Pith. sign in

REVIEW 4 major objections 4 minor 47 references

Unleashing the Potential of Model Bias for Generalized Category Discovery

T0 review · 4 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A frozen biased model can be reused, not just fought, to discover novel categories.

desk verdict Reasonable incremental GCD work with a real mechanism gap: the entropy gate suppresses the confident errors the transfer step is meant to exploit. read the letter →

arxiv 2412.12501 v1 pith:ZDOM4ELU submitted 2024-12-17 cs.LG cs.CLcs.CV

Wenbin An , Haonan Lin , Jiahao Nie , Feng Tian , Wenkai Shi , Yaqiang Wu , Qianying Wang , Ping Chen This is my paper · ORCID
classification cs.LGcs.CLcs.CV
keywords GeneralizedCategoryDiscoverylogitadjustmentmodelbiaspseudo-labelingnovelentropyweightingtextintentdetectionSinkhorn-Knopp
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

Generalized Category Discovery asks a model to recognize both known and previously unseen categories from unlabeled data, but models pre-trained only on known categories are biased toward them. This paper argues that this bias is not merely an obstacle: a frozen biased model's predictions encode exactly how much the model over-favors each known class, and they also encode similarity hints that can separate novel classes from each other. The proposed Self-Debiasing Calibration (SDC) subtracts the biased logits from known-category logits and adds transferred logits to novel-category logits, with a per-sample entropy weight $\alpha$ that decides how strongly to intervene. On three text benchmarks, the framework lifts novel-category accuracy by about 2.6 points on average over prior methods while keeping known-category accuracy roughly intact. If the paper is right, the standard practice of treating model bias as pure noise is leaving useful signal on the table.

What carries the argument

The central mechanism is a pair of logit-adjustment identities built from the frozen biased model. Category Bias Mitigation computes $L_{\mathrm{CBM}} = L_\theta[:M] - \alpha \cdot L_{\mathrm{bias}}$, subtracting the biased logits from the known-category slice of the current logits; Category Confusion Mitigation computes $L_{\mathrm{CCM}} = L_\theta[M:] + \alpha \cdot T^\top \times L_{\mathrm{bias}}$, where $T$ is a prototype-similarity transfer matrix from known to novel categories. The two outputs are concatenated into $L_c$, and the entropy-weighted $\alpha = \beta \cdot \mathrm{Sigmoid}(E - E_{\max})$ decides per-sample how much debiasing and transfer to apply. These identities turn the frozen model's predictions into a structured prior rather than a nuisance, and they are what the ablation study shows the framework depends on.

What would settle it

Construct a probe set where a novel class is semantically near a known class, so the biased model confidently predicts the known class (low entropy); if SDC's gains on novel accuracy come from the entropy gate, such a probe should show smaller or negative gains compared to a dataset with well-separated novel classes.

Watch

Extended reading notes

Core claim

The paper's central claim is that the output logits of a frozen model biased toward known categories can be repurposed as a calibration signal for the current training model. Two logit-level operations carry the argument: for known categories, subtracting the biased logits (scaled by $\alpha$) from the current model's logits removes the over-confidence the biased prior injects; for novel categories, adding the transpose of a prototype-similarity transfer matrix times the biased logits pushes the current logits along the directions the biased model already associates with each novel cluster. The per-sample weight $\alpha = \beta \cdot \mathrm{Sigmoid}(E - E_{\max})$ comes from the entropy of the biased softmax, which is low for known samples and high for novel samples. The calibrated logits feed the Sinkhorn-Knopp pseudo-labeling scheme, and the resulting pseudo-labels improve especially on novel categories. The paper reports that this yields an average improvement of 2.64% in novel-category accuracy over state-of-the-art baselines on BANKING, HWU64, and CLINC, while known-category accuracy improves by 0.28% on average.

Load-bearing premise

The argument depends on entropy of the biased model's softmax cleanly separating known from novel samples, so a single scalar alpha per sample can steer both the debiasing and the transfer; if a novel sample is confidently mispredicted as a known class, or a known sample has high entropy, the correction goes the wrong way.

