Pith. sign in

REVIEW 4 major objections 4 minor 40 references

Adaptive Prompting for Continual Relation Extraction: A Within-Task Variance Perspective

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 Per-task prompt pools plus Gaussian replay match memory-based continual relation extraction

desk verdict A competently executed rehearsal-free CRE method with a genuinely new prompt-pool plus Gaussian replay combination, whose load-bearing generative assumption needs more evidence before the memory-efficiency claim can be trusted. read the letter →

arxiv 2412.08285 v5 pith:ROQ5Y2R2 submitted 2024-12-11 cs.CL cs.LG

Minh Le , Tien Ngoc Luu , An Nguyen The , Thanh-Thien Le , Trang Nguyen , Tung Thanh Nguyen , Linh Ngo Van , Thien Huu Nguyen This is my paper · ORCID
classification cs.CLcs.LG
keywords continualrelationextractioncatastrophicforgettingrehearsal-freelearningpromptpoolprefix-tuningmixtureofexpertsgenerativereplaytaskpredictor
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that continual relation extraction can shed the memory buffer without losing accuracy, provided each task keeps a private prompt pool and old relations are replayed as synthetic latent samples rather than stored sentences. Existing prompt-based methods fail, the paper says, because prompt selection is inaccurate, shared parameters still forget, and neither cross-task nor within-task variation is handled well. WAVE-CRE assigns each task its own pool of one-expert prompts, chosen by matching a query vector to learned keys, and fits a separate Gaussian to the encoded representations and to the query vectors of every relation seen so far. Those Gaussians generate the training data for the relation classifier and for a task predictor that selects the correct pool at test time, so prior knowledge lives in means and covariances instead of a buffer. On FewRel and TACRED, the method beats the rehearsal-free and prompt-based baselines, nearly matches rehearsal-based methods, and tops them on FewRel's final task.

What carries the argument

The load-bearing mechanism is a task-specific prompt pool paired with per-relation Gaussian replay. Each pool $P_t$ contains $M$ prompts, each a single prefix expert with length $L_p=2$ and its own learnable key; the query vector $q(x)$ selects the top-$K$ prompts by cosine similarity (equation 13), and only those prompts and the shared classifier are trained for the current task. The design is justified by viewing prefix-tuning as inserting sparse experts into the mixture-of-experts structure of self-attention, so a separate key for each expert gives flexible selection and captures within-task variation while separate pools keep tasks apart. Forgetting in shared parameters is handled by the generative models: for each relation $r$, equations (16) and (17) fit Gaussians $\mathcal{N}(\mu^r_z,\Sigma^r_z)$ to the prompted representations and $\mathcal{N}(\mu^r_q,\Sigma^r_q)$ to the query vectors, and equations (18) and (19) train the task predictor $\psi$ and relation classifier $\phi$ on samples drawn from those Gaussians.

What would settle it

On a relation whose encoded examples form two or more well-separated clusters, fit a two-component Gaussian mixture to the training representations from $D^r_z$. If the relation classifier trained on samples from the single fitted Gaussian loses more than a few points relative to one trained on the mixture, while the mixture version does not, then the single-Gaussian assumption is the weak link and the replay will misrepresent that relation.

Watch

Extended reading notes

Core claim

The paper's central claim is that catastrophic forgetting in continual relation extraction can be controlled without storing any training instances by combining per-task prompt pools with generative replay of latent representations. WAVE-CRE freezes a BERT encoder, trains only the current task's prompt pool and the shared classifier head, and after each task fits a Gaussian per relation to the prompted representations $z^r=f_r(x^r_p)$ and another to the query vectors $q(x^r)$, following equations (16) and (17). Samples from these Gaussian distributions are then used in equations (18) and (19) to train the task predictor and the relation classifier, consolidating old tasks in the shared parameters. The experiments report that WAVE-CRE outperforms L2P, EPI, and HiDe-Prompt at every stage on both datasets, reaches a final accuracy of 85.0 on FewRel and 78.7 on TACRED, and on FewRel's last task exceeds the rehearsal-based CDec+ACA baseline.

