A Cross-Corpus Speech Emotion Recognition Method Based on Supervised Contrastive Learning

Xiang minjie

REVIEW 4 major objections 5 minor 19 references

A Cross-Corpus Speech Emotion Recognition Method Based on Supervised Contrastive Learning

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read This paper claims that a two-stage supervised contrastive fine-tuning method, which treats same-emotion samples across English and Chinese corpora as positive pairs, reaches 77.41% unweighted accuracy on IEMOCAP and 96.49% on CASIA…

desk verdict A plausible cross-corpus SER recipe whose main empirical claim is underdetermined by the reported ablations. read the letter →

arxiv 2411.19803 v1 pith:RROMW7OB submitted 2024-11-25 cs.SD cs.CLeess.AS

Xiang minjie This is my paper · ORCID

classification cs.SDcs.CLeess.AS

keywords speechemotionrecognitioncontrastivelearningself-supervisedrepresentationscross-corpustransfertwo-stagefine-tuningHuBERTWavLMInfoNCE

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 tries to show that the generalization problem in speech emotion recognition can be reduced by training the feature encoder with supervised contrastive learning across two corpora before training the classifier on the target corpus. It reports that WavLM fine-tuned this way reaches 77.41% unweighted accuracy on IEMOCAP and 96.49% on CASIA, while HuBERT reaches 76.92% and 97.19%, both above direct fine-tuning and published baselines. The implied payoff is a transfer recipe: self-supervised speech models can absorb emotion information from several datasets at once, then adapt cheaply to a new one.

What carries the argument

The machinery is a two-stage fine-tuning procedure. In stage one, the speech encoder and feature projection are frozen and only the transformer feature encoder is optimized with a total loss that sums InfoNCE contrastive loss and cosine margin loss: samples of the same emotion drawn from both corpora are positive pairs, and different-emotion samples are negative pairs. In stage two, the whole representation model is frozen and only the self-attention pooling layer and a two-layer classifier are trained with cross-entropy loss on the target corpus. The self-attention pooling layer, taken from a speaker-recognition design, aggregates time-step features into one utterance vector before classification.

What would settle it

Fine-tune the same two-stage setup on two corpora whose emotion labels are deliberately mismatched, for example by swapping 'happy' and 'sad' in one corpus before defining pairs, and check whether the IEMOCAP and CASIA accuracies stay high; if they do, the reported gains are not caused by cross-corpus emotion alignment.

Watch

Extended reading notes

Core claim

The central claim is that pulling together same-emotion samples from different languages and corpora, while pushing apart different-emotion samples, makes the learned speech features more emotion-discriminative. Positive pairs are defined by shared emotion labels across IEMOCAP and CASIA, and negatives by differing emotion labels. The paper reports this first fine-tuning stage raises UA by 10.31 points for HuBERT and 9.1 points for WavLM over direct fine-tuning on IEMOCAP, and that the final models exceed the compared methods on both datasets. It also presents feature-similarity evidence that after fine-tuning, positive pairs become more similar than negative pairs in the deeper transformer layers.

Load-bearing premise

The load-bearing premise is that the four shared emotion labels in the English IEMOCAP and Chinese CASIA corpora are semantically aligned, so a positive pair really is two samples of the same emotion.

Editorial extensions

If this is right

The two-stage recipe can be applied to any self-supervised speech representation model; both HuBERT and WavLM gain over their direct fine-tuning versions.
Cross-lingual emotion transfer works with only the four emotion categories shared by IEMOCAP and CASIA; adding more corpora with overlapping labels may push accuracy further.
The first stage alone changes the feature geometry: after roughly seven transformer layers, positive-pair similarity exceeds negative-pair similarity, showing the contrastive objective reorganizes deep features.
Training on multiple datasets simultaneously yields strong results on both target datasets, so multi-corpus contrastive fine-tuning does not sacrifice one language for the other.