Editorial extensions

If this is right

  • Novel-category accuracy on text GCD benchmarks rises by about 2.6 points on average, and known-category accuracy is not sacrificed, so the debiasing and transfer steps are complementary.
  • The classifier-based design allows online inference that is about 5.6 times faster than clustering-based inference, which matters for streaming intent data.
  • Estimating the category count $K$ with the dropout algorithm gives results close to using the true $K$, so the framework's benefit does not hinge on knowing $K$ in advance.
  • Ablations on HWU64 show that removing CBM, CCM, or the entropy weighting each drops the H-score, so all three components are load-bearing.

Reading between the lines

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

  • The logit-adjustment recipe should transfer to vision GCD if the frozen model's class-similarity structure is replaced by a vision-language model's text-conditional logits; this is a direct testable extension.
  • The method's boundary case is a novel sample that the biased model confidently places in a known class - such samples get little $\alpha$ and may not be recovered, so gains should shrink on datasets with high known-novel similarity.
  • One could close the loop by periodically fine-tuning the 'biased' model on the current pseudo-labels, letting the reference distribution co-evolve with the discovered novel classes rather than staying fixed.
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.

Referee Report

4 major / 4 minor

Summary. The paper proposes Self-Debiasing Calibration (SDC), a method for Generalized Category Discovery (GCD) on text. The core idea is to use a frozen pre-trained model biased toward known categories to adjust the logits of a trainable model: category bias mitigation subtracts the biased logits from known-category logits (Eq. 3), and category confusion mitigation adds a transfer-matrix-weighted version of the biased logits to novel-category logits (Eq. 5). An entropy-based weight (Eq. 8) modulates these adjustments per sample, and pseudo-labels are produced from the calibrated logits via Sinkhorn-Knopp optimal transport. The model is trained with a supervised loss on labeled data, a pseudo-label loss on unlabeled data, and a contrastive loss. Experiments on BANKING, HWU64, and CLINC report state-of-the-art results, with particular gains in novel-category accuracy.

Significance. If the reported findings hold, the paper offers a conceptually interesting reframing: instead of merely suppressing model bias toward known categories, the biased model's outputs can be reused as a source of knowledge for novel categories. The method is clearly specified and the open-source code is a reproducible artifact. However, the empirical evidence is narrow (three text intent datasets only), and the absence of standard deviations, significance tests, and hyperparameter sensitivity analysis substantially tempers the strength of the claims. The central mechanism also has a potential internal tension between the motivating example and the entropy gate (Eq. 8), which the authors would need to resolve for the conceptual claim to be fully convincing.

