Pith. sign in

REVIEW 4 major objections 9 minor 43 references

Cued-Agent: A Collaborative Multi-Agent System for Automatic Cued Speech Recognition

T0 review · 4 major / 9 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Cued-Agent is the first multi-agent system for Automatic Cued Speech Recognition, using four specialized agents to fuse hand and lip information without training a fusion network.

desk verdict Plausible multi-agent ACSR system and a genuinely useful hearing-impaired dataset, but the headline accuracy claim is contradicted by the paper's own Table 2. read the letter →

arxiv 2508.00391 v1 pith:EBHVHNN6 submitted 2025-08-01 cs.CV eess.AS

Guanjie Huang , Danny H.K. Tsang , Shan Yang , Guangzhi Lei , Li Liu This is my paper · ORCID
classification cs.CVeess.AS
keywords CuedSpeechrecognitionmulti-agentsystemhand-lipfusionmultimodallargelanguagemodelphoneme-to-wordconversionhearing-impaireddatasetlipreadingparameter-free
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 proposes Cued-Agent, a multi-agent system that converts Mandarin Cued Speech videos—hand gestures plus lip movements—into natural-language sentences. Its central claim is that the long-standing problem of asynchrony between hand and lip signals can be handled without training a fusion network: a multimodal large language model reads hand positions and shapes from selected keyframes, and those one-hot hand prompts are added directly into the CTC decoding scores of a lip-reading model. A separate LLM agent then revises the predicted phoneme sequence into a sentence using Cued Speech rules and semantic coherence. The authors also collect a new dataset from eight hearing-impaired cuers and report that Cued-Agent matches the previous state of the art on phoneme error rates while being the only method that produces sentence-level output. If correct, this lowers the cost of building ACSR systems for new speakers and languages.

What carries the argument

The load-bearing object is the hand prompt matrix $H \in \mathbb{R}^{T \times q}$, a one-hot encoding of the hand position and shape categories that GPT-4o assigns to each keyframe, propagated over the frames in each slow-motion group. During decoding, the agent forms the fused feature $L_H = L' + \lambda_{\text{prompt}} \cdot H$ by adding $H$ to the lip model's projected CTC feature $L'$, then computes the CTC prefix score from $L_H$ and combines it with the attention score in beam search. This is what makes hand-lip fusion parameter-free: the hand signal enters as a bias on the CTC logits rather than through trained cross-modal layers. The second mechanism is the Self-Correction P2W agent, which repeatedly rewrites the phoneme sequence under prompts encoding Mandarin CS conversion rules, in-context examples, and confusion-pair contrasts, finally emitting a sentence.

What would settle it

Run the Hand Recognition Agent alone on a held-out set of hearing-impaired cuers with known hand-code labels and compute position and shape classification accuracy; if accuracy on confusable codes is near chance, set $\lambda_{\text{prompt}} = 0$ and compare. A failure to show a significant CER or WER drop when the hand prompt is removed would indicate the hand pathway is not carrying the claimed information.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a Cued Speech recognizer does not need a trained multimodal fusion module. Cued-Agent decomposes recognition into four cooperating agents: a GPT-4o-based Hand Recognition agent that screens slow-motion keyframes and classifies hand position and shape; a pretrained Transformer Lip Recognition agent finetuned only on lip frames; a Hand Prompt Decoding agent that forms $L_H = L' + \lambda_{\text{prompt}} \cdot H$, adding a one-hot hand-prompt matrix to the projected lip features during beam search; and a DeepSeek-R1-based Self-Correction Phoneme-to-Word agent that turns phonemes into sentences. In experiments the system reaches 2.61% CER / 6.56% WER on the single normal-hearing cuer subset and 12.67% / 29.86% on the new eight-hearing-impaired-cuer set, comparable to the previous state of the art, and it reports sentence-level S-WER from 12.1% to 40.84% across settings, the first such results. The paper's key novelty claims are the parameter-free fusion via prompt-weighted decoding and the self-correcting phoneme-to-word conversion.

