Pith. sign in

REVIEW 4 major objections 5 minor 49 references

MEGA: Second-Order Gradient Alignment for Catastrophic Forgetting Mitigation in GFSCIL

T0 review · 4 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

Pith's one-line read Graph few-shot class-incremental learning models can avoid catastrophic forgetting if their meta-training stage simulates the incremental stage through an incremental second-order gradient and aligned loss functions, as the proposed MEGA…

desk verdict A genuinely stricter GFSCIL setting and a modular meta-training stage, but the no-query-leakage claim is compromised by unspecified transductive message passing, and the SOTA margins are less clean than the abstract suggests. read the letter →

arxiv 2504.13691 v3 pith:GIXYUKN7 submitted 2025-04-18 cs.LG

classification cs.LG
keywords graphfew-shotclass-incrementallearningcatastrophicforgettingmeta-learningsecond-ordergradientcontinualknowledgedistillationexperiencereplaynodeclassification
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

Graph few-shot class-incremental learning (GFSCIL) asks a model to keep recognizing old classes while learning new classes from only a handful of labeled nodes. This paper claims that existing GFSCIL methods leak information by fine-tuning on the new classes' query sets, and that this masks catastrophic forgetting. It proposes a stricter setting in which incremental training uses only the few support nodes, and a meta-learning framework called MEGA that computes an incremental second-order gradient over a sequence of pseudo-tasks drawn from the base classes. The claim is that this produces an initialization whose behavior matches the incremental stage, so the model adapts quickly to new classes without forgetting old ones. On four graph datasets, MEGA is reported to outperform adapted baselines and to improve existing graph continual-learning methods when they are wrapped in its meta-training stage.

What carries the argument

The central mechanism is the incremental second-order gradient, a variant of model-agnostic meta-learning's outer-loop update. Where standard meta-learning optimizes a single task's query loss, MEGA samples an ordered sequence of pseudo-tasks from the base label space, performs inner-loop adaptation on each, and computes the gradient of the cumulative sum of query losses across all tasks with respect to the initial parameters (Eq. 2). This injects second-order information about the whole task stream into the initialization. The second half of the machinery is behavioral alignment: the inner loop, outer loop, and incremental stage share the same loss components (cross-entropy, knowledge distillation, single-instance replay), so the update dynamics during meta-training match the dynamics during incremental fine-tuning. That alignment is what, according to the paper, activates the learned priors at test time.

What would settle it

Evaluate the incremental stage with message passing restricted to the support-induced subgraph only, so the model cannot see any unlabeled or test nodes before evaluation; if MEGA's large accuracy advantage over the baselines vanishes, the claim that the meta-learned prior alone drives forgetting mitigation is falsified.

Watch

Extended reading notes

Core claim

MEGA's core proposal is that catastrophic forgetting in GFSCIL can be mitigated by making the model's meta-training stage mimic the incremental-learning stage at the level of both the optimization trajectory and the loss functions. In the meta-training stage, the base class set $\mathbb{Y}_0$ is randomly split into an ordered sequence of pseudo-tasks; for each task the model performs $m$ inner-loop steps on the support set, then the outer loop computes the gradient of the sum of query losses over all tasks seen so far with respect to the original parameters (Eq. 2). This 'incremental second-order gradient' accumulates curvature information across the task stream, unlike single-task meta-learning. The same loss terms — cross-entropy, knowledge distillation, and a single-instance replay buffer — are used in the inner loop, outer loop, and incremental fine-tuning, so the model's behavior in all three phases is aligned. The paper claims this alignment is what turns the meta-learned initialization into a usable prior for sequential novel classes, and it presents experiments on Amazon-Clothing, DBLP, Cora-Full, and Reddit showing accuracy gains of 2.60% to 7.43% on the first novel task, along with improvements when existing continual-learning losses (FTF-ER, MAS, TWP) are embedded into the same framework.

Load-bearing premise

The base classes, randomly split into an ordered sequence of pseudo-tasks, must be a faithful stand-in for the future novel-task stream, because the meta-learned initialization is trusted to transfer from base-class gradients to genuinely new classes with different labels, features, and graph neighborhoods.

