Pith. sign in

REVIEW 4 major objections 6 minor 49 references

LH-Mix: Local Hierarchy Correlation Guided Mixup over Hierarchical Prompt Tuning

T0 review · 4 major / 6 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 similarity-guided Mixup ratio lifts hierarchical text classification on three benchmark datasets.

desk verdict A modest but sound empirical paper: HPT+Mixup is new in HTC, the gains are real yet small, and the one novel component (the guided ratio) lacks direct significance testing. read the letter →

arxiv 2412.16963 v2 pith:PPB5ZIK3 submitted 2024-12-22 cs.CL

Fanshuang Kong , Richong Zhang , Ziqiao Wang This is my paper · ORCID
classification cs.CL
keywords hierarchicaltextclassificationmixupprompttuninglocalhierarchylabelcorrelationdataaugmentationmulti-labelBERT
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 claims that hierarchical text classification (HTC) can be improved by mixing not just a single label representation but the entire text-specific label path, with the strength of mixing set by how similar those paths are in a learned latent space. The proposed LH-Mix model encodes each text's local hierarchy (the labels it actually uses) as a depth-aligned prompt sequence, then interpolates hidden states at every depth between pairs of examples. The novel piece is a hand-designed function that maps the cosine similarity between two local-hierarchy embeddings to a Mixup ratio: highly similar paths are blended toward a 50/50 mix, dissimilar paths are barely mixed. On WOS, NYT, and RCV1-V2 the method reports the best score in five of six F1 metrics, with statistically significant gains over the two strongest baselines on most metrics, and the gains grow as training data is sparser and the hierarchy deeper.

What carries the argument

The load-bearing object is the local-hierarchy correlation-guided Mixup ratio, $\lambda = -(\beta-0.5)s^\alpha + \beta$, with $\alpha > 0$ controlling how quickly the blend strength changes with similarity and $\beta \in (0.5, 1]$ capping the maximum $\lambda$ (minimum mixing). Here $s$ is the normalized cosine similarity (Eq. 3) between the frozen [CLS] embeddings of two gold-label hierarchy sequences, e.g. '[CLS][Dth1]CS[Dth2]Machine Learning[SEP]'. This function carries the argument: it turns a static, uniformly sampled Beta ratio into a per-pair intensity that approaches 0.5 for highly correlated local hierarchies (strong mixing) and approaches $\beta$ for dissimilar ones (light mixing), and when applied to the hidden [MASK] states at each depth it is what lets the model learn sibling/peer correlations that the parent-child prompt alone does not encode.

What would settle it

Run LH-Mix with a per-pair ratio that is randomly shuffled independently of $s$: if shuffled ratios perform as well as the correlation-guided ones, the guidance carries no information and the reported gains come from Mixup's regularization alone. A faster check is to construct a dataset where gold-label sequences are lexical twins (high $s$) but their true label sets are nearly disjoint — under Eq. 4 they would be mixed hard, and if performance drops versus vanilla Mixup on that data, the similarity proxy is misleading.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that treating the local hierarchy as a sequence and applying Mixup under a similarity-adapted ratio captures implicit label correlations that both vanilla Mixup and graph-encoder baselines miss. Concretely, LH-Mix builds on HPT's depth-level prompt, replaces the [MASK] token with the gold label to form a 'local hierarchy sentence' for each input, and takes the frozen BERT [CLS] embedding of that sentence as the hierarchy's representation. For a pair of inputs, the normalized cosine similarity s of these embeddings is plugged into $\lambda = -(\beta-0.5)s^\alpha + \beta$ to produce the Mixup ratio, with $\alpha$ and $\beta$ tuned from a small grid. The mixed hidden states are fed through the depth classifiers, and the zero-bounded multi-label cross-entropy (ZMLCE) loss is mixed in the same proportion, which the paper argues mirrors vanilla Mixup's gradient behavior. The authors report best-in-table results on five of six metrics across WOS, NYT, and RCV1-V2, with t-tests showing significance on all but one comparison against HPT and all but two against HBGL, and they show the advantage widens when training data is downsampled to 10% and for deep or rare labels.