Reading between the lines

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

The gain likely depends on the assumption that 'neutral', 'sad', 'angry', and 'happy' are culturally equivalent across English and Chinese; if annotation standards differ, the contrastive loss would reinforce mismatched labels, so a label-similarity audit across corpora would make the claim more solid.
A testable extension is to rotate or shuffle emotion labels during the first stage; if accuracy stays high under shuffled labels, the mechanism is not label alignment but some other corpus structure.
The reported parameter cost could be attacked with knowledge distillation from the fine-tuned WavLM or HuBERT into a smaller student, a direction the paper itself names.
The same pairing logic might transfer to multimodal emotion recognition by treating same-emotion audio-text pairs as positives, though the paper only gestures at this.

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: the paper describes a sensible two-stage fine-tuning recipe for cross-corpus SER—train a WavLM/Hubert feature encoder with supervised contrastive loss on IEMOCAP+CASIA, then fine-tune a classifier on the target corpus. The reported numbers (77.41 UA on IEMOCAP, 96.49 on CASIA) are decent, and the internal comparison against direct fine-tuning shows a large gap. But the paper doesn't actually show that the gap comes from contrastive learning. The FT baseline trains one-stage on the target dataset only; the proposed method trains first-stage on both datasets and then a second-stage classifier. Those three differences (extra data, two-stage schedule, contrastive objective) are never disentangled. Without an ablation that replaces the contrastive loss with plain cross-entropy on the same combined data, the central claim is overclaimed. This is the paper's main soft spot, and it's load-bearing.

What's genuinely useful: the cross-corpus positive-pair sampling idea (same emotion label across languages as positives) is a natural and potentially effective transfer heuristic, and the feature similarity analysis (positive pairs becoming more similar than negatives after fine-tuning) is a nice sanity check. The 5-fold CV protocol is appropriate, and reporting both Hubert and WavLM results is good practice.

The smaller issues: the equations are garbled (section numbering is off; equations 1–5 are hard to read), there are no error bars or significance tests, hyperparameters are chosen ad hoc (tau=0.07, alpha=0.5, m=0.4) with no sensitivity analysis, and the SOTA claim rests on only a handful of baselines from other papers, some using different protocols. The assumption that the four shared emotion labels are cross-lingually aligned is plausible but untested; if the annotation standards differ, the contrastive loss could reinforce label mismatches.

Overall: the paper is a reasonable empirical note, not a transformative contribution. A serious referee could make the ablation gap explicit and send the authors back to add a proper control. With that control, the result could be a solid workshop-type paper. As is, I would not cite it as evidence that contrastive learning helps.

Recommendation: send to peer review, but expect heavy revision. The idea is testable, the topic is relevant, and the current evidence is suggestive rather than conclusive.

Referee Report

4 major / 5 minor

Summary. The paper proposes a two-stage fine-tuning method for cross-corpus speech emotion recognition (SER). In the first stage, the feature encoder of a self-supervised speech representation model (HuBERT or WavLM) is fine-tuned with a supervised contrastive loss using jointly sampled batches from the English IEMOCAP and Chinese CASIA corpora; positive pairs are samples sharing the same emotion label across the two datasets, and negative pairs are samples with different labels. In the second stage, the self-attention pooling layer and a small classifier are fine-tuned on the target corpus with cross-entropy loss. The authors report unweighted accuracy (UA) of 76.92% (HuBERT) and 77.41% (WavLM) on IEMOCAP, and 97.19% (HuBERT) and 96.49% (WavLM) on CASIA, and claim these outperform state-of-the-art results. The paper also compares against direct fine-tuning (FT) baselines on the same two models.

Significance. If the central claim is established, the proposed recipe would be a simple and potentially effective way to leverage multiple SER corpora for transfer learning, a genuinely useful contribution to cross-corpus emotion recognition. The paper uses standard self-supervised models and standard loss functions (InfoNCE plus a cosine-margin term), and the idea of cross-corpus positive pairs defined by shared emotion labels is intuitive. However, the empirical evidence as presented does not yet isolate the effect of the supervised contrastive objective. The comparison against the FT baseline conflates several factors, external baselines are not protocol-matched, and no variance information is reported. With the required ablation and protocol clarifications, the paper could be a solid, if modest, empirical contribution; in its current form the central claim is underdetermined.

major comments (4)