Load-bearing premise

The whole pipeline rides on GPT-4o correctly recognizing hand positions and shapes from keyframes of unseen cuers, because those one-hot hand prompts are injected directly into the decoder's scores and no trained fusion layer exists to absorb a wrong hand classification.

Editorial extensions

If this is right

  • Adding a new cuer or language no longer requires collecting hand labels to train a fusion module; only the lip model needs finetuning.
  • Cued Speech output can be presented as readable sentences, with S-WER and Semantic Score as quantitative measures, making ACSR usable in assistive communication tools.
  • The method's hand information can be swapped: any multimodal LLM that can classify the CS hand code can replace GPT-4o without changing the fusion mechanism.
  • Hearing-impaired cuers, whose lip movements are harder to read, benefit most from the hand prompt pathway, since the ablation shows hand information gives the largest gains on hearing-impaired subsets.
  • The same four-agent recipe can be applied to other cued languages whose hand codes differ, as long as prompts and support sets are rebuilt.

Reading between the lines

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

  • A testable extension is to measure GPT-4o's hand position and shape accuracy directly; if it is high, the parameter-free fusion can be seen as injecting near-oracle hand information, and if low, Cued-Agent's gains would be expected to shrink on unseen cuers.
  • The confusion matrices show errors concentrate on phonemes with identical hand codes (for example, b/p and yu/w); one implication is that the lip model, not hand fusion, must resolve those, so pairing the hand prompt with a language-model prior over Mandarin syllables could close much of the remaining error.
  • The choices $\lambda_{\text{prompt}} = 4.5$ and $\lambda_{\text{decode}} = 0.5$ are reported without a validation sweep; a sensitivity analysis over these weights would reveal how much of the result depends on hand-prompt strength rather than on the hand signal itself.
  • Because sentence-level metrics are new, future work could compare human lip-reader performance on the same videos to calibrate how much of the semantic correction is doing the work.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 9 minor

Summary. This paper proposes Cued-Agent, a multi-agent system for Mandarin Automatic Cued Speech Recognition (ACSR). Four agents are combined: an MLLM-based hand recognition agent that converts keyframe hand classifications into a one-hot prompt matrix, a Transformer-based lip recognition agent finetuned with CTC/attention loss, a hand-prompt decoding agent that fuses hand prompts into lip logits by weighted addition during beam search, and an LLM-based self-correction phoneme-to-word agent. The authors also collect a new multi-hearing-impaired Mandarin CS dataset (MHI-MCCSD) with eight hearing-impaired cuers and propose two sentence-level metrics, S-WER and Semantic Score. The abstract and conclusion claim that the system 'performs superbly' and 'outperforms' state-of-the-art methods, while Section 5.2 more cautiously describes the result as comparable to STF-ACSR.

Significance. The paper has several strengths: the multi-agent decomposition of ACSR is novel, the parameter-free hand-lip fusion idea is simple and relevant to data-scarce settings, the new MHI-MCCSD dataset addresses a real gap in hearing-impaired evaluation, and the sentence-level output is a useful step beyond phoneme-only recognition. However, the headline performance claim is not supported by the paper's own Table 2, where Cued-Agent has higher CER than STF-ACSR on all four test settings and higher WER on three of four settings. The fusion weights are introduced without a described validation protocol, and Eq. (5) misstates the CTC prefix probability. If the performance claims are revised and the technical points are corrected, the dataset and the multi-agent pipeline could still make a moderate contribution, but the current text overstates what the experiments show.