Editorial extensions

If this is right

  • Under the stricter support-only incremental protocol, MEGA is reported to keep higher accuracy on both base and novel classes than adapted GFSCIL baselines, GFSL methods, and GCL methods on four graph datasets.
  • The reported 2.60–7.43 percent improvement on the first novel task indicates that the benefit of the meta-learned prior appears immediately after the first incremental step, not only late in the stream.
  • Existing graph continual-learning methods (FTF-ER, MAS, TWP) show consistent accuracy gains when their loss functions are used inside MCTF, supporting the model-agnostic claim rather than a KDSIR-specific effect.
  • Ablations show that the meta-training framework alone raises the performance ceiling, while the knowledge-distillation and single-instance-replay terms add further forgetting resistance; removing any component degrades the final accuracy.
  • Performance stays roughly flat as the support set size $K$ varies from 1 to 7, suggesting the method does not depend on having several labeled nodes per new class.

Reading between the lines

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

  • Editorial inference: If behavioral alignment is truly the active ingredient, the same incremental second-order gradient recipe should transfer to non-graph modalities, such as image or text few-shot class-incremental learning, where base classes can be reorganized into pseudo-task sequences; this is a direct test the paper does not run.
  • Editorial inference: The paper does not state whether incremental training propagates messages through the full graph (transductive) or only through the support-induced subgraph. If full-graph propagation is used, part of the reported advantage may come from seeing unlabeled nodes during training; an inductive evaluation would separate graph-structure effects from the meta-learned prior.
  • Editorial inference: The stricter support-only protocol invalidates direct comparison with prior GFSCIL numbers; adopting it as a community standard would require re-benchmarking existing methods, and some previously reported improvements may shrink once query-set fine-tuning is removed.
  • Editorial inference: The mechanism is described for node classification; a natural extension is to test the same MCTF wrapper on link prediction or graph classification tasks, where the definition of a 'task' over a label space differs and the transfer assumption may or may not hold.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. This paper introduces MEGA, a model-agnostic meta-learning framework for graph few-shot class-incremental learning (GFSCIL). The authors first propose a stricter protocol in which incremental training uses only novel support sets, without fine-tuning on novel query sets, to remove query-set leakage. MEGA has two main components: MCTF, a meta-training algorithm that samples ordered base-task sequences and updates initial parameters via the gradient of cumulative query losses through inner-loop adaptation (Eq. 2), and KDSIR, a continual-learning loss module combining knowledge distillation with per-class single-instance replay. The paper evaluates MEGA against GFSL, GCL, and adapted GFSCIL baselines on Amazon-Clothing, DBLP, Cora-Full, and Reddit, reporting that MEGA outperforms all baselines and that MCTF improves existing GCL methods when embedded.

Significance. If the empirical results hold, the paper makes two useful contributions: it highlights the query-set leakage in prior GFSCIL protocols and proposes a model-agnostic way to import GCL losses into meta-learned initializations. The ablation studies (Tables IV-VII) are a genuine strength: they isolate the contributions of MCTF, L_SIR, and L_KD and show additive gains, and Table VII demonstrates that MCTF can improve three different GCL baselines. The setting itself is a sensible step toward stricter evaluation. However, the main state-of-the-art claim is currently not fully supported: the authors adapt the GFSCIL baselines themselves and report no code, several tabulated differences overlap with standard deviations, and the Reddit results lack error bars. The paper therefore needs additional verification before the headline claims can be accepted.

