54% to 10.52% over three SP iterations, and further to 10.34% with WA inference.
The method especially helps rare and out-of-vocabulary words: rWER drops from 31.04% to 24.63% and oWER from 74.94% to 70.22% for the medium model with SP.
The Gini-weighted attention scheme consistently outperforms max- and entropy-based weighting, avoiding hallucinations and refining named-entity spelling.
Self-training without prompts harms the large model (WER up to 21.49%), while adding subtitle prompts reverses that degradation, showing prompts mitigate error propagation.

Reading between the lines

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

The method likely generalizes beyond Flemish: any language with available subtitles and a pre-trained generative ASR could use the same prompt-based weak supervision recipe.
The Gini-weighted attention mechanism could be applied to other prompt-conditioned generation tasks where prompts are noisy or loosely aligned with inputs, not just ASR.
A direct comparison against verbatim-supervised fine-tuning on the same data would clarify how close subtitle-prompting gets to fully supervised performance.
The mechanism of why fine-tuning teaches the model to separate prompt content from generation targets remains unexplained; ablation studies isolating prompt-format effects would sharpen the claim.

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

Short version: this is a genuinely different way to use TV subtitles for ASR—as decoder prompts instead of training targets—plus a Gini-weighted attention trick at inference. The core result, that prompt-conditioned self-training beats plain self-training on Flemish broadcast data, is plausible and probably real. But the headline 10.34% WER is not cleanly identified because the evaluation folds were used to choose both the layer configuration and the stopping iteration.

What's new and good: the idea that subtitles can constrain decoding without being verbatim labels is well motivated, and the authors' explanation of self-training failure—deletions at utterance starts, reinforced by the large model—is sensible. Using first-layer cross-attention concentration (Gini) to reweight prompt keys and values during inference is original, and the ablation against max- and entropy-based weights gives it some credibility. They also honestly report that directly prompting a pre-trained Whisper with subtitles produces truncated output, and they don't pretend the baseline is strong.

The biggest problem is experimental hygiene. Section V-B explicitly says they split the only verbatim test set (subs-annot) into folds to "ensure robust layer selection," and the all-layers configuration is then applied to the full set. Table IV then stops at three iterations after observing the test WER. That makes the improvement from 11.54% to 10.34% partly a selection artifact. The honest claim is "we can reach low 10s on this test set with test-set-informed choices," not "the method robustly improves WER by 1.2 absolute." An independent dev split or a pre-specified configuration would fix this.

Other soft spots: there is no baseline that simply uses subtitles as training targets, which is the natural alternative given the literature they cite; training details are thin (epochs, early stopping); no code or data is released; and the WA gain itself is modest (about 0.5 absolute WER) and only demonstrated on one model size.

Who this is for: ASR practitioners working on low-resource or under-resourced languages that have subtitle corpora. The paper deserves a serious referee—the idea is fresh, the flaws are fixable with better evaluation discipline, and the problem is practically important.

Recommendation: send it out, but the reviewers should push for a held-out dev set and a subtitle-as-target baseline before acceptance.

Referee Report

2 major / 5 minor

Summary. The paper proposes a weakly supervised ASR fine-tuning method that uses TV subtitles as decoder prompts rather than as direct training targets. Pseudo-transcripts generated by pre-trained Whisper serve as the training labels, and the model is fine-tuned with a decoder prompt of the form <|sop|>subtitles<|sot|><|nl|><|transcribe|><|notimestamps|>. The training is iterated, with pseudo-labels refreshed each round. At inference, a weighted-attention (WA) mechanism based on Gini coefficients computed from cross-attention weights re-weights the subtitle prompt keys/values in self-attention. Experiments on 760 hours of Flemish broadcast audio with a 6-hour manually annotated test set (subs-annot) report WER improvements from 18.75%/13.07% (pre-trained medium/large) to 11.54%/11.37% after one SP fine-tuning iteration, 11.02% with WA inference, and 10.34% after three iterative updates. The central claim is that subtitles as prompts, combined with pseudo-transcript targets, enable iterative transcript refinement without verbatim labels.