major comments (4)
  1. [Eq. (8) and Category Confusion Mitigation] The entropy-based weighting in Eq. (8) undermines the stated mechanism of Category Confusion Mitigation (CCM). The motivating example in the introduction is a novel 'cat' confidently misclassified as known 'dog'; such a sample has low entropy on the biased model, so alpha_i approaches zero and the transfer term alpha_i * T^T L_bias in Eq. (5) is suppressed. The batch-relative sigmoid further weakens the transfer for any sample whose entropy is below the batch maximum. The ablation in Table 2 shows that removing CCM lowers H-score by 1.4 points on HWU64, but it does not tell us whether that gain comes from confident misclassified novel samples or only from already-uncertain novel samples. Unless the authors provide evidence that CCM helps specifically for low-entropy novel samples (e.g., by binning pseudo-label accuracy by entropy of the biased model), the claim that bias transfer resolves Novel-Known errors is not established. This is an internal tension between the stated motivation and the implemented gate, not a tuning issue.
  2. [Table 1 and Experimental Setup] All numbers in Table 1 are averages over 3 runs without standard deviations, confidence intervals, or significance tests. The headline improvements (2.64% average on novel accuracy, 2.00% on H-score) are modest relative to typical run-to-run variability in deep clustering methods on these benchmarks. Without a measure of dispersion, the reader cannot judge whether the reported superiority over KTN, TAN, and DPN is systematic or within noise. I request per-run results or standard deviations for the main table and the ablation table, and, where feasible, a paired test over runs.
  3. [Implementation Details] The key hyperparameter beta is set to 0.03, 0.05, and 0.42 for BANKING, HWU64, and CLINC, respectively, which is a large variation and suggests strong per-dataset tuning. No sensitivity analysis is provided, so it is unclear whether the reported gains persist across reasonable choices of beta. Since beta scales the entire logit adjustment in Eqs. (3), (5), and (8), the authors should provide a plot or table of H-score and known/novel accuracy versus beta on at least one dataset, and ideally a principled selection rule. As it stands, the method may be over-fit to these three benchmarks.
  4. [Datasets (Section: Experiments)] The evaluation is limited to three text intent detection datasets (BANKING, HWU64, CLINC). Generalized Category Discovery is commonly evaluated on image benchmarks such as CIFAR-100, ImageNet-100, and the Semantic Shift Benchmark; by restricting the experiments to text, the paper does not demonstrate that the proposed mechanism generalizes beyond this narrow setting. Given that the abstract frames the method in terms of general 'category bias' without domain-specific assumptions, this is a significant gap. At minimum, the authors should either add one image-domain GCD experiment (even a representative one) or explicitly state and justify the text-only scope.
minor comments (4)
  1. [Model Training (Eq. 10)] The notation uses the hat symbol both for pseudo-labels and for the model's predictions (\hat{Y}^l and \hat{Y}^u), which is confusing. Please rename the pseudo-label variable (e.g., \tilde{Y}) or use distinct notation.
  2. [Evaluation Protocol] The paper states that testing is inductive, but Table 1 reports clustering accuracy, which typically relies on transductive clustering over the test set. The Real-world Applications section later says 'We report experimental results using clustering on the testing sets offline'. Please reconcile the terminology and clarify whether the main results come from a transductive clustering step or from the learned classifier.
  3. [K Estimation] The manuscript says that K estimation uses the dropout algorithm (Zhang et al., 2021a) but provides no description or algorithmic details. Since Table 2 reports results 'w/o Ground-truth K', the estimation method should be specified enough to be reproducible.
  4. [Comparison with GPT models] The GPT-3.5 and GPT-4 baselines are stated to be based on the turbo API, but the prompt template is not given. Including the exact prompt would make the zero-shot comparison reproducible.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: SDC's claims are empirically evaluated on external benchmarks, not derived from its own definitions.

full rationale

The paper's central claim—that logit adjustment using the frozen biased model's outputs improves novel-category GCD—is an empirical claim tested on held-out test splits of BANKING, HWU64, and CLINC. Equations (1)-(9) define a constructive procedure: Lbias is the frozen biased model's output, LCBM and LCCM are adjusted logits, and pseudo-labels are produced by Sinkhorn-Knopp. None of these equations defines the measured accuracy or H-score in terms of itself; the benchmark numbers in Table 1 are external. The entropy-based weighting in Eq. (8) is borrowed from Shi et al. (2024), a prior paper with overlapping authors, and is therefore a self-citation; however, it is not load-bearing as a circular proof because the paper corroborates the entropy separation empirically in Figure 3 and ablates the weighting in Table 2, while the overall comparison remains against external test sets. The self-training loop (calibrated logits generate pseudo-labels that train fθ and gϕ) is standard self-training and not a definitional equivalence. No uniqueness theorem or fitted parameter is renamed as a prediction. The internal tension noted by a skeptical reader—that Eq. (8) suppresses the low-entropy novel samples that motivate CCM—is a potential flaw in mechanism design, not a circularity. Therefore the paper is self-contained against external benchmarks and receives score 0.