major comments (4)
  1. [Section III-A and Section IV-A] The manuscript's central claim of a support-only incremental protocol is not operationalized: no statement is made about whether the GNN forward pass during incremental fine-tuning uses the full graph or a support-induced subgraph. Because standard GNN layers aggregate over all neighbors, using the full adjacency leaks query-node features and topology into support representations even if query labels are not in the loss. Please specify the masking or sampling procedure; if the full graph is used, the distinction from HAG-Meta and GEOMETER reduces to withholding labels rather than excluding query sets, and the experimental comparisons need to be redone under a genuinely inductive protocol.
  2. [Section IV-A and Section IV-B] The authors adapt Geometer, HAG-Meta, and Mecoin to their new setting, acknowledge that this adaptation 'reveals a partial degradation in their catastrophic forgetting mitigation capabilities, thereby precluding equitable performance comparison,' and yet use these degraded versions in Table III to claim state-of-the-art performance. This is not a controlled comparison: the main SOTA claim is established against baselines that the authors themselves defined and weakened, with no released code to verify the adaptation. Report the original methods in their native settings as an upper reference, or restrict the SOTA claim to the unmodified GCL/GFSL baselines.
  3. [Table III and Table XI] Several of the reported improvements are within one standard deviation of the best baseline, and Table XI provides no error bars for Reddit despite the statement in Section IV-A that all accuracy scores are means over five random seeds. For example, Amazon Clothing Task 9 shows MEGA at 49.11±1.75% versus Mecoin at 48.82±1.96%, a 0.29% gap, and Cora-Full Task 1 shows overlapping intervals (52.74±3.89% vs 45.90±6.29%). Add per-task significance tests or paired-seed comparisons, and provide error bars for Reddit, before claiming that MEGA 'consistently outperforms all baselines.'
  4. [Abstract and Conclusion] The claimed 'improvement of 2.60% to 7.43% in the first novel task' does not match the tables: DBLP Task 1 shows +1.83% in Table III and Reddit Task 1 shows +11.55% in Table XI. Correct the reported range and specify the baseline used for each improvement.
minor comments (5)
  1. [Equation (4)] The SIR loss is written as CE(f_θ(x_b), f_θ(y_b)), which is not a valid cross-entropy expression; it should be CE(f_θ(x_b), y_b), with y_b as the label set. The same correction applies to Eq. (7).
  2. [Section IV-B vs Table III caption] The text states MEGA consistently outperforms all baselines, but GAT-FT's base accuracy is higher on DBLP (48.08±2.07 vs 46.99±0.94) and Cora-Full (72.64±0.57 vs 58.08±4.09), as the caption itself notes; please reconcile the wording.
  3. [Table II] The inner-loop step is set to 1, but the complexity analysis in Section III-E and the ablation in Table VIII do not examine the effect of the number of inner steps m on the second-order gradient quality; a short sensitivity study would clarify the role of m.
  4. [Section III-D] The role of the buffer B_m in Eq. (5) is described only as 'cleared at the end of each epoch'; please clarify how the single-instance replay buffer is populated during meta-training and why the same buffer is not used in the outer-loop objective.
  5. [Related work] FTF-ER [45] is the authors' own prior method and serves both as a baseline and as the basis for MCTF-enhanced comparisons in Table VII; a brief independent description of FTF-ER's mechanism would help readers assess potential overlap with KDSIR.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: MEGA's meta-gradient objective is explicitly constructed from MAML-style updates, and the empirical comparisons are self-contained against external datasets and baselines.

full rationale

The paper contains no derivation step in which a claimed prediction reduces by construction to its inputs. Equations (1)-(2) define MAML-style inner/outer updates over a task sequence; the outer gradient is taken with respect to the original parameters after unrolling inner updates, which is a standard second-order meta-learning objective using only base-task data and not fitted to incremental query accuracy. The loss-alignment design in Equations (5)-(7) is an algorithmic choice, and KDSIR's two components are knowledge distillation and single-instance replay, both cited to external prior works ([38], [39]); the versatility experiments plug in external GCL losses (FTF-ER, MAS, TWP). The FTF-ER citation [45] is a self-citation, but it functions as a comparison baseline and a plug-in module rather than as the load-bearing argument for MEGA's central claim, so it does not constitute circularity under the stated rules. No fitted constant is relabeled as a prediction; hyperparameters are either fixed or tuned per baseline in the standard way, and improvements are reported on held-out novel query sets. The paper's own admission that adapted GFSCIL baselines may not be equitably compared ('precluding equitable performance comparison') is an experimental-fairness caveat, not a logical circularity. Similarly, the unresolved question of whether the GNN forward pass uses full-graph message passing during incremental fine-tuning concerns the validity of the proposed support-only protocol, not a definitional equivalence. The empirical evaluation is therefore self-contained relative to the paper's stated claims.