[Section 4.2, Table 1] The comparison that supports the central claim contrasts the proposed two-stage supervised contrastive fine-tuning (rows 8 and 9) against a single-stage direct fine-tuning baseline (rows 1 and 2). This comparison conflates at least four factors: (i) exposure to additional cross-corpus training data, (ii) the two-stage fine-tuning schedule, (iii) the contrastive loss itself, and (iv) different optimization settings (the proposed first stage uses an initial learning rate of 1e-4, while the FT baseline uses 1e-3). No ablation holds factors (i)-(iii) fixed while removing only the contrastive objective. Please add a control experiment that uses the same union of training data and the same two-stage protocol, but replaces the contrastive loss in the first stage with a standard supervised cross-entropy loss on the source-dataset labels (or a multi-task CE). Without such a control, the reported gains of 10.31% and 9.1% cannot be attributed to supervised contrastive learning.
[Tables 1 and 2, Section 4.2] The external baselines (rows 3-7 of Table 1 and all rows of Table 2) are taken from other papers, but the manuscript does not state whether those results were obtained under the same protocol: the same four-class partition of IEMOCAP, the same five-fold cross-validation splits, the same evaluation metric (UA), and the same input preprocessing. For example, the claimed superiority over the nearest external baselines is only about 0.5-1.8 percentage points, which can easily be explained by differences in folds or class definitions. Please provide a detailed protocol table for each baseline or re-implement the baselines under the authors' evaluation setup to make the state-of-the-art claim defensible.
[Section 3.1] The positive-pair construction treats the four emotion labels (neutral, sad, angry, happy) as semantically interchangeable across IEMOCAP and CASIA. This is a load-bearing design assumption: if the two corpora use different annotation guidelines, annotator agreement levels, or cultural display rules, the supervised contrastive loss will align mismatched labels and could reinforce cross-corpus discrepancies rather than learning emotion-invariant features. The paper provides no empirical evidence of label equivalence, such as label distribution statistics, annotator agreement, or a manual inspection of positive pairs. Please either validate this assumption or discuss it explicitly as a limitation, and ideally include an ablation in which positive pairs are drawn only within each dataset to gauge the effect of cross-corpus pairing.
[Section 4.2] The paper reports only the mean UA over the five cross-validation folds. No standard deviations, per-fold results, or significance tests are given. Given that the claimed advantage over some external baselines is about 1-2 percentage points (e.g., 77.41 vs. 76.78 in Table 1), the absence of variance information makes it difficult to judge whether the differences are statistically reliable. Please report per-fold results and standard deviations, and, where appropriate, a paired significance test across folds.

minor comments (5)

[General] The manuscript has broken section and equation numbering: sections 2.2, 2.3, 3.3, 3.4, and 3.5 appear after Section 4, and equations are numbered inconsistently (e.g., Eq. (1) appears in Section 3.1 but Sections 2.2 and 2.3 contain equations (2)-(5)). Please renumber all sections and equations for readability.
[Equations (1)-(5)] Equations (2)-(5) are garbled in the current typesetting; for example, Eq. (2) reads 'exp sim , /' with unclear symbols, and Eq. (4) is unreadable. Please typeset them cleanly and define every symbol, including the meaning of k, i, j, and the positive/negative sample superscripts. Eq. (1) also needs a clearer definition of the self-attention pooling operation.
[Section 4.2] The hyperparameters tau=0.07, alpha=0.5, m=0.4, learning rates, and epoch counts are stated without justification or sensitivity analysis. Please add a brief explanation of how they were chosen (e.g., from prior work, grid search, or intuition) and report at least a small sensitivity study for the three main loss hyperparameters.
[Figure 4] Figure 4 is difficult to interpret: the axes are not clearly labeled, the curves are not captioned with a legend, and the text does not specify how 'positive' and 'negative' pairs are defined for this similarity analysis. Please revise the figure for clarity and explain the sampling used to compute the plotted similarities.
[Tables 1 and 2] The relationship between Table 1 and Table 2 is unclear: Table 2 lists 'Hubert-base' and 'WavLM-base' without the asterisk used in Table 1, and the text does not explicitly state whether these rows correspond to the proposed two-stage method or to a different fine-tuning procedure. Please clarify which configuration each row reports.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the reported UA values are external test-set measurements, not quantities fixed by construction.