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

The method is heuristic and data-driven: its behavior is controlled by per-dataset hyperparameters and by assumptions about the biased model's entropy and prototype geometry. No new physical or conceptual entities are introduced; the 'biased model' is a frozen copy of the pre-trained model and the transfer matrix T is a mathematical construct, not an additional entity.

free parameters (4)
  • beta (entropy weighting factor) = 0.03 (BANKING), 0.05 (HWU64), 0.42 (CLINC)
    Scales alpha in Eq. 8 and therefore controls the strength of both CBM and CCM. Set to a different value per dataset with no sensitivity analysis; the main results depend on these chosen values.
  • lambda1 (supervised loss balance) = 0.6 increasing to 0.7
    Balances labeled vs unlabeled cross-entropy in Eq. 10; fixed by hand across datasets.
  • lambda2 (contrastive loss weight) = 0.01
    Weights the instance-level contrastive loss in Eq. 11; fixed by hand.
  • epsilon (Sinkhorn entropy regularization) = not reported, adopted from Caron et al. 2020
    Regularization strength in Eq. 9; taken from prior work, not tuned in this paper, but still a hyperparameter of the pseudo-label generation.
assumptions (5)
  • domain assumption Entropy of the biased model's softmax outputs separates known from novel samples.
    Used in Eq. 8 to set alpha per sample; Figure 3 only illustrates the separation qualitatively and no overlap statistic is provided.
  • domain assumption The frozen biased model's logits L_bias accurately measure category bias toward known classes.
    Central to CBM in Eq. 3; the paper asserts this without direct validation of the bias measure.
  • domain assumption Cosine similarity between known and novel prototypes is a valid basis for transferring knowledge.
    Used to build transfer matrix T in Eq. 4 and to boost novel logits in Eq. 5; assumes the geometry of initial prototypes is meaningful for semantic transfer.
  • domain assumption Uniform distribution of pseudo-labels across all K categories is a desirable constraint.
    Enforced by the Sinkhorn-Knopp formulation in Eq. 9; this may not hold for imbalanced real-world data, and the paper does not test imbalanced settings.
  • domain assumption The total number of categories K is known for the main experiments.
    The classifier is built with K output units and the Sinkhorn balancing uses K; a separate section estimates K, but the headline results use ground-truth K.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Unleashing the Potential of Model Bias for Generalized Category Discovery." pith.science (2026). https://pith.science/paper/ZDOM4ELU

@misc{pith2026241212501,
  author       = {Pith},
  title        = {Pith review of: Unleashing the Potential of Model Bias for Generalized Category Discovery},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZDOM4ELU}},
  note         = {Machine review of arXiv:2412.12501}
}
read the original abstract

Generalized Category Discovery is a significant and complex task that aims to identify both known and undefined novel categories from a set of unlabeled data, leveraging another labeled dataset containing only known categories. The primary challenges stem from model bias induced by pre-training on only known categories and the lack of precise supervision for novel ones, leading to category bias towards known categories and category confusion among different novel categories, which hinders models' ability to identify novel categories effectively. To address these challenges, we propose a novel framework named Self-Debiasing Calibration (SDC). Unlike prior methods that regard model bias towards known categories as an obstacle to novel category identification, SDC provides a novel insight into unleashing the potential of the bias to facilitate novel category learning. Specifically, the output of the biased model serves two key purposes. First, it provides an accurate modeling of category bias, which can be utilized to measure the degree of bias and debias the output of the current training model. Second, it offers valuable insights for distinguishing different novel categories by transferring knowledge between similar categories. Based on these insights, SDC dynamically adjusts the output logits of the current training model using the output of the biased model. This approach produces less biased logits to effectively address the issue of category bias towards known categories, and generates more accurate pseudo labels for unlabeled data, thereby mitigating category confusion for novel categories. Experiments on three benchmark datasets show that SDC outperforms SOTA methods, especially in the identification of novel categories. Our code and data are available at \url{https://github.com/Lackel/SDC}.

Figures

Figures reproduced from arXiv: 2412.12501 by the authors.

Figure 1
Figure 1. Four types of GCD errors. For instance, Novel￾Novel Error indicates samples from novel categories are mis￾classified into wrong novel categories. An et al. 2024c) and image categorization (Vaze, Vedaldi, and Zisserman 2024; Rastegar, Doughty, and Snoek 2024), where well-trained models may encounter unlabeled data from pre￾viously unseen novel categories. To address the requirements of open-world scenarios, Generaliz… view at source ↗
Figure 2
Figure 2. The overall framework of our model, which mainly contains Bias Mitigation, Confusion Mitigation, and Entropy-based [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Entropy distribution for known and Novel cate [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: Visualization of the learned features and model [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 27 canonical work pages