Assumptions & free parameters 5 free parameters · 3 assumptions · 0 invented entities

The central empirical claim rests on hyperparameters (learning rates, inner steps, fine-tuning steps, sequence length) and on three domain assumptions: that base-task sequences proxy novel-class streams, that the meta-learned prior transfers, and that graph message passing can be used without label leakage. No new physical entities are postulated.

free parameters (5)
  • inner learning rate alpha = 0.005
    Chosen for SGD inner loop; controls adaptation strength in MCTF (Algorithm 1).
  • outer learning rate beta = 0.005
    Adam outer update; controls meta-learned initialization speed.
  • inner loop steps m = 1
    Only one inner gradient step per task in all experiments (Table II), reducing the effective second-order signal; larger m not tested.
  • incremental fine-tuning steps = 5
    Number of gradient steps on novel support sets; baselines tuned between 1 and 10, MEGA fixed at 5.
  • task sequence length p = not reported
    Length of the sampled base task sequence in MCTF; not listed in Table II, so the number of simulated tasks per meta-epoch is unspecified.
assumptions (3)
  • standard math Backpropagation through the inner gradient updates is well-defined and produces usable second-order gradients.
    MCTF outer update (Eq. 2) requires differentiating through Eq. (1); this is standard MAML.
  • domain assumption Random partitions of the base classes form a representative task distribution for novel incremental classes.
    Algorithm 1 samples Tbase from Y0; the transfer of the learned prior to unseen classes is assumed, not tested across domains.
  • domain assumption GNN message passing during incremental training may use graph structure without leaking label information.
    Losses in Eqs. (5)-(7) do not restrict the message-passing neighborhood; using full-graph structure is transductive and could make the task easier, but this is not discussed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MEGA: Second-Order Gradient Alignment for Catastrophic Forgetting Mitigation in GFSCIL." pith.science (2026). https://pith.science/paper/GIXYUKN7

@misc{pith2026250413691,
  author       = {Pith},
  title        = {Pith review of: MEGA: Second-Order Gradient Alignment for Catastrophic Forgetting Mitigation in GFSCIL},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GIXYUKN7}},
  note         = {Machine review of arXiv:2504.13691}
}
read the original abstract

Graph Few-Shot Class-Incremental Learning (GFSCIL) enables models to continually learn from limited samples of novel tasks after initial training on a large base dataset. Existing GFSCIL approaches typically utilize Prototypical Networks (PNs) for metric-based class representations and fine-tune the model during the incremental learning stage. However, these PN-based methods oversimplify learning via novel query set fine-tuning and fail to integrate Graph Continual Learning (GCL) techniques due to architectural constraints. To address these challenges, we propose a more rigorous and practical setting for GFSCIL that excludes query sets during the incremental training phase. Building on this foundation, we introduce Model-Agnostic Meta Graph Continual Learning (MEGA), aimed at effectively alleviating catastrophic forgetting for GFSCIL. Specifically, by calculating the incremental second-order gradient during the meta-training stage, we endow the model to learn high-quality priors that enhance incremental learning by aligning its behaviors across both the meta-training and incremental learning stages. Extensive experiments on four mainstream graph datasets demonstrate that MEGA achieves state-of-the-art results and enhances the effectiveness of various GCL methods in GFSCIL. We believe that our proposed MEGA serves as a model-agnostic GFSCIL paradigm, paving the way for future research.

Figures

Figures reproduced from arXiv: 2504.13691 by the authors.