Load-bearing premise

The load-bearing premise is that the hand-designed mapping from local-hierarchy similarity to Mixup intensity — $\lambda = -(\beta-0.5)s^\alpha + \beta$ — correctly expresses how strongly any pair of label paths should be blended, and that the frozen BERT [CLS] embedding of the gold-label sequence is a trustworthy measure of that similarity; the paper itself notes there is no well-established theoretical framework for this relationship.

Editorial extensions

If this is right

  • Any hierarchical multi-label model that already uses depth-level prompts can add the same correlation-guided Mixup with a small grid search over $\alpha$ and $\beta$, and the paper's ablations suggest it should beat vanilla Mixup and graph encoders.
  • In low-resource settings, the relative gain of LH-Mix over both prompt-only and vanilla-Mixup training grows as training data shrinks to 50%, 25%, and 10%, so the method is a candidate for HTC with sparse labels.
  • Macro-F1 improves more than Micro-F1, which the authors attribute to better label-level correlation; if correct, this means Mixup-style regularization is a viable way to address rare and deep labels without changing the architecture.
  • The two-parameter family in Eq. 4 subsumes linear ($\alpha=1$) and nonlinear ($\alpha \neq 1$) similarity-to-ratio mappings, so the choice of $\alpha$ can be tuned per dataset; the paper's parameter study shows NYT and RCV1-V2 prefer different $\alpha$, meaning the family has enough capacity to express the difference.

Reading between the lines

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

  • Because the similarity measure uses a frozen encoder on gold-label sequences, LH-Mix is computationally cheap at train time but cannot adapt its correlation measure as the classifier learns; a learnable but still frozen-like encoder (e.g., a stop-gradient projection) is a natural next test.
  • The same ratio-law could transfer to other structured prediction tasks where labels form a tree, such as taxonomic image classification or hierarchical retrieval, wherever a text-like sequence can be built from the label path.
  • The paper's heuristic mapping suggests a testable hypothesis about Mixup generally: instance-pair similarity should control blend strength along a monotone family; the exact exponent $\alpha$ may be a per-task free parameter, which the reported $\alpha$ sensitivity on NYT versus RCV1-V2 supports.
  • If the gains indeed come from peer/sibling correlation rather than from extra regularization, then LH-Mix should pair well with graph-based global hierarchy encoders, which the paper does not combine with HBGL; an ensemble or hybrid remains untested.
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 / 6 minor

Summary. The paper presents LH-Mix, a method for hierarchical text classification that represents a text-specific local hierarchy as a depth-level prompt sequence and then applies Mixup to the [MASK] hidden states at each depth. The Mixup ratio is not sampled from a Beta distribution; instead, it is computed by Eq. (4) from the normalized cosine similarity of local-hierarchy representations. The authors evaluate on WOS, NYT, and RCV1-V2, compare with hierarchy-aware, LLM, and BERT-based baselines, and report that LH-Mix achieves the best performance on five of six metrics, with larger gains on Macro-F1 and in low-resource settings. They also provide five-seed statistics and t-tests against HPT and HBGL in Appendix A.2.

Significance. If the reported gains are robust, the paper makes a useful and simple contribution: it shows that a hierarchy-correlation-guided Mixup ratio, on top of hierarchical prompt tuning, improves HTC over strong baselines, especially for rare and deep labels and in sparse training regimes. The gradient derivation in Eq. (8) is a correct chain-rule application, and the inclusion of five-seed runs and t-tests against SOTA baselines is a strength. However, the central novelty, namely the mechanism of Eq. (4), is not directly statistically validated, and the headline numbers in Table 2 are inconsistent with the multi-seed statistics in Table 6, so the strength of the empirical claim currently exceeds the evidence.