full rationale

The paper's central claims are empirical accuracy values (UA 77.41% on IEMOCAP and 96.49% on CASIA) obtained by evaluating a fixed fine-tuning protocol under 5-fold cross-validation. These numbers are not fitted parameters, and no equation in the paper defines the reported accuracy in terms of the training loss or the datasets used for fitting; the loss functions (InfoNCE and cosine margin) are standard and their hyperparameters are set manually. The comparison against direct fine-tuning is under-controlled because the FT rows do not share the two-stage protocol or the cross-corpus training data, so the specific contribution of the contrastive loss is not isolated, but that is an experimental-design weakness rather than circular reasoning. Likewise, the similarity analysis in Figure 4 merely confirms that the contrastive training objective moved positive-pair and negative-pair cosine similarities in the intended direction; the paper does not present that diagnostic as an independent prediction. There are no load-bearing self-citations and no imported uniqueness arguments. The derivation chain is therefore self-contained as an empirical study.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The method relies on standard contrastive loss functions and the assumption that emotion labels are comparable across the two corpora. No new entities, forces, or dimensions are introduced. The main free parameters are the manually chosen hyperparameters, which are not fitted to the test set.

free parameters (5)

tau (temperature in InfoNCE loss) = 0.07
Chosen hyperparameter for the contrastive loss in Eq. 2; controls the sharpness of the softmax distribution and the scale of gradients.
alpha (cosine margin loss weight) = 0.5
Weight of the cosine margin loss in Eq. 4; balances the two loss terms.
m (cosine margin) = 0.4
Margin for negative pairs in Eq. 4; controls how far negative samples must be separated.
Learning rates and schedules = 1e-4 (stage1), 5e-4 (stage2), 1e-3 (FT baseline); decay multipliers
Selected manually; central to the reported results but no hyperparameter search is documented.
Epoch counts = 50 (stage1), 10 (stage2), 50 (FT baseline)
Training duration in each stage, chosen without a stated validation criterion.

assumptions (4)

standard math The InfoNCE contrastive loss correctly implements supervised contrastive learning
Eqs. 2-3 use a standard formulation for pulling positive pairs together and pushing negative pairs apart.
standard math The cosine margin loss combines with InfoNCE to improve separation
Eq. 4 follows a common margin-based objective for metric learning.
domain assumption Emotion labels in IEMOCAP and CASIA are semantically aligned for the four shared categories
Section 3.1 defines positive pairs as samples with the same emotion label across languages; if the labels are not comparable, the contrastive loss may learn the wrong alignment.
domain assumption Pretrained HuBERT/WavLM representations capture transferable speech features for emotion
Section 4.2 loads base versions from HuggingFace and assumes these self-supervised features are a good starting point for the emotion task.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Cross-Corpus Speech Emotion Recognition Method Based on Supervised Contrastive Learning." pith.science (2026). https://pith.science/paper/RROMW7OB

@misc{pith2026241119803,
  author       = {Pith},
  title        = {Pith review of: A Cross-Corpus Speech Emotion Recognition Method Based on Supervised Contrastive Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RROMW7OB}},
  note         = {Machine review of arXiv:2411.19803}
}

read the original abstract

Research on Speech Emotion Recognition (SER) often faces challenges such as the lack of large-scale public datasets and limited generalization capability when dealing with data from different distributions. To solve this problem, this paper proposes a cross-corpus speech emotion recognition method based on supervised contrast learning. The method employs a two-stage fine-tuning process: first, the self-supervised speech representation model is fine-tuned using supervised contrastive learning on multiple speech emotion datasets; then, the classifier is fine-tuned on the target dataset. The experimental results show that the WavLM-based model achieved unweighted accuracy (UA) of 77.41% on the IEMOCAP dataset and 96.49% on the CASIA dataset, outperforming the state-of-the-art results on the two datasets.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

19 extracted references · 18 canonical work pages

[1]

Emotion recognition combining acoustic and linguistic features based on speech recognition results[C]

Sakurai M, Kosaka T. Emotion recognition combining acoustic and linguistic features based on speech recognition results[C]. 2021 IEEE 10th Global Conference on Consumer Electronics (GCCE). IEEE, 2021: 824-827

work page 2021
[2]