Load-bearing premise

The method assumes that the encoded representations of each relation sit in one bell-shaped cluster, so a mean vector and covariance matrix can replace the real training examples; if a relation's examples split into several distinct groups or skew sharply, the replayed samples misrepresent it and the forgetting protection fails.

Editorial extensions

If this is right

  • Because only per-relation means and covariances are stored, the approach removes the storage and privacy costs of rehearsal buffers while keeping most of their accuracy.
  • The per-task prompt pool with a single expert per prompt keeps the learnable parameter count low (3.8M in the reported setup), so the method remains parameter-efficient as the number of tasks grows.
  • Training the task predictor on per-relation query distributions rather than on per-task aggregates improves task-identity precision, which removes the train/test prompt-selection mismatch that limits earlier prompt methods.
  • On the reported benchmarks the final accuracy lands close to the best rehearsal-based systems (85.0 on FewRel, 78.7 on TACRED), so the method is a practical substitute when buffers are undesirable.

Reading between the lines

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

  • If the per-relation Gaussian fit is adequate, the same representation-replay recipe should transfer to other continual text classification problems that use a frozen encoder, not just relation extraction.
  • The task predictor's per-relation output can be read as a soft assignment over relations; one could threshold its confidence to reject low-certainty inputs instead of always committing to a prompt pool.
  • A head-to-head comparison against a variant that replays the real stored representations would isolate exactly how much accuracy the bell-shaped approximation costs.
  • Because the encoder stays frozen, the method inherits whatever biases the off-the-shelf features carry; unfreezing the encoder would change the relation representation geometry and could break the fitted Gaussians.
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 WAVE-CRE, a rehearsal-free continual relation extraction method that combines task-specific prompt pools with per-relation Gaussian generative models. For each task, a dedicated prompt pool is learned with key-query selection (Section 3.1). After training, per-relation Gaussian distributions are fitted to prompted representations and query features (Section 3.2), and synthetic samples from these Gaussians are used to train a task predictor and the shared relation classifier (Section 3.3). Experiments on FewRel and TACRED report consistent gains over prompt-based and rehearsal-free baselines, and near-parity or slight superiority over rehearsal-based methods on the final task (Table 1).

Significance. If the reported results are reliable, WAVE-CRE offers a genuinely rehearsal-free alternative with competitive accuracy, and the within-task variance framing of prompt pools is a useful conceptual contribution. The paper covers two standard benchmarks, consistently improves over rehearsal-free baselines, and includes ablations isolating the prompt-pool design. However, the central empirical claim is not yet fully supported: the Gaussian generative assumption is unvalidated, no statistical significance is reported for the key comparisons, and a core ablation is run only on one dataset in a simplified setting. These gaps are load-bearing because the method's rehearsal-free mechanism depends entirely on the fidelity of the fitted Gaussians.