major comments (4)
  1. [Section 5.4 and Appendix A.2 (Tables 2 and 6)] The main results in Table 2 do not match the five-seed means in Table 6. For example, WOS Macro-F1 is 82.07 in Table 2 but 81.82±0.18 in Table 6, and RCV1-V2 Macro-F1 is 71.93 versus 71.49±0.50. Since Table 2 is used to claim the best performance in five out of six metrics, the paper should either report the multi-seed mean in Table 2 or explicitly state that Table 2 reports a single run and identify the corresponding seed. This is not a cosmetic issue because several reported margins are smaller than the standard deviations in Table 6.
  2. [Table 3 and Appendix A.2] The only comparison that isolates the contribution of Eq. (4) is Prompt+Mixup versus Prompt+LH-Mix in Table 3, with single-run gains of +0.14/+0.25 on WOS, +0.22/+0.63 on NYT, and +0.02/+0.54 on RCV1-V2 in Micro-F1/Macro-F1. These numbers are not accompanied by standard deviations or paired significance tests. The five-seed statistics in Table 6 and the t-tests in Table 7 compare only against HPT and HBGL, not against the vanilla Mixup baseline. Given that Section 5.9 itself cautions that Micro-F1 changes are small and may not accurately reflect the final results, the claim that the correlation-guided ratio is responsible for the observed improvement is not statistically established. Please add multi-seed results and paired significance tests for LH-Mix versus the vanilla Mixup variant.
  3. [Section 4.2 and Section 5.2] Eq. (4) is explicitly heuristic, and its two parameters alpha and beta are chosen from grids on the same datasets. The manuscript does not state the selected per-dataset values, the selection criterion, or whether the grid search was performed on a held-out split. Because the paper concedes in Section 4.2 that there is no well-established theoretical framework for the similarity-to-ratio map, the possibility that the reported gains reflect tuned hyperparameters rather than the proposed mechanism cannot be ruled out. Please report the chosen alpha and beta for each dataset, the validation procedure, and results for neighboring grid values, at least for the datasets shown in Figure 5.
  4. [Section 4.1] The guiding signal s is not validated. The paper assumes that the frozen BERT [CLS] embedding of the gold-label sequence is a trustworthy measure of local-hierarchy correlation, but no evidence is provided that this cosine similarity tracks actual hierarchy distance or that it is what drives the gains. Figure 6 is a qualitative t-SNE case study after training and does not isolate the similarity measure. A concrete test would be to compare s against a structural distance such as the depth of the lowest common ancestor on a sample of label pairs, or to ablate Eq. (4) by replacing s with a random or constant value. Without such a test, the correlation-guided interpretation remains an assumption.
minor comments (6)
  1. [Section 4.1] The text says local hierarchy representations use the same pre-trained model encoder and then says the encoder used for similarity calculation is distinct from the one used for classification; please clarify whether the similarity encoder is frozen and separate from the fine-tuned prompt encoder.
  2. [Section 5.2] The sentence 'we will release code our the Github' should read 'we will release our code on Github.'
  3. [Section 5.10] 'Noted, for conciseness and clarity' should be 'Note that, for conciseness and clarity.'
  4. [Table 2] The underlining of current SOTA entries is not visible in the manuscript text; please ensure the formatting is correct.
  5. [Figures 4, 5, and 7] These figures report curves without error bars or multi-seed aggregation; please state whether they are single-run results.
  6. [Eq. (8)] The expression shown is the derivative of the mixed loss with respect to the mixed prediction score ep^d, not the full gradient with respect to model parameters; the latter also involves the chain rule through Eq. (5). Please state this explicitly to avoid confusion.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: LH-Mix's novelty is a heuristic Mixup-ratio map evaluated on held-out test sets, not derived from its inputs.

full rationale

The paper's claimed derivation chain is empirical rather than circular. The novel component, Eq. 4, is an explicitly heuristic mapping from a computed local-hierarchy similarity s to a Mixup ratio lambda, with alpha and beta selected by grid search. Eq. 4 is not derived from the target result, and the central claims are evaluated on held-out test sets of WOS, NYT, and RCV1-V2 against external baselines and ablations. The similarity s uses gold-label sequences through a frozen BERT encoder, but only during training; this is standard supervised training-time label use, not test-time leakage. The self-citations ([12], [13], [44]) support generic Mixup methodology and are not load-bearing; no uniqueness theorem or external result is invoked to force the design. The parameter analysis and t-SNE case study are post-hoc empirical observations, not reductions of the prediction to its inputs by construction. The skeptical concern about the absence of paired significance tests against the +Mixup ablation is a matter of evidence strength, not circularity. Therefore, no circular step is present, and the derivation is self-contained for the purposes of this review.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The core new mechanism (Eq. 4) rests on two tuned hyperparameters and two domain assumptions about the similarity measure and the interpolation space; no new physical or ontological entities are introduced.