major comments (4)
  1. [Abstract; Table 2; Section 6] The claim that Cued-Agent 'performs superbly' and 'outperforms SOTA methods' is directly contradicted by Table 2. For all four test settings, Cued-Agent has higher CER than STF-ACSR: 2.61 vs 1.82 on MCCSD (1-H), 6.72 vs 4.62 on MCCSD (1-HI), 9.05 vs 8.35 on MCCSD (6-H), and 12.67 vs 10.96 on MHI-MCCSD (8-HI). The WER is also higher in three of the four settings, with only MCCSD (6-H) WER being better (20.54 vs 21.06). The abstract and conclusion therefore overstate the results, and Section 5.2's word 'comparable' is not equivalent to 'superb.' The performance claims must be retracted or substantially weakened, or new experiments must support the claimed superiority.
  2. [Section 5.1; Eqs. (6), (8)] The reported results depend on the free hyperparameters lambda_prompt=4.5 and lambda_decode=0.5, as well as on the keyframe thresholds sigma=6 and theta=2, but the paper does not describe a validation split or a tuning procedure. If these values were selected after inspecting test-set results, the CER/WER numbers in Tables 2-4 are fitted values rather than independent predictions. The authors should specify how each hyperparameter is chosen, report results on a held-out validation split, and include a sensitivity analysis over lambda_prompt and lambda_decode.
  3. [Section 3.5, Eq. (5)] Eq. (5) is not a correct statement of the CTC prefix probability. The prefix probability is the cumulative probability over all label sequences that have hypothesis h as a prefix, including arbitrarily long continuations with blank and non-blank transitions; it is not the sum over a single next label nu in Q union {<eos>} of p_ctc(h·nu|L). Because Eqs. (7)-(8) use this quantity in the beam-search scoring, the derivation as written is technically unsound and should be corrected to the standard recursive prefix-probability definition.
  4. [Section 3.3; Table 4] The hand recognition agent is load-bearing because its one-hot matrix H is directly added to the CTC logits with weight 4.5 in Eq. (6). The paper never reports the accuracy of the GPT-4o hand position and shape classification on the keyframe set, so the error injected into the decoder is unknown. The ablation in Table 4 shows only an end-to-end effect; an intermediate evaluation of hand recognition accuracy, or at least a confusion matrix for position and shape categories, is needed to justify the design and to quantify error propagation.
minor comments (9)
  1. [Section 5.2 and Table 2 caption] The text should be aligned with the numbers: Table 2's caption says 'good performance,' the main text says 'comparable,' and the abstract says 'superb'; these are inconsistent and should be harmonized.
  2. [Section 3.5] The phrase 'we innovatively add' is editorializing; also, while the fusion is parameter-free in the sense of no learned weights, lambda_prompt is a hyperparameter that must be tuned, so the 'parameter-free' claim should be qualified.
  3. [Sections 3.3 and 3.6] The exact prompt templates for the hand recognition and self-correction agents are not provided; the descriptions in the text are not sufficient for reproduction. The prompts should be included in a supplementary file.
  4. [Section 4] The paper claims that MHI-MCCSD is 'publicly available,' but only a GitHub URL for the implementation is given; the dataset release mechanism should be stated explicitly.
  5. [Section 3.4, Eq. (3)] The value of lambda_train in the joint CTC/attention loss is not reported; this is a tunable parameter that affects the lip recognizer and should be specified.
  6. [Section 3.7] The Semantic Score uses Sentence-BERT [31], but the authors should specify the exact model checkpoint and confirm that it provides appropriate Mandarin embeddings, since the original Sentence-BERT models are primarily English-oriented.
  7. [Figure 7] The confusion matrices would benefit from a colorbar and a description of the color scale; the current figure is difficult to interpret quantitatively.
  8. [References] References [11] and [12] are identical in title, authors, and venue; this duplication should be removed.
  9. [Section 6] The phrase 'training-free multimodal alignment' is misleading because the lip recognition agent is finetuned; a more precise phrasing would be 'training-free hand-lip fusion.'