major comments (4)
  1. [Section 3.2, Eqs. (16)-(17)] The Gaussian assumption is load-bearing and unvalidated. The task predictor (Eq. 18) and relation classifier (Eq. 19) are trained exclusively on samples drawn from the per-relation Gaussians, while at test time they receive real q and z. If the true latent distributions are multi-modal or skewed, the synthetic replay will misrepresent past relations, degrading both the task predictor and the relation classifier. The paper reports no fit diagnostics (e.g., goodness-of-fit, visualization, or comparison to stored-feature replay) and no ablation with a more flexible generative model. Additionally, storing full covariance matrices costs O(D^2) per relation (about 376 MB for 80 relations at D=768), which weakens the stated memory-efficiency motivation; switching to diagonal covariance would make the fit even more fragile. Please validate the Gaussian assumption or bound the harm by ablating synthetic replay against stored-feature replay.
  2. [Table 1, Section 4.1 (Evaluation Metrics)] The central empirical claims are not supported by statistical evidence. The paper reports "mean accuracy on 5 different random seeds" but gives no standard deviations, confidence intervals, or significance tests. The key comparisons are close: on TACRED T10, WAVE-CRE (78.7) is below CRE-DAS (79.1), and on FewRel T10 the margin over CDec+ACA is only 0.2 points (85.0 vs 84.8). Without variance information, the claims of "superior performance" and "nearly equivalent to rehearsal-based baselines" are not established. Please report error bars and, where appropriate, paired significance tests across seeds.
  3. [Section 4.3, Table 2] The ablation supporting the task-specific prompt pool is run only on TACRED in the task-incremental setting, and the 1.8-point improvement on the final task (85.2 vs 83.4) has no confidence interval. Since the prompt pool is a core component of the proposed method, the ablation should be repeated on both datasets and in the class-incremental setting (the main evaluation protocol), or the claim about within-task variance should be weakened accordingly.
  4. [Section 3.1, Eq. (15) and Section 4.1 (Implementation Details)] The hyperparameters M (prompt pool size), K (number of selected prompts), and λ (regularization weight) are not reported; the paper states only that they were tuned via random search. This prevents reproducibility and makes it difficult to assess sensitivity to these choices. Please report the selected values and the search ranges.
minor comments (4)
  1. [Table 1] In the TACRED block, the EA-EMR value at T5 is reported as "24" while all other values use two decimal places; please format consistently.
  2. [Section 3.2] The sentence "Future works may explore alternative generative models" is the only acknowledgment of the Gaussian assumption's limitation; please add a sentence explicitly noting the risk of distribution mismatch between synthetic and real representations at test time.
  3. [Throughout] The method name appears with varying spacing as "W A VE-CRE", "WAVE-CRE", and "W A VE-CRE"; please unify the notation throughout the text, figures, and captions.
  4. [Algorithm 1] Step 3 uses the notation "xB ∈ Dt" but Dt is not listed among the inputs; please clarify that Dt denotes the t-th training dataset.

Circularity Check

0 steps flagged · score 2.0 of 10

No load-bearing circularity; central claims rest on held-out benchmark evaluation, not on fitted parameters or self-cited results.

full rationale

The paper's central claims are empirical: WAVE-CRE outperforms rehearsal-free baselines and approaches rehearsal-based methods on FewRel and TACRED held-out test sets (Table 1). The generative replay in Section 3.2 fits per-relation Gaussian distributions (Eqs. 16-17) to training-set prompted representations and queries, then trains the task predictor (Eq. 18) and relation classifier (Eq. 19) on samples drawn from those fitted Gaussians. This is a standard pseudorehearsal construction; the fit is not renamed as a prediction, and evaluation is on real test features, so the benchmark result is not forced by the fitted parameters. The weakest assumption, that each relation's latent z and q are approximately Gaussian, is a correctness/robustness risk rather than a circularity: a poor fit would degrade the method but would not make the derivation equivalent to its inputs. The MoE-prefix connection is motivated by a self-citation (Le et al. 2024a), but the paper reproduces the mathematical derivation in Eqs. 6-11 rather than importing an unverified uniqueness claim, and this connection is used as design inspiration, not as evidence for the reported accuracy. No ansatz is smuggled in via citation, and no known result is merely renamed. The only circularity-adjacent feature is the presence of several self-citations among the authors (Le et al. 2024a, 2024b, 2024d); none is load-bearing because the cited results are either reproduced in the text or serve only as background. Accordingly, the paper is substantially self-contained against external benchmarks, and the circularity score is 2 reflecting only minor, non-load-bearing self-citation.