free parameters (2)
  • alpha (Eq. 4) = not reported per dataset
    Controls how quickly the Mixup weight lambda decreases as local-hierarchy similarity s increases. Chosen from {0.1, 0.3, 0.6, 1, 2, 5, 10} by grid search on the three datasets; final per-dataset values are not given in the paper.
  • beta (Eq. 4) = not reported per dataset
    Sets the upper bound of lambda and therefore the minimum impact of Mixup. Chosen from {0.7, 0.75, 0.8, 0.85, 0.9, 0.95, 1} by grid search; final values not reported.
assumptions (4)
  • ad hoc to paper Eq. 4: lambda = -(beta-0.5)*s^alpha+beta is a valid model of the relationship between local-hierarchy similarity and beneficial Mixup intensity.
    The paper explicitly states this is a heuristic design motivated by intuition, without theoretical justification (Section 4.2).
  • domain assumption The [CLS] representation of a gold-label local-hierarchy sequence, from a frozen BERT encoder, measures the correlation between local hierarchies.
    Section 4.1 builds the similarity s on these embeddings and assumes they capture semantic and taxonomic closeness relevant to HTC; this is asserted, not validated.
  • domain assumption Depth-level [MASK] hidden states from the hierarchical prompt are linearly interpolable across documents, so Mixup in that space is meaningful.
    The method mixes hidden states at each depth (Eq. 5) without evidence that the prompt alignment makes the per-depth representation spaces comparable across texts.
  • domain assumption The ZMLCE loss from HPT is an appropriate base loss for hierarchical text classification.
    The paper adopts zero-bounded multi-label cross-entropy from HPT (Section 3.2) and relies on its behavior for the loss-mixing formulation in Eq. 7.

how reviews work

0 comments
Cite this review

Pith. "Pith review of LH-Mix: Local Hierarchy Correlation Guided Mixup over Hierarchical Prompt Tuning." pith.science (2026). https://pith.science/paper/PPB5ZIK3

@misc{pith2026241216963,
  author       = {Pith},
  title        = {Pith review of: LH-Mix: Local Hierarchy Correlation Guided Mixup over Hierarchical Prompt Tuning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PPB5ZIK3}},
  note         = {Machine review of arXiv:2412.16963}
}
read the original abstract

Hierarchical text classification (HTC) aims to assign one or more labels in the hierarchy for each text. Many methods represent this structure as a global hierarchy, leading to redundant graph structures. To address this, incorporating a text-specific local hierarchy is essential. However, existing approaches often model this local hierarchy as a sequence, focusing on explicit parent-child relationships while ignoring implicit correlations among sibling/peer relationships. In this paper, we first integrate local hierarchies into a manual depth-level prompt to capture parent-child relationships. We then apply Mixup to this hierarchical prompt tuning scheme to improve the latent correlation within sibling/peer relationships. Notably, we propose a novel Mixup ratio guided by local hierarchy correlation to effectively capture intrinsic correlations. This Local Hierarchy Mixup (LH-Mix) model demonstrates remarkable performance across three widely-used datasets.

Figures

Figures reproduced from arXiv: 2412.16963 by the authors.

Figure 1
Figure 1. (a) A toy example of global hierarchy in HTC. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Illustration of LH-Mix. The light orange color scheme represents elements of [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Curves corresponding to Eq. 4. We separately plot [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 5
Figure 5. Figure 5: Performance on different 𝛽 and 𝛼, when fixing 𝛼 = 1 and 𝛽 = 1 respectively. the final results. Therefore, we temporarily focus on the Macro-F1 for consideration. 5.9.1 Effect of 𝛼. When 𝛽 = 1 is fixed, for NYT, the Macro-F1 initially increases and then decreases with i…
Figure 4
Figure 4. Figure 4: Performance on different downsample ratios to [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 6
Figure 6. Figure 6: (a) Example of hierarchically related labels on “Hockey”. (b) The visualization of label correlations related to “Hockey” [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Performance of comparable models on different [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 32 canonical work pages

  1. [1]