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the hand-lip fusion and phoneme-to-word correction are not tautological.

full rationale

I reviewed the derivation chain from hand/lip ROI preprocessing through Eq. (6)–(8) and the Self-Correction P2W agent. The hand prompt matrix H is built from MLLM keyframe classification results, and the decoding feature L_H = L' + lambda_prompt*H is a weighted sum of a learned lip logit map and the hand one-hot prompt; the final phoneme sequence is obtained by beam search over this combined score. This is a fusion of two independent signal sources, not a definitional restatement of either input. The Self-Correction P2W agent is prompted with training-set in-context examples and CS conversion rules, and it is not described as receiving test labels, so the reported CER/WER improvements are not forced by construction. The claims of 'first', 'training-free', and 'parameter-free' are architectural novelty claims, not circular reductions. The main self-citation [15] supplies the MLLM prompt design; although [15] shares overlapping authors, it is used here as a system component rather than as an external theorem invoked to forbid alternatives, and no load-bearing result is derived solely from that citation. No equation in the paper reduces to its own input; lambda_prompt and lambda_decode are stated as fixed constants (4.5 and 0.5) without a described validation split, which is an experimental-reporting and generalization concern, not a circularity. I therefore assign a circularity score of 0. Separately, and not as circularity, the abstract's 'superb' claim is not supported by Table 2, where Cued-Agent has higher CER than STF-ACSR in all four settings and higher WER in three of four settings; that is a correctness/consistency issue outside the scope of this pass.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The system relies on four free hyperparameters and several domain assumptions. The two most consequential are lambda_prompt and lambda_decode, which control the hand-lip fusion and the CTC/attention balance; these are set without a validation split and directly influence all reported accuracy numbers. The hand recognition is assumed to work zero-shot via GPT-4o, and the LLM self-correction is assumed to improve accuracy without hallucination. No new physical or artificial entities are introduced beyond the hand-prompt matrix representation.

free parameters (4)
  • lambda_prompt = 4.5
    Weight for adding the hand-prompt matrix to CTC logits. Set without a validation split; may be tuned on the test set.
  • lambda_decode = 0.5
    Interpolation weight between CTC and attention scores in beam search. Set without a validation split.
  • sigma = 6
    Slow-motion speed threshold for keyframe filtering in the Hand Recognition agent.
  • theta = 2
    Frame index distance threshold for grouping slow-motion frames.
assumptions (4)
  • domain assumption GPT-4o with the designed prompts can accurately classify Mandarin CS hand positions and shapes from keyframes.
    The hand recognition agent relies entirely on zero-shot MLLM classification; the paper reports no hand recognition accuracy and inherits the approach from [15].
  • ad hoc to paper Adding a one-hot hand-prompt matrix H to the CTC logits (with a constant weight lambda_prompt) is a valid fusion of hand and lip information.
    This linear logit bias is introduced as the core fusion mechanism without theoretical justification or comparison to alternative fusion methods.
  • domain assumption The LLM (DeepSeek-R1) corrects phoneme sequences using language priors without degrading accuracy.
    The self-correction agent assumes the LLM's edits improve the recognition; no hallucination analysis or human evaluation is provided.
  • domain assumption The MHI-MCCSD dataset is correctly annotated and representative of hearing-impaired cuers.
    The dataset is new, but no annotation quality metrics, per-cuer accuracy, or release availability are provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Cued-Agent: A Collaborative Multi-Agent System for Automatic Cued Speech Recognition." pith.science (2026). https://pith.science/paper/EBHVHNN6

@misc{pith2026250800391,
  author       = {Pith},
  title        = {Pith review of: Cued-Agent: A Collaborative Multi-Agent System for Automatic Cued Speech Recognition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EBHVHNN6}},
  note         = {Machine review of arXiv:2508.00391}
}
read the original abstract