Significance. If valid, the method is practically useful for low-resource and under-represented domains where verbatim transcripts are unavailable but TV subtitles exist. The idea of converting subtitle mismatch into a prompt-conditioning problem is novel and clearly presented, and the reported rare-word and OOV error reductions are plausible and potentially valuable. The paper also explicitly discusses the relation to and difference from self-training, and it provides a concrete inference-time mechanism (Gini-weighted attention) that is inexpensive and model-agnostic. However, the empirical evidence for the headline numbers is weakened by the use of the evaluation set for configuration and stopping decisions, as detailed in the major comments. Because that issue is fixable with a held-out development split, the contribution is defensible pending revision.

major comments (2)

[V-B/V-C, Table IV] The headline WERs are obtained after using the evaluation set for model selection. Section V-B states that because "only the test set (subs-annot) contains verbatim transcripts, we divide this set into five folds to simulate different test scenarios and ensure robust layer selection"; the all-layers WA configuration is then applied to the full subs-annot set. The number of iterations is also chosen after observing test WER ("training is halted after three updates", Section V-C), and Table IV reports the resulting 10.34% WER. Since WA layer choice and iteration count are selected on the same set used to report the final accuracy, the monotonic decrease 11.54→10.82→10.52 and the inference WERs 11.02→10.66→10.34 are optimistically biased; the iter2→iter3 gain of 0.32% absolute may be noise. Fix: hold out a development portion of subs-annot (or use nested cross-validation) for layer/iteratio
[V-A, III-A] The mechanism underlying SP is not tested. The paper argues that pre-trained Whisper truncates with subtitle prompts, while SP fine-tuning learns to use subtitles as context; however, no analysis is provided to show that the model actually conditions on subtitle content rather than exploiting the prefix structure or copying subtitle tokens. The rWER/oWER improvements (Table II) are suggestive, but a control with shuffled or corrupted subtitle prompts (e.g., random tokens from the same distribution, or reversed subtitles) would test whether the content of the prompt matters. Without such a control, the claim that "subtitles provide additional information to guide generation" (Section V-A) remains an interpretation of the WER difference rather than a demonstrated property of the trained model.

minor comments (5)

[IV-A] The sentence "6 hours of speech (approximately 2,600 utterances) from each genre" is ambiguous: is it 6 hours per genre (24 hours total) or 6 hours overall divided across genres? Please clarify.
[III-B, Eq. (2)] In Eq. (2), G is described as the Gini coefficient over the prompt sequence, but if G is a vector, the product G Kp should be defined as element-wise (per-token) scaling of the key vectors, not standard matrix multiplication. Please clarify the notation and the exact dimension of G.
[IV-B] Training details are incomplete: the number of epochs, early stopping criterion, and the pseudo-transcript length filter threshold are not specified. These are needed for reproducibility.
[V-A, V-B] No confidence intervals or significance tests are reported. Given the small differences in Table IV (e.g., 0.32% between iter2 and iter3), error bars or a paired test would help assess whether the improvements are reliable.
[General] Typos and formatting issues: "frame work" in the Introduction, "W A" and "SP fine-tuning" spacing inconsistencies, "I NTRODUCTION" heading, and "subs-annot set" redundancy. Figure 2 caption should describe the colors and the meaning of the folds.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the core WER claims are evaluated against manually annotated, training-excluded references, so the reported gains are not forced by construction.

full rationale

The derivation is self-contained against an external benchmark. Pseudo-transcripts (Ypt) and Gini-weighted attention are internal to the model, but the paper never uses those internal quantities as its evidence of correctness: every headline WER (Tables I-IV) is computed against the manually annotated subs-annot set, which is explicitly 'excluded from the training set.' The paper also reports the failure mode that would appear if the procedure were merely self-confirming: no-prompt fine-tuning of Whisper-large raises WER to 21.49% (Table I), while subtitle prompting lowers it, and the iterative improvement in Table IV is an empirical trajectory, not a definitional consequence of the loss. The only noteworthy weakness is evaluation hygiene rather than circularity: in Section V-B, 'only the test set (subs-annot) contains verbatim transcripts, we divide this set into five folds to simulate different test scenarios and ensure robust layer selection,' and in Section V-C the iteration count is chosen after observing test WER ('training is halted after three updates'), meaning the final 10.34% may be optimistically biased by selection on the evaluation set. That is a statistical validity caveat, not a case of a prediction being identical to an input by construction; with no load-bearing self-citation and no imported uniqueness/ansatz, the circularity score is 0.