  1. [1]

    L.; Almeida, D.; Altenschmidt, J.; Altman, S.; Anadkat, S.; et al

    Achiam, J.; Adler, S.; Agarwal, S.; Ahmad, L.; Akkaya, I.; Aleman, F. L.; Almeida, D.; Altenschmidt, J.; Altman, S.; Anadkat, S.; et al. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774

    arXiv 2023

  2. [2]

    An, W.; Shi, W.; Tian, F.; Lin, H.; Wang, Q.; Wu, Y.; Cai, M.; Wang, L.; Chen, Y.; Zhu, H.; and Chen, P. 2024 a . Generalized Category Discovery with Large Language Models in the Loop. In Findings of the Association for Computational Linguistics: ACL 2024, 8653--8665. Bangkok, Thailand: Association for Computational Linguistics

    work page 2024

  3. [3]

    An, W.; Tian, F.; Chen, P.; Tang, S.; Zheng, Q.; and Wang, Q. 2022. Fine-grained Category Discovery under Coarse-grained supervision with Hierarchical Weighted Self-contrastive Learning. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, 1314--1323

    work page 2022

  4. [4]

    An, W.; Tian, F.; Chen, P.; Zheng, Q.; and Ding, W. 2023 a . New User Intent Discovery with Robust Pseudo Label Training and Source Domain Joint-training. IEEE Intelligent Systems

    work page 2023

  5. [5]

    An, W.; Tian, F.; Shi, W.; Chen, Y.; Wu, Y.; Wang, Q.; and Chen, P. 2024 b . Transfer and alignment network for generalized category discovery. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 10856--10864

    work page 2024

  6. [6]

    An, W.; Tian, F.; Shi, W.; Chen, Y.; Zheng, Q.; Wang, Q.; and Chen, P. 2023 b . DNA: Denoised Neighborhood Aggregation for Fine-grained Category Discovery. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, 12292--12302

    work page 2023

  7. [7]

    An, W.; Tian, F.; Shi, W.; Lin, H.; Wu, Y.; Cai, M.; Wang, L.; Wen, H.; Yao, L.; and Chen, P. 2024 c . Down: Dynamic order weighted network for fine-grained category discovery. Knowledge-Based Systems, 293: 111666

    work page 2024

  8. [8]

    An, W.; Tian, F.; Zheng, Q.; Ding, W.; Wang, Q.; and Chen, P. 2023 c . Generalized category discovery with decoupled prototypical network. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, 12527--12535

    work page 2023

Show all 47 references
  1. [9]

    M.; Rupprecht, C.; and Vedaldi, A

    Asano, Y. M.; Rupprecht, C.; and Vedaldi, A. 2019. Self-labelling via simultaneous clustering and representation learning. arXiv preprint arXiv:1911.05371

    2019 arXiv

  2. [10]

    D.; Dhariwal, P.; Neelakantan, A.; Shyam, P.; Sastry, G.; Askell, A.; et al

    Brown, T.; Mann, B.; Ryder, N.; Subbiah, M.; Kaplan, J. D.; Dhariwal, P.; Neelakantan, A.; Shyam, P.; Sastry, G.; Askell, A.; et al. 2020. Language models are few-shot learners. Advances in neural information processing systems, 33: 1877--1901

    2020

  3. [11]

    Caron, M.; Bojanowski, P.; Joulin, A.; and Douze, M. 2018. Deep clustering for unsupervised learning of visual features. In Proceedings of the European Conference on Computer Vision (ECCV), 132--149

    2018

  4. [12]

    Caron, M.; Misra, I.; Mairal, J.; Goyal, P.; Bojanowski, P.; and Joulin, A. 2020. Unsupervised learning of visual features by contrasting cluster assignments. Advances in neural information processing systems, 33: 9912--9924

    2020

  5. [13]

    Casanueva, I.; Tem c inas, T.; Gerz, D.; Henderson, M.; and Vuli \'c , I. 2020. Efficient intent detection with dual sentence encoders. arXiv preprint arXiv:2003.04807

    2020 arXiv

  6. [14]