Figure 1
Figure 1. Graph few-shot class-incremental learning problem (N-way K￾shot R-query). In our proposed setting, each novel task is divided into a support set and a query set, both containing N classes. Each class within these sets comprises K sample nodes in the support set and R sample nodes in the query set, respectively. these stages. Specifically, in the meta-training stage, we intro￾duce an incremental second-order gradient… view at source ↗
Figure 2
Figure 2. Comparison of dataset partitioning strategies and task complexity in [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Overall architecture of MEGA. In the meta-training stage, we train a model by calculating the incremental second-order gradient. Then, in the incremental learning stage, the model is fine-tuned using the novel support set. By aligning the model’s behavior across both stages, we enable it to acquire prior knowledge about optimal initial parameters across a task stream. we propose the MEGA method, driven by its core c… view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Evolution of the accuracy throughout the learning process on the task sequences of three datasets. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Evolution of the accuracy throughout the learning process on the task [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Evolution of the accuracy throughout the learning process on the task [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 8
Figure 8. Figure 8: Comparison of MCTF-enhanced GCL methods against original GCL [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 7
Figure 7. Figure 7: Evolution of the accuracy throughout the learning process on the task [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 10
Figure 10. Figure 10: The influence of support set size K on Amazon Clothing and DBLP datasets. TABLE IX EXPERIMENTAL RESULTS WITH DIFFERENT SHOT SETTINGS AND THEIR CORRESPONDING RANGE VALUES ON THE AMAZON CLOTHING DATASET (3-WAY). THE RANGE, DEFINED AS THE DIFFERENCE BETWEEN MAXIMUM AND M…
Figure 9
Figure 9. Figure 9: Comparison of various meta-learning methods on DBLP dataset. [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]
Figure 11
Figure 11. Figure 11: 2-D t-SNE projections of embeddings in two models on Amazon [PITH_FULL_IMAGE:figures/full_fig_p011_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 41 canonical work pages

  1. [1]

    The graph neural network model,

    F. Scarselli, M. Gori, A. C. Tsoi, M. Hagenbuchner, and G. Monfardini, “The graph neural network model,” IEEE Transactions on Neural Networks, pp. 61–80, 2008

  2. [2]

    Semi-supervised classification with graph convolutional networks,

    T. N. Kipf and M. Welling, “Semi-supervised classification with graph convolutional networks,” arXiv preprint arXiv:1609.02907 , 2016

  3. [3]

    Inductive representation learning on large graphs,

    W. Hamilton, Z. Ying, and J. Leskovec, “Inductive representation learning on large graphs,” Advances in Neural Information Processing Systems, 2017

  4. [4]

    Graph neural networks with continual learning for fake news detection from social media,

    Y . Han, S. Karunasekera, and C. Leckie, “Graph neural networks with continual learning for fake news detection from social media,” arXiv preprint arXiv:2007.03316, 2020

  5. [5]

    Graduate employment prediction with bias,

    T. Guo, F. Xia, S. Zhen, X. Bai, D. Zhang, Z. Liu, and J. Tang, “Graduate employment prediction with bias,” inProceedings of the AAAI Conference on Artificial Intelligence , 2020, pp. 670–677

  6. [6]

    Predicting multicellular function through multi-layer tissue networks,

    M. Zitnik and J. Leskovec, “Predicting multicellular function through multi-layer tissue networks,” Bioinformatics, pp. i190–i198, 2017

  7. [7]

    Protein interface prediction using graph convolutional networks,

    A. Fout, J. Byrd, B. Shariat, and A. Ben-Hur, “Protein interface prediction using graph convolutional networks,” Advances in Neural Information Processing Systems , 2017

  8. [8]

    Meta- gnn: On few-shot node classification in graph meta-learning,

    F. Zhou, C. Cao, K. Zhang, G. Trajcevski, T. Zhong, and J. Geng, “Meta- gnn: On few-shot node classification in graph meta-learning,” in ACM International Conference on Information and Knowledge Management , 2019, pp. 2357–2360

Show all 49 references
  1. [9]

    Graph proto- typical networks for few-shot learning on attributed networks,

    K. Ding, J. Wang, J. Li, K. Shu, C. Liu, and H. Liu, “Graph proto- typical networks for few-shot learning on attributed networks,” in ACM International Conference on Information and Knowledge Management , 2020, pp. 295–304. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 12

  2. [10]

    Meta-learning with graph neural networks: Methods and applications,

    D. Mandal, S. Medya, B. Uzzi, and C. Aggarwal, “Meta-learning with graph neural networks: Methods and applications,” ACM SIGKDD Explorations Newsletter, pp. 13–22, 2022

  3. [11]

    Meta-inductive node classification across graphs,

    Z. Wen, Y . Fang, and Z. Liu, “Meta-inductive node classification across graphs,” in Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval , 2021, pp. 1219–1228