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

The ledger is mostly empty of fitted parameters because the method is a fine-tuning procedure, not a closed-form derivation. The central claim rests on domain assumptions about the reliability of attention-based relevance and the utility of pseudo labels, rather than on newly introduced free parameters or entities.

free parameters (3)

Number of refinement iterations = 3
Training is halted after three updates; no stopping criterion or sensitivity analysis is given, so the reported 10.34% WER depends on this hand-chosen stopping point.
Pseudo transcript length filter threshold = not specified
The paper removes hallucinated pseudo transcripts by filtering on transcript length but never gives the threshold, so the training target set depends on an unreported hand-chosen value.
WA layer configuration = all layers
The decision to apply Gini weights to all self-attention layers was made by observing the five folds of the test set (Section V-B), a post-hoc selection that affects the reported WA gain.

assumptions (3)

domain assumption The Gini coefficient of first-layer cross-attention weights measures subtitle token relevance to the audio.
This is the premise of the weighted attention mechanism (Section III-B, Eq. 1); no validation is provided beyond the held-out WER results.
domain assumption Fine-tuning with the decoder prefix <|sop|>subtitles<|sot|><|nl|><|transcribe|><|notimestamps|> teaches the model to use subtitle context without memorizing it.
This is the core SP training assumption (Section III-A); if the model memorized subtitle strings, the gains on held-out data would not appear.
domain assumption Pseudo transcripts generated by pre-trained Whisper-large-v3 are accurate enough to serve as training targets and improve with iteration.
The iterative refinement loop (Section V-C) depends on this; the paper shows that self-training without prompts can hurt for the large model, so the target quality is not guaranteed.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Refining Transcripts With TV Subtitles by Prompt-Based Weakly Supervised Training of ASR." pith.science (2026). https://pith.science/paper/BXPERG4B

@misc{pith2026250904491,
  author       = {Pith},
  title        = {Pith review of: Refining Transcripts With TV Subtitles by Prompt-Based Weakly Supervised Training of ASR},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BXPERG4B}},
  note         = {Machine review of arXiv:2509.04491}
}

read the original abstract

This study proposes a novel approach to using TV subtitles within a weakly supervised (WS) Automatic Speech Recognition (ASR) framework. Although TV subtitles are readily available, their imprecise alignment with corresponding audio limits their applicability as supervised targets for verbatim transcription. Rather than using subtitles as direct supervision signals, our method reimagines them as context-rich prompts. This design enables the model to handle discrepancies between spoken audio and subtitle text. Instead, generated pseudo transcripts become the primary targets, with subtitles acting as guiding cues for iterative refinement. To further enhance the process, we introduce a weighted attention mechanism that emphasizes relevant subtitle tokens during inference. Our experiments demonstrate significant improvements in transcription accuracy, highlighting the effectiveness of the proposed method in refining transcripts. These enhanced pseudo-labeled datasets provide high-quality foundational resources for training robust ASR systems.

Figures

Figures reproduced from arXiv: 2509.04491 by the authors.

**Figure 2.** WERs (%) on the 5 folds of subs-annot, obtained by applying Gini attention weights to individual layers or all layers of transformers. Gini weights are applied to all layers. Under this configuration, the WER on the entire subs-annot set improves from 11.54% to 11.02%. For ablation, we also experiment with max- and entropybased weights derived from the cross-attention weights between prompt tokens and speech frame… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

20 extracted references · 20 canonical work pages

[1]

Summary of chatgpt-related research and perspective towards the future of large language models,

Yiheng Liu, Tianle Han, Siyuan Ma, Jiayue Zhang, Yuanyuan Yang, Jiaming Tian, Hao He, Antong Li, Mengshen He, Zhengliang Liu, et al., “Summary of chatgpt-related research and perspective towards the future of large language models,” Meta-Radiology, p. 100017, 2023

work page 2023
[2]

Foun- dation models for generalist medical artificial intelligence,

Michael Moor, Oishi Banerjee, Zahra Shakeri Hossein Abad, Harlan M Krumholz, Jure Leskovec, Eric J Topol, and Pranav Rajpurkar, “Foun- dation models for generalist medical artificial intelligence,” Nature, vol. 616, no. 7956, pp. 259–265, 2023

work page 2023
[3]

Will affective computing emerge from foundation models and general artificial intelli- gence? a first evaluation of chatgpt,

Mostafa M. Amin, Erik Cambria, and Bj ¨orn W. Schuller, “Will affective computing emerge from foundation models and general artificial intelli- gence? a first evaluation of chatgpt,” IEEE Intelligent Systems , vol. 38, no. 2, pp. 15–23, 2023

work page 2023
[4]