    Chen, D.; Shen, Y.; Zhang, H.; and Torr, P. H. 2022. Zero-shot logit adjustment. arXiv preprint arXiv:2204.11822

    2022 arXiv

  7. [15]

    Cui, Y.; Jia, M.; Lin, T.-Y.; Song, Y.; and Belongie, S. 2019. Class-balanced loss based on effective number of samples. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 9268--9277

    2019

  8. [16]

    Cuturi, M. 2013. Sinkhorn distances: Lightspeed computation of optimal transport. Advances in neural information processing systems, 26

    2013

  9. [17]

    Devlin, J.; Chang, M.-W.; Lee, K.; and Toutanova, K. 2018. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805

    2018 arXiv

  10. [18]

    Gidaris, S.; and Komodakis, N. 2018. Dynamic few-shot visual learning without forgetting. In Proceedings of the IEEE conference on computer vision and pattern recognition, 4367--4375

    2018

  11. [19]

    Han, K.; Vedaldi, A.; and Zisserman, A. 2019. Learning to discover novel visual categories via deep transfer clustering. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 8401--8409

    2019

  12. [20]

    J.; Clarke, C.; Lee, A.; Hill, P.; Kummerfeld, J

    Larson, S.; Mahendran, A.; Peper, J. J.; Clarke, C.; Lee, A.; Hill, P.; Kummerfeld, J. K.; Leach, K.; Laurenzano, M. A.; Tang, L.; et al. 2019. An evaluation dataset for intent classification and out-of-scope prediction. arXiv preprint arXiv:1909.02027

    2019 arXiv

  13. [21]

    Lin, H.; An, W.; Chen, Y.; Tian, F.; Yao, Y.; Ding, W.; Wang, Q.; and Chen, P. 2024 a . A Tri-Branch Network with Prototype-aware Matching for Universal Category Discovery. In 2024 IEEE International Conference on Multimedia and Expo (ICME), 1--6. IEEE Computer Society

    2024

  14. [22]

    Lin, H.; An, W.; Wang, J.; Chen, Y.; Tian, F.; Wang, M.; Wang, Q.; Dai, G.; and Wang, J. 2024 b . Flipped Classroom: Aligning Teacher Attention with Student in Generalized Category Discovery. In The Thirty-eighth Annual Conference on Neural Information Processing Systems

    2024

  15. [23]

    Lin, T.-E.; Xu, H.; and Zhang, H. 2020. Discovering new intents via constrained deep adaptive clustering with cluster refinement. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, 8360--8367

    2020

  16. [24]

    Liu, X.; Eshghi, A.; Swietojanski, P.; and Rieser, V. 2021. Benchmarking natural language understanding services for building conversational agents. In Increasing Naturalness and Flexibility in Spoken Dialogue Interaction. Springer

    2021

  17. [25]

    MacQueen, J.; et al. 1967. Some methods for classification and analysis of multivariate observations. In Proceedings of the fifth Berkeley symposium on mathematical statistics and probability, volume 1, 281--297. Oakland, CA, USA

    1967

  18. [26]

    K.; Jayasumana, S.; Rawat, A

    Menon, A. K.; Jayasumana, S.; Rawat, A. S.; Jain, H.; Veit, A.; and Kumar, S. 2020. Long-tail learning via logit adjustment. arXiv preprint arXiv:2007.07314

    2020 arXiv

  19. [27]

    Pennington, J.; Socher, R.; and Manning, C. D. 2014. Glove: Global vectors for word representation. In Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), 1532--1543

    2014

  20. [28]

    Rastegar, S.; Doughty, H.; and Snoek, C. 2024. Learn to categorize or categorize to learn? self-coding for generalized category discovery. Advances in Neural Information Processing Systems, 36

    2024

  21. [29]

    Saito, K.; and Saenko, K. 2021. Ovanet: One-vs-all network for universal domain adaptation. In Proceedings of the ieee/cvf international conference on computer vision, 9000--9009

    2021

  22. [30]

    Shi, W.; An, W.; Tian, F.; Chen, Y.; Wu, Y.; Wang, Q.; and Chen, P. 2024. A Unified Knowledge Transfer Network for Generalized Category Discovery. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 18961--18969