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

The method introduces no new physical or conceptual entities. It relies on fitted Gaussian summaries and a set of hyperparameters, plus several background assumptions, notably the self-cited MoE-prefix equivalence and the Gaussian distribution of latent representations.

free parameters (4)
  • Prompt pool size M = not reported
    Number of prompts per task is a design choice; the paper says it is consistent across tasks but does not state its value.
  • Number of selected prompts K = 8 (with L=1 in Table 3)
    Top-K prompt selection; the paper notes that K is chosen to keep total experts equal in ablations, but the main runs use a value selected by random search.
  • Regularization weight lambda = not reported
    Hyperparameter in equation (15) balancing classification loss and key-query distance; tuned via random search without reporting the final value.
  • Per-relation Gaussian means and covariances = estimated from training data
    For each relation, mean and covariance of prompted and query representations are fit to the training set and used to generate replay samples; the central claim depends on these fits.
assumptions (4)
  • domain assumption BERT provides fixed, useful representations for relation extraction
    The encoder is frozen; the method assumes its representations are sufficient for all tasks encountered in the benchmark.
  • domain assumption Prefix-tuning can be viewed as adding experts to Mixture-of-Experts models
    Adopted from Le et al. 2024a by the same authors, used as motivation for task-specific prompt pools; no independent derivation is given in this paper.
  • ad hoc to paper Representations of each relation are Gaussian
    Gaussian is chosen for memory efficiency; no evidence is provided that latent representations are normally distributed, and the paper does not test alternative densities.
  • domain assumption The task predictor can reliably infer task identity at test time
    Wrong task predictions select the wrong prompt pool and cause forgetting; the paper evaluates this only on two benchmarks and does not analyze failure modes.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Adaptive Prompting for Continual Relation Extraction: A Within-Task Variance Perspective." pith.science (2026). https://pith.science/paper/ROQ5Y2R2

@misc{pith2026241208285,
  author       = {Pith},
  title        = {Pith review of: Adaptive Prompting for Continual Relation Extraction: A Within-Task Variance Perspective},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ROQ5Y2R2}},
  note         = {Machine review of arXiv:2412.08285}
}
read the original abstract

To address catastrophic forgetting in Continual Relation Extraction (CRE), many current approaches rely on memory buffers to rehearse previously learned knowledge while acquiring new tasks. Recently, prompt-based methods have emerged as potent alternatives to rehearsal-based strategies, demonstrating strong empirical performance. However, upon analyzing existing prompt-based approaches for CRE, we identified several critical limitations, such as inaccurate prompt selection, inadequate mechanisms for mitigating forgetting in shared parameters, and suboptimal handling of cross-task and within-task variances. To overcome these challenges, we draw inspiration from the relationship between prefix-tuning and mixture of experts, proposing a novel approach that employs a prompt pool for each task, capturing variations within each task while enhancing cross-task variances. Furthermore, we incorporate a generative model to consolidate prior knowledge within shared parameters, eliminating the need for explicit data storage. Extensive experiments validate the efficacy of our approach, demonstrating superior performance over state-of-the-art prompt-based and rehearsal-free methods in continual relation extraction.

Figures

Figures reproduced from arXiv: 2412.08285 by the authors.

Figure 1
Figure 1. Overall framework of WAVE-CRE. To prevent information loss across tasks, we use a task-specific prompt pool Pt for each task and a representation generator to synthesize past-task information, strengthening the relation classifier’s knowledge retention. and Mixture of Experts (MoE) models (Jacobs et al. 1991; Jordan and Jacobs 1994). The study demonstrates that self￾attention can be seen as embodying multiple MoE mo… view at source ↗
Figure 2
Figure 2. Data Flow Diagram: Initially, the task predictor predicts the task identity of the input x, enabling the selection of the corresponding prompt pool. Subsequently, the input x queries this prompt pool to identify prompts whose corresponding keys are closest to the query q(x). The chosen prompt is then prepended to the embedded input xe, creating the prompted input xp. The combined xp is fed into the BERT Encoder, whe… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 29 canonical work pages

  1. [1]

    , " * 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 mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    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 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Chaudhry, A.; Ranzato, M.; Rohrbach, M.; and Elhoseiny, M. 2019. Efficient Lifelong Learning with A-GEM. In International Conference on Learning Representations (ICLR)

    work page 2019

  4. [4]