Cued Speech (CS) is a visual communication system that combines lip-reading with hand coding to facilitate communication for individuals with hearing impairments. Automatic CS Recognition (ACSR) aims to convert CS hand gestures and lip movements into text via AI-driven methods. Traditionally, the temporal asynchrony between hand and lip movements requires the design of complex modules to facilitate effective multimodal fusion. However, constrained by limited data availability, current methods demonstrate insufficient capacity for adequately training these fusion mechanisms, resulting in suboptimal performance. Recently, multi-agent systems have shown promising capabilities in handling complex tasks with limited data availability. To this end, we propose the first collaborative multi-agent system for ACSR, named Cued-Agent. It integrates four specialized sub-agents: a Multimodal Large Language Model-based Hand Recognition agent that employs keyframe screening and CS expert prompt strategies to decode hand movements, a pretrained Transformer-based Lip Recognition agent that extracts lip features from the input video, a Hand Prompt Decoding agent that dynamically integrates hand prompts with lip features during inference in a training-free manner, and a Self-Correction Phoneme-to-Word agent that enables post-process and end-to-end conversion from phoneme sequences to natural language sentences for the first time through semantic refinement. To support this study, we expand the existing Mandarin CS dataset by collecting data from eight hearing-impaired cuers, establishing a mixed dataset of fourteen subjects. Extensive experiments demonstrate that our Cued-Agent performs superbly in both normal and hearing-impaired scenarios compared with state-of-the-art methods. The implementation is available at https://github.com/DennisHgj/Cued-Agent.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 41 canonical work pages

  1. [1]

    Orin Cornett

    R. Orin Cornett. 1967. Cued Speech. American Annals of the Deaf 112, 1 (1967), 3–13

    work page 1967

  2. [2]

    Orin Cornett

    R. Orin Cornett. 1994. Adapting Cued Speech to additional languages. Cued Speech Journal 5 (1994), 19–29

    work page 1994

  3. [3]

    Yue Fan, Xiaojian Ma, Rujie Wu, Yuntao Du, Jiaqi Li, Zhi Gao, and Qing Li. 2024. Videoagent: A memory-augmented multimodal agent for video understanding. In ECCV

    work page 2024

  4. [4]

    Lufei Gao, Shan Huang, and Li Liu. 2023. A novel interpretable and generalizable re-synchronization model for Cued Speech based on a multi-cuer corpus. In Interspeech

    work page 2023

  5. [5]

    Alex Graves. 2012. Supervised sequence labelling. InSupervised sequence labelling with recurrent neural networks . Springer, 5–13

    work page 2012

  6. [6]

    Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, et al . 2025. DeepSeek-R1: Incentivizing reasoning capability in LLMs via reinforcement learning. arXiv (2025)

    work page 2025

  7. [7]

    Taicheng Guo, Xiuying Chen, Yaqi Wang, Ruidi Chang, Shichao Pei, Nitesh V Chawla, Olaf Wiest, and Xiangliang Zhang. 2024. Large language model based multi-agents: A survey of progress and challenges. arXiv (2024)

    work page 2024

  8. [8]

    Shanshan Han, Qifan Zhang, Yuhang Yao, Weizhao Jin, Zhaozhuo Xu, and Chaoyang He. 2024. LLM multi-agent systems: Challenges and open problems. arXiv (2024)

    work page 2024

Show all 43 references
  1. [9]

    Junda He, Christoph Treude, and David Lo. 2024. LLM-Based multi-agent systems for software engineering: literature review, vision and the road ahead. ACM Transactions on Software Engineering and Methodology (2024)

    2024

  2. [10]

    Panikos Heracleous, Noureddine Aboutabit, and Denis Beautemps. 2009. Lip shape and hand position fusion for automatic vowel recognition in Cued Speech for French. IEEE Signal Processing Letters 16, 5 (2009), 339–342

    2009

  3. [11]

    Panikos Heracleous, Denis Beautemps, and Noureddine Aboutabit. 2010. Cued Speech automatic recognition in normal-hearing and deaf subjects. Speech Communication 52, 6 (2010), 504–512

    2010

  4. [12]