A comprehensive survey of few-shot learning: Evolution, applications, challenges, and opportunities,

Yisheng Song, Ting Wang, Puyu Cai, Subrota K Mondal, and Jy- oti Prakash Sahoo, “A comprehensive survey of few-shot learning: Evolution, applications, challenges, and opportunities,” ACM Computing Surveys, vol. 55, no. 13s, pp. 1–40, 2023

work page 2023
[5]

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. PMLR, 2023, pp. 28492–28518

work page 2023
[6]

Analysis of whisper automatic speech recognition performance on low resource language,

Riefkyanov Surya Adia Pratama and Agit Amrullah, “Analysis of whisper automatic speech recognition performance on low resource language,” Jurnal Pilar Nusa Mandiri , vol. 20, no. 1, pp. 1–8, 2024

work page 2024
[7]

Comparison of multilingual self-supervised and weakly-supervised speech pre-training for adaptation to unseen languages,

Andrew Rouditchenko, Sameer Khurana, Samuel Thomas, Rogerio Feris, Leonid Karlinsky, Hilde Kuehne, David Harwath, Brian Kings- bury, and James Glass, “Comparison of multilingual self-supervised and weakly-supervised speech pre-training for adaptation to unseen languages,” in Proc. Interspeech 2023 , 2023, pp. 2268–2272

work page 2023
[8]

Benchmarking whisper for low-resource speech recognition: An n- shot evaluation on pashto, punjabi, and urdu,

Najm Ul Sehar, Ayesha Khalid, Farah Adeeba, and Sarmad Hussain, “Benchmarking whisper for low-resource speech recognition: An n- shot evaluation on pashto, punjabi, and urdu,” in Proceedings of the First Workshop on Challenges in Processing South Asian Languages (CHiPSAL 2025) , 2025, pp. 202–207

work page 2025

Show all 20 references

[9]

Exploring native and non-native english child speech recognition with whisper,

Rishabh Jain, Andrei Barcovschi, Mariam Yahayah Yiwere, Peter Cor- coran, and Horia Cucu, “Exploring native and non-native english child speech recognition with whisper,” IEEE Access , vol. 12, pp. 41601– 41610, 2024

2024
[10]

The mgb- 5 challenge: Recognition and dialect identification of dialectal arabic speech,

Ahmed Ali, Suwon Shon, Younes Samih, Hamdy Mubarak, Ahmed Abdelali, James Glass, Steve Renals, and Khalid Choukri, “The mgb- 5 challenge: Recognition and dialect identification of dialectal arabic speech,” in IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)....

2019
[11]

Large scale weakly and semi- supervised learning for low-resource video asr,

Kritika Singh, Vimal Manohar, Alex Xiao, Sergey Edunov, Ross Gir- shick, Vitaliy Liptchinsky, Christian Fuegen, Yatharth Saraf, Geoffrey Zweig, and Abdelrahman Mohamed, “Large scale weakly and semi- supervised learning for low-resource video asr,” pp. 3770–3774, 2020

2020
[12]

Weakly supervised construction of asr systems with massive video data,

Mengli Cheng, Chengyu Wang, Xu Hu, Jun Huang, and Xiaobo Wang, “Weakly supervised construction of asr systems with massive video data,” arXiv preprint arXiv:2008.01300 , 2020

2008 arXiv
[13]

Prompting the hidden talent of web-scale speech models for zero-shot task generalization,

Puyuan Peng, Brian Yan, Shinji Watanabe, and David Harwath, “Prompting the hidden talent of web-scale speech models for zero-shot task generalization,” arXiv preprint arXiv:2305.11095 , 2023

2023 arXiv
[14]

Prompting whisper for qa-driven zero-shot end-to-end spoken language understanding,

Mohan Li, Simon Keizer, and Rama Doddipatla, “Prompting whisper for qa-driven zero-shot end-to-end spoken language understanding,” in Proc. Interspeech 2024 , 2024, pp. 1330–1334

2024
[15]

Zero-shot domain-sensitive speech recognition with prompt-conditioning fine-tuning,

Feng-Ting Liao, Yung-Chieh Chan, Yi-Chang Chen, Chan-Jan Hsu, and Da-shan Shiu, “Zero-shot domain-sensitive speech recognition with prompt-conditioning fine-tuning,” in 2023 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU). IEEE, 2023, pp. 1–8

2023
[16]

Extending whisper with prompt tuning to target-speaker asr,

Hao Ma, Zhiyuan Peng, Mingjie Shao, Jing Li, and Ju Liu, “Extending whisper with prompt tuning to target-speaker asr,” in ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024, pp. 12516–12520

2024
[17]

Whisper-based transfer learning for alzheimer disease classification: Leveraging speech segments with full transcripts as prompts,

Jinpeng Li and Wei-Qiang Zhang, “Whisper-based transfer learning for alzheimer disease classification: Leveraging speech segments with full transcripts as prompts,” in ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . IEEE, 20...