    2024

  23. [31]

    Shi, W.; An, W.; Tian, F.; Zheng, Q.; Wang, Q.; and Chen, P. 2023. A Diffusion Weighted Graph Framework for New Intent Discovery. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, 8033--8042

    2023

  24. [32]

    Vaze, S.; Han, K.; Vedaldi, A.; and Zisserman, A. 2022. Generalized category discovery. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 7492--7501

    2022

  25. [33]

    Vaze, S.; Vedaldi, A.; and Zisserman, A. 2024. No representation rules them all in category discovery. Advances in Neural Information Processing Systems, 36

    2024

  26. [34]

    Wen, X.; Zhao, B.; and Qi, X. 2022. A Simple Parametric Classification Baseline for Generalized Category Discovery. arXiv preprint arXiv:2211.11727

    2022 arXiv

  27. [35]

    Wolf, T.; Debut, L.; Sanh, V.; Chaumond, J.; Delangue, C.; Moi, A.; Cistac, P.; Rault, T.; Louf, R.; Funtowicz, M.; et al. 2019. HuggingFace's Transformers: State-of-the-art natural language processing. arXiv preprint arXiv:1910.03771

    2019 arXiv

  28. [36]

    Xie, J.; Girshick, R.; and Farhadi, A. 2016. Unsupervised deep embedding for clustering analysis. In International conference on machine learning, 478--487. PMLR

    2016

  29. [37]

    D.; and Hong, M

    Yang, B.; Fu, X.; Sidiropoulos, N. D.; and Hong, M. 2017. Towards k-means-friendly spaces: Simultaneous deep learning and clustering. In international conference on machine learning, 3861--3870. PMLR

    2017

  30. [38]

    Zhang, H.; Xu, H.; Lin, T.-E.; and Lyu, R. 2021 a . Discovering New Intents with Deep Aligned Clustering. In Proceedings of the AAAI Conference on Artificial Intelligence

    2021

  31. [39]

    H.; Chang, W.; and Yu, P

    Zhang, J.; Bui, T.; Yoon, S.; Chen, X.; Liu, Z.; Xia, C.; Tran, Q. H.; Chang, W.; and Yu, P. 2021 b . Few-shot intent detection via contrastive pre-training and fine-tuning. arXiv preprint arXiv:2109.06349

    2021 arXiv

  32. [40]

    Zhang, S.; Yan, C.; Yang, J.; Ren, C.; Bai, J.; Li, T.; and Li, Z. 2024 a . RoNID: New Intent Discovery with Generated-Reliable Labels and Cluster-friendly Representations. arXiv preprint arXiv:2404.08977

    2024 arXiv

  33. [41]

    Zhang, S.; Yang, J.; Bai, J.; Yan, C.; Li, T.; Yan, Z.; and Li, Z. 2024 b . New Intent Discovery with Attracting and Dispersing Prototype. arXiv preprint arXiv:2403.16913

    2024 arXiv

  34. [42]

    Zhang, Y.; Zhang, H.; Zhan, L.-M.; Wu, X.-M.; and Lam, A. 2022. New intent discovery with pre-training and contrastive learning. arXiv preprint arXiv:2205.12914

    2022 arXiv

  35. [43]

    Zhao, Y.; Chen, W.; Tan, X.; Huang, K.; and Zhu, J. 2022. Adaptive logit adjustment loss for long-tailed visual recognition. In Proceedings of the AAAI conference on artificial intelligence, volume 36, 3472--3480

    2022

  36. [44]

    Zhao, Z.; Li, X.; Zhai, Z.; and Chang, Z. 2024. Pseudo-supervised contrastive learning with inter-class separability for generalized category discovery. Knowledge-Based Systems, 289: 111477

    2024

  37. [45]

    Zhong, Z.; Fini, E.; Roy, S.; Luo, Z.; Ricci, E.; and Sebe, N. 2021. Neighborhood contrastive learning for novel class discovery. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 10867--10875

    2021

  38. [46]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...

  39. [47]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.