    Panikos Heracleous, Denis Beautemps, and Noureddine Aboutabit. 2010. Cued speech automatic recognition in normal-hearing and deaf subjects. Speech Com- munication 52, 6 (2010), 504–512

    2010

  5. [13]

    Panikos Heracleous, Denis Beautemps, and Norihiro Hagita. 2012. Continuous phoneme recognition in Cued Speech for French. In EUSIPCO

    2012

  6. [14]

    Wen-Lian Hsu and Yi-Shiou Chen. 1999. On phoneme-to-character conversion systems in Chinese processing. Journal of the Chinese Institute of Engineers 22, 5 (1999), 573–579

    1999

  7. [15]

    Guanjie Huang, Danny Hin Kwok Tsang, and Li Liu. 2025. Lend a hand: Semi training-free Cued Speech recognition via MLLM-driven hand modeling for barrier-free communication. arXiv (2025)

    2025

  8. [16]

    Rongjie Huang, Mingze Li, Dongchao Yang, Jiatong Shi, Xuankai Chang, Zhenhui Ye, Yuning Wu, Zhiqing Hong, Jiawei Huang, Jinglin Liu, et al. 2024. AudioGPT: Understanding and generating speech, music, sound, and talking head. In AAAI

    2024

  9. [17]

    Mingcheng Li, Xiaolu Hou, Ziyang Liu, Dingkang Yang, Ziyun Qian, Jiawei Chen, Jinjie Wei, Yue Jiang, Qingyao Xu, and Lihua Zhang. 2025. MCCD: Multi- agent collaboration-based compositional diffusion for complex text-to-image generation. In CVPR

    2025

  10. [18]

    Xinyi Li, Sai Wang, Siqi Zeng, Yu Wu, and Yi Yang. 2024. A survey on LLM-based multi-agent systems: workflow, infrastructure, and challenges. Vicinagearth 1, 1 (2024), 9

    2024

  11. [19]

    Yunxin Li, Haoyuan Shi, Baotian Hu, Longyue Wang, Jiashun Zhu, Jinyi Xu, Zhen Zhao, and Min Zhang. 2024. Anim-director: A large multimodal model powered agent for controllable animation video generation. In SIGGRAPH Asia

    2024

  12. [20]

    Jinhua Liang, Huan Zhang, Haohe Liu, Yin Cao, Qiuqiang Kong, Xubo Liu, Wenwu Wang, Mark D Plumbley, Huy Phan, and Emmanouil Benetos. 2024. WavCraft: Audio editing and generation with natural language prompts. InICLR Workshop

    2024

  13. [21]

    Daniel Ling and Bryan R Clarke. 1975. Cued speech: An evaluative study. Ameri- can Annals of the deaf (1975), 480–488

    1975

  14. [22]

    Li Liu and Gang Feng. 2019. A pilot study on Mandarin Chinese Cued Speech. American Annals of the Deaf 164, 4 (2019), 496–518

    2019

  15. [23]

    Li Liu, Gang Feng, and Denis Beautemps. 2018. Automatic temporal segmentation of hand movements for hand positions recognition in French Cued Speech. In ICASSP

    2018

  16. [24]

    Li Liu, Thomas Hueber, Gang Feng, and Denis Beautemps. 2018. Visual recog- nition of continuous Cued Speech using a tandem CNN-HMM approach. In Interspeech

    2018

  17. [25]

    Lei Liu and Li Liu. 2023. Cross-modal mutual learning for Cued Speech recogni- tion. In ICASSP

    2023

  18. [26]

    Lei Liu, Li Liu, and Haizhou Li. 2024. Computation and parameter efficient multi- modal fusion transformer for Cued Speech recognition. IEEE/ACM Transactions on Audio, Speech, and Language Processing 32 (2024), 1559–1572

    2024

  19. [27]

    Camillo Lugaresi, Jiuqiang Tang, Hadon Nash, Chris McClanahan, Esha Uboweja, Michael Hays, Fan Zhang, Chuo-Ling Chang, Ming Guang Yong, Juhyun Lee, Wan-Teh Chang, Wei Hua, Manfred Georg, and Matthias Grundmann. 2019. MediaPipe: A framework for building perception pipelines. ar...