Deep implicit distribution alignment networks for cross-corpus speech emotion recognition[C]

Zhao Y , Wang J, Zong Y , et al. Deep implicit distribution alignment networks for cross-corpus speech emotion recognition[C]. ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2023: 1-5

work page 2023
[3]

Exploring wav2vec 2.0 fine tuning for improved speech emotion recognition[C]

Chen L W, Rudnicky A. Exploring wav2vec 2.0 fine tuning for improved speech emotion recognition[C]. ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2023: 1-5

work page 2023
[4]

A Comprehensive Exploration of Fine-Tuning WavLM for Enhancing Speech Emotion Recognition[C]

Ali F, Arymurthy A M, Prasojo R E. A Comprehensive Exploration of Fine-Tuning WavLM for Enhancing Speech Emotion Recognition[C]. 2023 6th International Seminar on Research of Information Technology and Intelligent Systems (ISRITI). IEEE, 2023: 295-300

work page 2023
[5]

Revealing Emotional Clusters in Speaker Embeddings: A Contrastive Learning Strategy for Speech Emotion Recognition[C]

Ulgen I R, Du Z, Busso C, et al. Revealing Emotional Clusters in Speaker Embeddings: A Contrastive Learning Strategy for Speech Emotion Recognition[C]. ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2024: 12081-12085

work page 2024
[6]

IEMOCAP: Interactive emotional dyadic motion capture database[J]

Busso C, Bulut M, Lee C C, et al. IEMOCAP: Interactive emotional dyadic motion capture database[J]. Language resources and evaluation, 2008, 42: 335-359

work page 2008
[7]

The CASIA audio emotion recognition method for audio/visual emotion challenge 2011[C]

Pan S, Tao J, Li Y . The CASIA audio emotion recognition method for audio/visual emotion challenge 2011[C]. Affective Computing and Intelligent Interaction: Fourth International Conference, ACII 2011, Memphis, TN, USA, October 9–12, 2011, Proceedings, Part II. Springer Berlin Heidelberg, 2011: 388-395

work page 2011
[8]

Hubert: Self-supervised speech representation learning by masked prediction of hidden units[J]

Hsu W N, Bolte B, Tsai Y H H, et al. Hubert: Self-supervised speech representation learning by masked prediction of hidden units[J]. IEEE/ACM transactions on audio, speech, and language processing, 2021, 29: 3451-3460

work page 2021

Show all 19 references

[9]

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

Chen S, Wang C, Chen Z, et al. Wavlm: Large-scale self-supervised pre-training for full stack speech processing[J]. IEEE Journal of Selected Topics in Signal Processing, 2022, 16(6): 1505-1518

2022
[10]

wav2vec 2.0: A framework for self-supervised learning of speech representations[J]

Baevski A, Zhou Y , Mohamed A, et al. wav2vec 2.0: A framework for self-supervised learning of speech representations[J]. Advances in neural information processing systems, 2020, 33: 12449-12460

2020
[11]

Speech emotion recognition using self-supervised features[C]

Morais E, Hoory R, Zhu W, et al. Speech emotion recognition using self-supervised features[C]. ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2022: 6922-6926

2022
[12]

Ecapa-tdnn: Emphasized channel attention, propagation and aggregation in tdnn based speaker verification[J]

Desplanques B, Thienpondt J, Demuynck K. Ecapa-tdnn: Emphasized channel attention, propagation and aggregation in tdnn based speaker verification[J]. arXiv preprint arXiv:2005.07143, 2020

2005 arXiv
[13]

Speech-based emotion recognition with self-supervised models using attentive channel-wise correlations and label smoothing[C]

Kakouros S, Stafylakis T, Mošner L, et al. Speech-based emotion recognition with self-supervised models using attentive channel-wise correlations and label smoothing[C]. ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2023: 1-5

2023
[14]

Contrastive unsupervised learning for speech emotion recognition[C]

Li M, Yang B, Levy J, et al. Contrastive unsupervised learning for speech emotion recognition[C]. ICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2021: 6329-6333

2021
[15]

MCM-CSD: Multi-Granularity Context Modeling with Contrastive Speaker Detection for Emotion Recognition in Real-Time Conversation[C]