    Raphael Baena, Lucas Drumetz, and Vincent Gripon. 2022. Preventing manifold intrusion with locality: Local mixup. arXiv preprint arXiv:2201.04368 (2022)

    work page Pith review arXiv 2022

  2. [2]

    Siddhartha Banerjee, Cem Akkaya, Francisco Perez-Sorrosal, and Kostas Tsiout- siouliklis. 2019. Hierarchical Transfer Learning for Multi-label Text Classification.. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. https://doi.org/10.18653/v1/p19-1633

    doi:10.18653/v1/p19-1633 2019

  3. [3]

    Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. Language models are few-shot learners. Advances in neural information processing systems 33 (2020), 1877–1901

    2020

  4. [4]

    Oscar Chang, Dung N Tran, and Kazuhito Koishida. 2021. Single-Channel Speech Enhancement Using Learnable Loss Mixup.. In Interspeech. 2696–2700

    work page 2021

  5. [5]

    Boli Chen, Xin Huang, Lin Xiao, Zixin Cai, and Liping Jing. 2020. Hyperbolic interaction model for hierarchical multi-label classification. In Proceedings of the AAAI conference on artificial intelligence . 7496–7503

    work page 2020

  6. [6]

    Haibin Chen, Qianli Ma, Zhenxi Lin, and Jiangyue Yan. 2021. Hierarchy-aware Label Semantics Matching Network for Hierarchical Text Classification. In Pro- ceedings of the 59th Annual Meeting of the Association for Computational Linguis- tics and the 11th International Joint Conference on Natural Language Processing . 4370–4379

    work page 2021

  7. [7]

    Jiaao Chen, Zichao Yang, and Diyi Yang. 2020. MixText: Linguistically-Informed Interpolation of Hidden Space for Semi-Supervised Text Classification. InProceed- ings of the 58th Annual Meeting of the Association for Computational Linguistics . 2147–2157. https://doi.org/10.18653/v1/2020.acl-main.194

    doi:10.18653/v1/2020.acl-main.194 2020

  8. [8]

    Hsin-Ping Chou, Shih-Chieh Chang, Jia-Yu Pan, Wei Wei, and Da-Cheng Juan

    work page

Show all 49 references
  1. [9]

    Zhongfen Deng, Hao Peng, Dongxiao He, Jianxin Li, and Philip Yu. 2021. HTCIn- foMax: A Global Model for Hierarchical Text Classification via Information Maximization. In Proceedings of the 2021 Conference of the North American Chap- ter of the Association for Computational Lin...

    2021

  2. [10]

    Susan Dumais and Hao Chen. 2000. Hierarchical classification of Web content. In Proceedings of the 23rd annual international ACM SIGIR conference on Research and development in information retrieval . https://doi.org/10.1145/345508.345593

    2000

  3. [11]

    Hongyu Guo. 2020. Nonlinear mixup: Out-of-manifold data augmentation for text classification. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 34. 4044–4051

    2020

  4. [12]

    Hongyu Guo, Yongyi Mao, and Richong Zhang. 2019. Augmenting Data with Mixup for Sentence Classification: An Empirical Study. CoRR abs/1905.08941 (2019). arXiv:1905.08941 http://arxiv.org/abs/1905.08941

    2019 arXiv

  5. [13]

    Hongyu Guo, Yongyi Mao, and Richong Zhang. 2019. Mixup as locally linear out-of-manifold regularization. In Proceedings of the AAAI conference on artificial intelligence, Vol. 33. 3714–3722

    2019

  6. [14]

    Jie He, Víctor Gutiérrez-Basulto, Jeff Z Pan, et al . 2023. Instances and Labels: Hierarchy-aware Joint Supervised Contrastive Learning for Hierarchical Multi- Label Text Classification. arXiv preprint arXiv:2310.05128 (2023)

    2023 arXiv

  7. [15]

    SangHun Im, GiBaeg Kim, Heung-Seon Oh, Seongung Jo, and Dong Hwan Kim

  8. [16]

    Ting Jiang, Deqing Wang, Leilei Sun, Zhongzhi Chen, Fuzhen Zhuang, and Qinghong Yang. 2022. Exploiting Global and Local Hierarchies for Hierarchical Text Classification. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing. 4030–4039