    2019

  20. [28]

    Pingchuan Ma, Alexandros Haliassos, Adriana Fernandez-Lopez, Honglie Chen, Stavros Petridis, and Maja Pantic. 2023. Auto-avsr: Audio-visual speech recogni- tion with automatic labels. In ICASSP

    2023

  21. [29]

    OpenAI. 2024. Hello GPT-4o. https://openai.com/index/hello-gpt-4o/

    2024

  22. [30]

    Katerina Papadimitriou and Gerasimos Potamianos. 2021. A fully convolutional sequence learning approach for Cued Speech recognition from videos. In EU- SIPCO

    2021

  23. [31]

    Nils Reimers and Iryna Gurevych. 2019. Sentence-BERT: Sentence embeddings using siamese BERT-networks. In EMNLP

    2019

  24. [32]

    Sanjana Sankar, Denis Beautemps, and Thomas Hueber. 2022. Multistream neural architectures for Cued Speech recognition using a pre-trained visual feature extractor and constrained CTC decoding. In ICASSP

    2022

  25. [33]

    Sefik Ilkin Serengil and Alper Ozpinar. 2020. LightFace: A hybrid deep face recognition framework. In ASYU

    2020

  26. [34]

    Beverly J Trezek. 2017. Cued Speech and the development of reading in English: Examining the evidence. The Journal of Deaf Studies and Deaf Education 22, 4 (2017), 349–364

    2017

  27. [35]

    Rong-Cheng Tu, Wenhao Sun, Zhao Jin, Jingyi Liao, Jiaxing Huang, and Dacheng Tao. 2024. SPAgent: adaptive task decomposition and model selection for general video generation and editing. arXiv (2024)

    2024

  28. [36]

    Wiebe Van der Hoek and Michael Wooldridge. 2008. Multi-agent systems. Foun- dations of artificial intelligence 3 (2008), 887–928

    2008

  29. [37]

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. Advances in NeurIPS 30 (2017)

    2017

  30. [38]

    Jianrong Wang, Nan Gu, Mei Yu, Xuewei Li, Qiang Fang, and Li Liu. 2021. An attention self-supervised contrastive learning based three-stage model for hand shape feature representation in Cued Speech. arXiv (2021)

    2021

  31. [39]

    Jianrong Wang, Ge Zhang, Zhenyu Wu, Xuewei Li, and Li Liu. 2021. Self- supervised depth estimation via implicit cues from videos. In ICASSP

    2021

  32. [40]

    Zhenyu Wang, Aoxue Li, Zhenguo Li, and Xihui Liu. 2024. Genartist: Multimodal llm as an agent for unified image generation and editing. Advances in NeurIPS 37 (2024), 128374–128395

    2024

  33. [41]

    Shinji Watanabe, Takaaki Hori, Suyoun Kim, John R Hershey, and Tomoki Hayashi. 2017. Hybrid CTC/attention architecture for end-to-end speech recogni- tion. IEEE Journal of Selected Topics in Signal Processing 11, 8 (2017), 1240–1253

    2017

  34. [42]

    Yehang Zhang, Xinli Xu, Xiaojie Xu, Li Liu, and Yingcong Chen. 2025. Long-video audio synthesis with multi-agent collaboration. arXiv (2025)

    2025

  35. [43]

    Tianchen Zhou, Zhongjie Duan, Cen Chen, Wenmeng Zhou, Yanhao Wang, and Yaliang Li. 2025. AgentStory: A multi-agent system for story visualization with multi-subject consistent text-to-image generation. In ICMR

    2025

Pith tools

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