    Cui, L.; Yang, D.; Yu, J.; Hu, C.; Cheng, J.; Yi, J.; and Xiao, Y. 2021. Refining Sample Embeddings with Relation Prototypes to Enhance Continual Relation Extraction. In Zong, C.; Xia, F.; Li, W.; and Navigli, R., eds., Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics, 232--243. Online

    work page 2021

  5. [5]

    Dao, V.; Pham, V.-C.; Tran, Q.; Le, T.-T.; Ngo, L.; and Nguyen, T. 2024. Lifelong Event Detection via Optimal Transport. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, 12610--12621

    work page 2024

  6. [6]

    Devlin, J.; Chang, M.-W.; Lee, K.; and Toutanova, K. 2019. BERT : Pre-training of Deep Bidirectional Transformers for Language Understanding. In Proceedings of the 2019 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies , 4171--4186

    2019

  7. [7]

    M.; Tong, S.; Lepikhin, D.; Xu, Y.; Krikun, M.; Zhou, Y.; Yu, A

    Du, N.; Huang, Y.; Dai, A. M.; Tong, S.; Lepikhin, D.; Xu, Y.; Krikun, M.; Zhou, Y.; Yu, A. W.; Firat, O.; Zoph, B.; Fedus, L.; Bosma, M. P.; Zhou, Z.; Wang, T.; Wang, Y. E.; Webster, K.; Pellat, M.; Robinson, K.; Meier - Hellstern, K. S.; Duke, T.; Dixon, L.; Zhang, K.; Le, Q. V.; Wu, Y.; Chen, Z.; and Cui, C. 2022. GLaM: Efficient Scaling of Language Mo...

    work page 2022

  8. [8]

    L.; Nguyen, T.; Van, L

    Hai, N. L.; Nguyen, T.; Van, L. N.; Nguyen, T. H.; and Than, K. 2024. Continual variational dropout: a view of auxiliary local variables in continual learning. Machine Learning, 113(1): 281--323

    work page 2024

Show all 40 references
  1. [9]

    Han, X.; Zhu, H.; Yu, P.; Wang, Z.; Yao, Y.; Liu, Z.; and Sun, M. 2018. F ew R el: A Large-Scale Supervised Few-Shot Relation Classification Dataset with State-of-the-Art Evaluation. In Riloff, E.; Chiang, D.; Hockenmaier, J.; and Tsujii, J., eds., Proceedings of the 2018 Conf...

    2018

  2. [10]

    A.; Jordan, M

    Jacobs, R. A.; Jordan, M. I.; Nowlan, S. J.; and Hinton, G. E. 1991. Adaptive mixtures of local experts. Neural Computation, 3

    1991

  3. [11]

    Ji, B.; Yu, J.; Li, S.; Ma, J.; Wu, Q.; Tan, Y.; and Liu, H. 2020. Span-based Joint Entity and Relation Extraction with Attention-based Span-specific and Contextual Semantic Representations. In Scott, D.; Bel, N.; and Zong, C., eds., Proceedings of the 28th International Confe...

    2020

  4. [12]

    I.; and Jacobs, R

    Jordan, M. I.; and Jacobs, R. A. 1994. Hierarchical mixtures of experts and the EM algorithm. Neural computation, 6(2): 181--214

    1994

  5. [13]

    Ke, Z.; and Liu, B. 2022. Continual learning of natural language processing tasks: A survey. arXiv preprint arXiv:2211.12701

    2022 arXiv

  6. [14]

    V.; and Ho, N

    Le, M.; Nguyen, A.; Nguyen, H.; Nguyen, T.; Pham, T.; Ngo, L. V.; and Ho, N. 2024 a . Mixture of Experts Meets Prompt-Based Continual Learning. In Advances in Neural Information Processing Systems

    2024

  7. [15]

    Le, M.; Nguyen, C.; Nguyen, H.; Tran, Q.; Le, T.; and Ho, N. 2024 b . Revisiting Prefix-tuning: Statistical Benefits of Reparameterization among Prompts. arXiv preprint arXiv:2410.02200

    2024 arXiv

  8. [16]

    Le, T.-T.; Dao, V.; Nguyen, L.; Nguyen, T.-N.; Ngo, L.; and Nguyen, T. 2024 c . SharpSeq: Empowering Continual Event Detection through Sharpness-Aware Sequential-task Learning. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computati...