    2022

  9. [17]

    Diederik P Kingma and Jimmy Ba. 2014. Adam: A method for stochastic opti- mization. arXiv preprint arXiv:1412.6980 (2014)

    2014 arXiv

  10. [18]

    Thomas N Kipf and Max Welling. 2016. Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907 (2016)

    2016 arXiv

  11. [19]

    Kamran Kowsari, Donald E Brown, Mojtaba Heidarysafa, Kiana Jafari Meimandi, Matthew S Gerber, and Laura E Barnes. 2017. Hdltex: Hierarchical deep learning for text classification. In 2017 16th IEEE international conference on machine learning and applications (ICMLA) . IEEE, 364–371

    2017

  12. [20]

    Lewis, Yiming Yang, TonyG

    DavidD. Lewis, Yiming Yang, TonyG. Rose, and Fan Li. 2004. RCV1: A New Benchmark Collection for Text Categorization Research. Journal of Machine Learning Research,Journal of Machine Learning Research (Dec 2004)

    2004

  13. [21]

    Xudong Mao, Yun Ma, Zhenguo Yang, Yangbin Chen, and Qing Li. 2019. Virtual Mixup Training for Unsupervised Domain Adaptation. arXiv: Computer Vision and Pattern Recognition,arXiv: Computer Vision and Pattern Recognition (May 2019)

    2019

  14. [22]

    Yuning Mao, Jingjing Tian, Jiawei Han, and Xiang Ren. 2019. Hierarchical Text Classification with Reinforced Label Assignment. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Langua...

    2019 doi

  15. [23]

    OpenAI. 2022. Chatgpt: Optimizing language models for dialogue. https://openai. com/blog/chatgpt

    2022

  16. [24]

    Hao Peng, Jianxin Li, Qiran Gong, Senzhang Wang, Lifang He, Bo Li, Lihong Wang, and PhilipS. Yu. 2019. Hierarchical Taxonomy-Aware and Attentional Graph Capsule RCNNs for Large-Scale Multi-Label Text Classification. IEEE Transactions on Knowledge and Data Engineering,IEEE Tran...

    2019

  17. [25]

    Nils Reimers and Iryna Gurevych. 2019. Sentence-bert: Sentence embeddings using siamese bert-networks. arXiv preprint arXiv:1908.10084 (2019)

    2019 arXiv

  18. [26]

    Evan Sandhaus. 2008. The New York Times Annotated Corpus. Linguistic Data Consortium (Oct 2008)

    2008

  19. [27]

    Ramit Sawhney, Megh Thakkar, Shrey Pandit, Ritesh Soun, Di Jin, Diyi Yang, and Lucie Flek. 2022. DMIX: Adaptive Distance-aware Interpolative Mixup. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) . 606–612

    2022

  20. [28]

    Kazuya Shimura, Jiyi Li, and Fumiyo Fukumoto. 2018. HFT-CNN: Learning Hierarchical Category Structure for Multi-label Short Text Categorization.. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. https://doi.org/10.18653/v1/d18-1093

    2018 doi

  21. [29]

    Carlos N Silla and Alex A Freitas. 2011. A survey of hierarchical classification across different application domains. Data mining and knowledge discovery 22 (2011), 31–72

    2011

  22. [30]

    Junru Song, Feifei Wang, and Yang Yang. 2023. Peer-Label Assisted Hierarchical Text Classification. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) . 3747–3758

    2023

  23. [31]

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, et al. 2023. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971 (2023)

    2023 arXiv

  24. [32]

    Laurens Van der Maaten and Geoffrey Hinton. 2008. Visualizing data using t-SNE. Journal of machine learning research 9, 11 (2008)

    2008

  25. [33]

    Petar Velickovic, Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Lio, Yoshua Bengio, et al. 2017. Graph attention networks. stat 1050, 20 (2017), 10–48550

    2017

  26. [34]

    Zihan Wang, Peiyi Wang, Lianzhe Huang, Xin Sun, and Houfeng Wang. 2022. Incorporating Hierarchy into Text Encoder: a Contrastive Learning Approach for Hierarchical Text Classification. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (...