  4. [12]

    Graph lifelong learning: A survey,

    F. G. Febrinanto, F. Xia, K. Moore, C. Thapa, and C. Aggarwal, “Graph lifelong learning: A survey,” IEEE Computational Intelligence Magazine, pp. 32–51, 2023

  5. [13]

    Self-supervised continual graph learning in adaptive riemannian spaces,

    L. Sun, J. Ye, H. Peng, F. Wang, and S. Y . Philip, “Self-supervised continual graph learning in adaptive riemannian spaces,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2023, pp. 4633–4642

  6. [14]

    Cat: Balanced continual graph learning with graph condensation,

    Y . Liu, R. Qiu, and Z. Huang, “Cat: Balanced continual graph learning with graph condensation,” in IEEE International Conference on Data Mining, 2023, pp. 1157–1162

  7. [15]

    Continual learning on dynamic graphs via parameter isolation,

    P. Zhang, Y . Yan, C. Li, S. Wang, X. Xie, G. Song, and S. Kim, “Continual learning on dynamic graphs via parameter isolation,” in Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval , 2023, pp. 601– 611

  8. [16]

    Graph few-shot class-incremental learning,

    Z. Tan, K. Ding, R. Guo, and H. Liu, “Graph few-shot class-incremental learning,” in ACM International Conference on Web Search and Data Mining, 2022, pp. 987–996

  9. [17]

    Geometer: Graph few-shot class-incremental learning via prototype representation,

    B. Lu, X. Gan, L. Yang, W. Zhang, L. Fu, and X. Wang, “Geometer: Graph few-shot class-incremental learning via prototype representation,” in ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2022, pp. 1152–1161

  10. [18]

    Prototypical networks for few-shot learning,

    J. Snell, K. Swersky, and R. Zemel, “Prototypical networks for few-shot learning,” Advances in Neural Information Processing Systems , 2017

  11. [19]

    Cglb: Benchmark tasks for continual graph learning,

    X. Zhang, D. Song, and D. Tao, “Cglb: Benchmark tasks for continual graph learning,” Advances in Neural Information Processing Systems , pp. 13 006–13 021, 2022

  12. [20]

    Sparsified subgraph memory for continual graph representation learning,

    ——, “Sparsified subgraph memory for continual graph representation learning,” in IEEE International Conference on Data Mining , 2022, pp. 1335–1340

  13. [21]

    Learning without forgetting,

    Z. Li and D. Hoiem, “Learning without forgetting,” IEEE Transactions on Pattern Analysis and Machine Intelligence , pp. 2935–2947, 2017

  14. [22]

    Incremental few-shot learning via vector quantization in deep embedded space,

    K. Chen and C.-G. Lee, “Incremental few-shot learning via vector quantization in deep embedded space,” in International Conference on Learning Representations, 2021

  15. [23]

    For- ward compatible few-shot class-incremental learning,

    D.-W. Zhou, F.-Y . Wang, H.-J. Ye, L. Ma, S. Pu, and D.-C. Zhan, “For- ward compatible few-shot class-incremental learning,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 9046–9056

  16. [24]

    Semantics- driven generative replay for few-shot class incremental learning,

    A. Agarwal, B. Banerjee, F. Cuzzolin, and S. Chaudhuri, “Semantics- driven generative replay for few-shot class incremental learning,” in Proceedings of the 30th ACM International Conference on Multimedia , 2022, pp. 5246–5254

  17. [25]

    On the soft-subnetwork for few-shot class incremental learning,

    H. Kang, J. Yoon, S. R. H. Madjid, S. J. Hwang, and C. D. Yoo, “On the soft-subnetwork for few-shot class incremental learning,” in International Conference on Learning Representations , 2023

  18. [26]

    A survey on few- shot class-incremental learning,

    S. Tian, L. Li, W. Li, H. Ran, X. Ning, and P. Tiwari, “A survey on few- shot class-incremental learning,” Neural Networks, pp. 307–324, 2024

  19. [27]

    Few-shot class-incremental learning,

    X. Tao, X. Hong, X. Chang, S. Dong, X. Wei, and Y . Gong, “Few-shot class-incremental learning,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2020, pp. 12 180–12 189