Xu Y , Yang M. MCM-CSD: Multi-Granularity Context Modeling with Contrastive Speaker Detection for Emotion Recognition in Real-Time Conversation[C]. ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2024: 11956-11960

2024
[16]

Self-attention encoding and pooling for speaker recognition[J]

Safari P, India M, Hernando J. Self-attention encoding and pooling for speaker recognition[J]. arXiv preprint arXiv:2008.01077, 2020

2008 arXiv
[17]

Dual-tbnet: Improving the robustness of speech features via dual-transformer-bilstm for speech emotion recognition[J]

Liu Z, Kang X, Ren F. Dual-tbnet: Improving the robustness of speech features via dual-transformer-bilstm for speech emotion recognition[J]. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2023, 31: 2193-2203

2023
[18]

Temporal modeling matters: A novel temporal emotional modeling approach for speech emotion recognition[C]

Ye J, Wen X C, Wei Y , et al. Temporal modeling matters: A novel temporal emotional modeling approach for speech emotion recognition[C]. ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2023: 1-5

2023
[19]

emodarts: Joint optimisation of cnn & sequential neural network architectures for superior speech emotion recognition[J]

Rajapakshe T, Rana R, Khalifa S, et al. emodarts: Joint optimisation of cnn & sequential neural network architectures for superior speech emotion recognition[J]. IEEE Access, 2024

2024

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Emotion recognition combining acoustic and linguistic features based on speech recognition results[C]

Sakurai M, Kosaka T. Emotion recognition combining acoustic and linguistic features based on speech recognition results[C]. 2021 IEEE 10th Global Conference on Consumer Electronics (GCCE). IEEE, 2021: 824-827

work page 2021

[2] [2]

Deep implicit distribution alignment networks for cross-corpus speech emotion recognition[C]

Zhao Y , Wang J, Zong Y , et al. Deep implicit distribution alignment networks for cross-corpus speech emotion recognition[C]. ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2023: 1-5

work page 2023

[3] [3]

Exploring wav2vec 2.0 fine tuning for improved speech emotion recognition[C]

Chen L W, Rudnicky A. Exploring wav2vec 2.0 fine tuning for improved speech emotion recognition[C]. ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2023: 1-5

work page 2023

[4] [4]

A Comprehensive Exploration of Fine-Tuning WavLM for Enhancing Speech Emotion Recognition[C]

Ali F, Arymurthy A M, Prasojo R E. A Comprehensive Exploration of Fine-Tuning WavLM for Enhancing Speech Emotion Recognition[C]. 2023 6th International Seminar on Research of Information Technology and Intelligent Systems (ISRITI). IEEE, 2023: 295-300

work page 2023

[5] [5]

Revealing Emotional Clusters in Speaker Embeddings: A Contrastive Learning Strategy for Speech Emotion Recognition[C]

Ulgen I R, Du Z, Busso C, et al. Revealing Emotional Clusters in Speaker Embeddings: A Contrastive Learning Strategy for Speech Emotion Recognition[C]. ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2024: 12081-12085

work page 2024

[6] [6]

IEMOCAP: Interactive emotional dyadic motion capture database[J]

Busso C, Bulut M, Lee C C, et al. IEMOCAP: Interactive emotional dyadic motion capture database[J]. Language resources and evaluation, 2008, 42: 335-359

work page 2008

[7] [7]

The CASIA audio emotion recognition method for audio/visual emotion challenge 2011[C]

Pan S, Tao J, Li Y . The CASIA audio emotion recognition method for audio/visual emotion challenge 2011[C]. Affective Computing and Intelligent Interaction: Fourth International Conference, ACII 2011, Memphis, TN, USA, October 9–12, 2011, Proceedings, Part II. Springer Berlin Heidelberg, 2011: 388-395

work page 2011

[8] [8]

Hubert: Self-supervised speech representation learning by masked prediction of hidden units[J]

Hsu W N, Bolte B, Tsai Y H H, et al. Hubert: Self-supervised speech representation learning by masked prediction of hidden units[J]. IEEE/ACM transactions on audio, speech, and language processing, 2021, 29: 3451-3460

work page 2021