    2022

  27. [35]

    Zihan Wang, Peiyi Wang, Tianyu Liu, Binghuai Lin, Yunbo Cao, Zhifang Sui, and Houfeng Wang. 2022. HPT: Hierarchy-aware prompt tuning for hierarchical text classification. arXiv preprint arXiv:2204.13413 (2022)

    2022 arXiv

  28. [36]

    Barros, and Ricardo Cerri

    Jonatas Wehrmann, RodrigoC. Barros, and Ricardo Cerri. 2018. Hierarchical Multi-Label Classification Networks. International Conference on Machine Learn- ing,International Conference on Machine Learning (Jul 2018)

    2018

  29. [37]

    Yeming Wen, Ghassen Jerfel, Rafael Muller, Michael W Dusenberry, Jasper Snoek, Balaji Lakshminarayanan, and Dustin Tran. 2020. Combining ensembles and data augmentation can harm your calibration. arXiv preprint arXiv:2010.09875 (2020)

    2020 arXiv

  30. [38]

    Jiawei Wu, Wenhan Xiong, and WilliamYang Wang. 2019. Learning to Learn and Predict: A Meta-Learning Approach for Multi-Label Classification. Cornell University - arXiv,Cornell University - arXiv (Sep 2019)

    2019

  31. [39]

    Yuan Wu, Diana Inkpen, and Ahmed El-Roby. 2020. Dual mixup regularized learning for adversarial domain adaptation. In Computer Vision–ECCV 2020: 16th European Conference, Proceedings, Part XXIX 16 . 540–555

    2020

  32. [40]

    Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, and Tie-Yan Liu. 2021. Do transformers really perform badly for graph representation? Advances in neural information processing systems 34 (2021), 28877–28888

    2021

  33. [41]

    Sangdoo Yun, Dongyoon Han, Seong Joon Oh, Sanghyuk Chun, Junsuk Choe, and Youngjoon Yoo. 2019. CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features. In International Conference on Computer Vision (ICCV)

    2019

  34. [42]

    Dauphin, and David Lopez-Paz

    Hongyi Zhang, Moustapha Cisse, YannN. Dauphin, and David Lopez-Paz. 2017. mixup: Beyond Empirical Risk Minimization. arXiv,Learning (Oct 2017)

    2017

  35. [43]

    Shaofeng Zhang, Meng Liu, Junchi Yan, Hengrui Zhang, Lingxiao Huang, Xi- aokang Yang, and Pinyan Lu. 2022. M-mix: Generating hard negatives via multi- sample mixing for contrastive learning. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Minin...

    2022

  36. [44]

    Yanzhao Zhang, Richong Zhang, Samuel Mensah, Xudong Liu, and Yongyi Mao

  37. [45]

    Jie Zhou, Chunping Ma, Dingkun Long, Guangwei Xu, Ning Ding, Haoyu Zhang, Pengjun Xie, and Gongshen Liu. 2020. Hierarchy-Aware Global Model for Hi- erarchical Text Classification. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics . 1106–1117

    2020

  38. [46]

    He Zhu, Chong Zhang, Junjie Huang, Junran Wu, and Ke Xu. 2023. HiTIN: Hierarchy-aware Tree Isomorphism Network for Hierarchical Text Classification. arXiv preprint arXiv:2305.15182 (2023). LH-Mix: Local Hierarchy Correlation Guided Mixup over Hierarchical Prompt Tuning KDD ’25...

    2023 arXiv

  39. [2020]

    In Computer Vision–ECCV 2020 Workshops: Glasgow, UK, August 23–28, 2020, Proceedings, Part VI 16

    Remix: rebalanced mixup. In Computer Vision–ECCV 2020 Workshops: Glasgow, UK, August 23–28, 2020, Proceedings, Part VI 16 . Springer, 95–110

    2020

  40. [2022]

    In Proceedings of the AAAI Conference on Artificial Intelligence , Vol

    Unsupervised sentence representation via contrastive learning with mixing negatives. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 36. 11730–11738

  41. [2023]

    In Proceedings of the AAAI Conference on Artificial Intelligence

    Hierarchical text classification as sub-hierarchy sequence generation. In Proceedings of the AAAI Conference on Artificial Intelligence . 12933–12941

Pith tools

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