2024
[18]

Careless whisper: Speech-to-text hallucination harms,

Allison Koenecke, Anna Seo Gyeong Choi, Katelyn X Mei, Hilke Schellmann, and Mona Sloane, “Careless whisper: Speech-to-text hallucination harms,” in Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency , 2024, pp. 1672–1681

2024
[19]

Self-training: A survey,

Massih-Reza Amini, Vasilii Feofanov, Loic Pauletto, Lies Hadjadj, Emilie Devijver, and Yury Maximov, “Self-training: A survey,” Neuro- computing, vol. 616, pp. 128904, 2025

2025
[20]

A formula for the gini coefficient,

Robert Dorfman, “A formula for the gini coefficient,” The review of economics and statistics , pp. 146–149, 1979

1979

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [1]

Summary of chatgpt-related research and perspective towards the future of large language models,

Yiheng Liu, Tianle Han, Siyuan Ma, Jiayue Zhang, Yuanyuan Yang, Jiaming Tian, Hao He, Antong Li, Mengshen He, Zhengliang Liu, et al., “Summary of chatgpt-related research and perspective towards the future of large language models,” Meta-Radiology, p. 100017, 2023

work page 2023

[2] [2]

Foun- dation models for generalist medical artificial intelligence,

Michael Moor, Oishi Banerjee, Zahra Shakeri Hossein Abad, Harlan M Krumholz, Jure Leskovec, Eric J Topol, and Pranav Rajpurkar, “Foun- dation models for generalist medical artificial intelligence,” Nature, vol. 616, no. 7956, pp. 259–265, 2023

work page 2023

[3] [3]

Will affective computing emerge from foundation models and general artificial intelli- gence? a first evaluation of chatgpt,

Mostafa M. Amin, Erik Cambria, and Bj ¨orn W. Schuller, “Will affective computing emerge from foundation models and general artificial intelli- gence? a first evaluation of chatgpt,” IEEE Intelligent Systems , vol. 38, no. 2, pp. 15–23, 2023

work page 2023

[4] [4]

A comprehensive survey of few-shot learning: Evolution, applications, challenges, and opportunities,

Yisheng Song, Ting Wang, Puyu Cai, Subrota K Mondal, and Jy- oti Prakash Sahoo, “A comprehensive survey of few-shot learning: Evolution, applications, challenges, and opportunities,” ACM Computing Surveys, vol. 55, no. 13s, pp. 1–40, 2023

work page 2023

[5] [5]

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. PMLR, 2023, pp. 28492–28518

work page 2023

[6] [6]

Analysis of whisper automatic speech recognition performance on low resource language,

Riefkyanov Surya Adia Pratama and Agit Amrullah, “Analysis of whisper automatic speech recognition performance on low resource language,” Jurnal Pilar Nusa Mandiri , vol. 20, no. 1, pp. 1–8, 2024

work page 2024

[7] [7]

Comparison of multilingual self-supervised and weakly-supervised speech pre-training for adaptation to unseen languages,

Andrew Rouditchenko, Sameer Khurana, Samuel Thomas, Rogerio Feris, Leonid Karlinsky, Hilde Kuehne, David Harwath, Brian Kings- bury, and James Glass, “Comparison of multilingual self-supervised and weakly-supervised speech pre-training for adaptation to unseen languages,” in Proc. Interspeech 2023 , 2023, pp. 2268–2272

work page 2023

[8] [8]

Benchmarking whisper for low-resource speech recognition: An n- shot evaluation on pashto, punjabi, and urdu,

Najm Ul Sehar, Ayesha Khalid, Farah Adeeba, and Sarmad Hussain, “Benchmarking whisper for low-resource speech recognition: An n- shot evaluation on pashto, punjabi, and urdu,” in Proceedings of the First Workshop on Challenges in Processing South Asian Languages (CHiPSAL 2025) , 2025, pp. 202–207

work page 2025