    2024

  9. [17]

    T.; Van, L

    Le, T.-T.; Nguyen, M.; Nguyen, T. T.; Van, L. N.; and Nguyen, T. H. 2024 d . Continual relation extraction via sequential multi-task learning. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 18444--18452

    2024

  10. [18]

    Lester, B.; Al-Rfou, R.; and Constant, N. 2021. The Power of Scale for Parameter-Efficient Prompt Tuning. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, 3045--3059. Online and Punta Cana, Dominican Republic: Association for Computati...

    2021

  11. [19]

    L.; and Liang, P

    Li, X. L.; and Liang, P. 2021. Prefix-Tuning: Optimizing Continuous Prompts for Generation. arXiv:2101.00190

    2021 arXiv

  12. [20]

    Liu, X.; Ji, K.; Fu, Y.; Tam, W.; Du, Z.; Yang, Z.; and Tang, J. 2022. P -Tuning: Prompt Tuning Can Be Comparable to Fine-tuning Across Scales and Tasks. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), 61--68...

    2022

  13. [21]

    Ma, J.; Zhao, Z.; Yi, X.; Chen, J.; Hong, L.; and Chi, E. H. 2018. Modeling Task Relationships in Multi-task Learning with Multi-gate Mixture-of-Experts. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, KDD '18, 1930–1939. Ne...

    2018

  14. [22]

    Nguyen, H.; Nguyen, C.; Ngo, L.; Luu, A.; and Nguyen, T. 2023. A spectral viewpoint on continual relation extraction. In Findings of the Association for Computational Linguistics: EMNLP 2023, 9621--9629

    2023

  15. [23]

    P.; Nguyen, S.; Linh, N

    Phan, H.; Tuan, A. P.; Nguyen, S.; Linh, N. V.; and Than, K. 2022. Reducing catastrophic forgetting in neural networks via gaussian mixture approximation. In Pacific-Asia Conference on Knowledge Discovery and Data Mining, 106--117. Springer

    2022

  16. [24]

    S.; Keysers, D.; and Houlsby, N

    Riquelme, C.; Puigcerver, J.; Mustafa, B.; Neumann, M.; Jenatton, R.; Pinto, A. S.; Keysers, D.; and Houlsby, N. 2021. Scaling Vision with Sparse Mixture of Experts. In Ranzato, M.; Beygelzimer, A.; Dauphin, Y. N.; Liang, P.; and Vaughan, J. W., eds., Advances in Neural Inform...

    2021

  17. [25]

    Shazeer, N.; Mirhoseini, A.; Maziarz, K.; Davis, A.; Le, Q.; Hinton, G.; and Dean, J. 2017. Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer. In International Conference on Learning Representations (ICLR)

    2017

  18. [26]

    K.; Kim, J.; and Kim, J

    Shin, H.; Lee, J. K.; Kim, J.; and Kim, J. 2017. Continual learning with deep generative replay. Advances in neural information processing systems, 30

    2017

  19. [27]

    Tran, Q.; Le, M.; Truong, T.; Phung, D.; Ngo, L.; Nguyen, T.; Ho, N.; and Le, T. 2024 a . Leveraging Hierarchical Taxonomies in Prompt-based Continual Learning. arXiv preprint arXiv:2410.04327

    2024 arXiv

  20. [28]

    Tran, Q.; Thanh, N.; Anh, N.; Hai, N.; Le, T.; Ngo, L.; and Nguyen, T. 2024 b . Preserving Generalization of Language models in Few-shot Continual Relation Extraction. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, 13771--13784

    2024

  21. [29]

    N.; Hai, N

    Van, L. N.; Hai, N. L.; Pham, H.; and Than, K. 2022. Auxiliary local variables for improving regularization/prior approach in continual learning. In Pacific-Asia conference on knowledge discovery and data mining, 16--28. Springer

    2022

  22. [30]

    M.; Tuytelaars, T.; and Tolias, A

    van de Ven, G. M.; Tuytelaars, T.; and Tolias, A. S. 2022. Three types of incremental learning. Nature Machine Intelligence, 4: 1185--1197

    2022

  23. [31]

    Wang, H.; Xiong, W.; Yu, M.; Guo, X.; Chang, S.; and Wang, W. Y. 2019. Sentence Embedding Alignment for Lifelong Relation Extraction. In Burstein, J.; Doran, C.; and Solorio, T., eds., Proceedings of the 2019 Conference of the North A merican Chapter of the Association for Com...

    2019

  24. [32]

    Wang, L.; Xie, J.; Zhang, X.; Huang, M.; Su, H.; and Zhu, J. 2023 a . Hierarchical Decomposition of Prompt-Based Continual Learning: Rethinking Obscured Sub-optimality. Advances in Neural Information Processing Systems

    2023

  25. [33]

    Wang, Z.; Liu, Y.; Ji, T.; Wang, X.; Wu, Y.; Jiang, C.; Chao, Y.; Han, Z.; Wang, L.; Shao, X.; and Zeng, W. 2023 b . Rehearsal-free Continual Language Learning via Efficient Parameter Isolation. In Rogers, A.; Boyd - Graber, J. L.; and Okazaki, N., eds., Proceedings of the 61s...

    2023

  26. [34]

    Wang, Z.; Zhang, Z.; Ebrahimi, S.; Sun, R.; Zhang, H.; Lee, C.-Y.; Ren, X.; Su, G.; Perot, V.; Dy, J.; et al. 2022 a . DualPrompt: Complementary Prompting for Rehearsal-free Continual Learning. European Conference on Computer Vision

    2022

  27. [35]

    Wang, Z.; Zhang, Z.; Lee, C.-Y.; Zhang, H.; Sun, R.; Ren, X.; Su, G.; Perot, V.; Dy, J.; and Pfister, T. 2022 b . Learning to prompt for continual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 139--149

    2022

  28. [36]

    Xia, H.; Wang, P.; Liu, T.; Lin, B.; Cao, Y.; and Sui, Z. 2023. Enhancing Continual Relation Extraction via Classifier Decomposition. In Proceedings of the 61th Annual Meeting of the Association for Computational Linguistics)

    2023

  29. [37]

    Zhang, Y.; Zhong, V.; Chen, D.; Angeli, G.; and Manning, C. D. 2017. Position-aware Attention and Supervised Data Improve Slot Filling. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing (EMNLP 2017), 35--45

    2017

  30. [38]

    Zhao, K.; Xu, H.; Yang, J.; and Gao, K. 2022. Consistent Representation Learning for Continual Relation Extraction. In Muresan, S.; Nakov, P.; and Villavicencio, A., eds., Findings of the Association for Computational Linguistics: ACL 2022, 3402--3411. Dublin, Ireland: Associa...

    2022

  31. [39]

    Zhao, W.; Cui, Y.; and Hu, W. 2023. Improving Continual Relation Extraction by Distinguishing Analogous Semantics. In Proceedings of the 61th Annual Meeting of the Association for Computational Linguistics)

    2023

  32. [40]

    R.; Dai, A

    Zhou, Y.; Du, N.; Huang, Y.; Peng, D.; Lan, C.; Huang, D.; Shakeri, S.; So, D. R.; Dai, A. M.; Lu, Y.; Chen, Z.; Le, Q. V.; Cui, C.; Laudon, J.; and Dean, J. 2023. Brainformers: Trading Simplicity for Efficiency. In Krause, A.; Brunskill, E.; Cho, K.; Engelhardt, B.; Sabato, S...

    2023

Pith tools

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