  20. [28]

    S3c: Self-supervised stochastic classifiers for few-shot class-incremental learning,

    J. Kalla and S. Biswas, “S3c: Self-supervised stochastic classifiers for few-shot class-incremental learning,” in Proceedings of the European Conference on Computer Vision , 2022, pp. 432–448

  21. [29]

    Few-shot learning on graphs: from meta-learning to pre-training and prompting,

    X. Yu, Y . Fang, Z. Liu, Y . Wu, Z. Wen, J. Bo, X. Zhang, and S. C. Hoi, “Few-shot learning on graphs: from meta-learning to pre-training and prompting,” arXiv preprint arXiv:2402.01440 , 2024

  22. [30]

    Advances and challenges in meta-learning: A technical re- view,

    A. Vettoruzzo, M.-R. Bouguelia, J. Vanschoren, T. Rognvaldsson, and K. Santosh, “Advances and challenges in meta-learning: A technical re- view,”IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024

  23. [31]

    Meta-gps++: Enhancing graph meta-learning with contrastive learning and self-training,

    Y . Liu, M. Li, X. Li, L. Huang, F. Giunchiglia, Y . Liang, X. Feng, and R. Guan, “Meta-gps++: Enhancing graph meta-learning with contrastive learning and self-training,” ACM Transactions on Knowledge Discovery from Data, pp. 1–30, 2024

  24. [32]

    Relative and absolute location embedding for few-shot node classification on graph,

    Z. Liu, Y . Fang, C. Liu, and S. C. Hoi, “Relative and absolute location embedding for few-shot node classification on graph,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2021, pp. 4267–4275

  25. [33]

    Graph few-shot learning via knowledge transfer,

    H. Yao, C. Zhang, Y . Wei, M. Jiang, S. Wang, J. Huang, N. Chawla, and Z. Li, “Graph few-shot learning via knowledge transfer,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2020, pp. 6656–6663

  26. [34]

    Adaptive attentional network for few-shot knowledge graph completion,

    J. Sheng, S. Guo, Z. Chen, J. Yue, L. Wang, T. Liu, and H. Xu, “Adaptive attentional network for few-shot knowledge graph completion,” in Pro- ceedings of the Conference on Empirical Methods in Natural Language Processing, 2020, pp. 1681–1691

  27. [35]

    Model-agnostic meta-learning for fast adaptation of deep networks,

    C. Finn, P. Abbeel, and S. Levine, “Model-agnostic meta-learning for fast adaptation of deep networks,” in International Conference on Machine Learning, 2017, pp. 1126–1135

  28. [36]

    Graph meta learning via local subgraphs,

    K. Huang and M. Zitnik, “Graph meta learning via local subgraphs,” Advances in Neural Information Processing Systems , pp. 5862–5874, 2020

  29. [37]

    Self-supervised auxiliary learning for graph neural networks via meta- learning,

    D. Hwang, J. Park, S. Kwon, K.-M. Kim, J.-W. Ha, and H. J. Kim, “Self-supervised auxiliary learning for graph neural networks via meta- learning,” arXiv preprint arXiv:2103.00771 , 2021

  30. [38]

    Lifelong learning via progressive distillation and retrospection,

    S. Hou, X. Pan, C. C. Loy, Z. Wang, and D. Lin, “Lifelong learning via progressive distillation and retrospection,” in Proceedings of the European Conference on Computer Vision , 2018, pp. 437–452

  31. [39]

    End-to-end incremental learning,

    F. M. Castro, M. J. Mar ´ın-Jim´enez, N. Guil, C. Schmid, and K. Alahari, “End-to-end incremental learning,” in Proceedings of the European conference on computer vision , 2018, pp. 233–248

  32. [40]

    Inferring networks of substi- tutable and complementary products,

    J. McAuley, R. Pandey, and J. Leskovec, “Inferring networks of substi- tutable and complementary products,” in ACM SIGKDD Conference on Knowledge Discovery and Data Mining , 2015, pp. 785–794

  33. [41]

    Arnetminer: extraction and mining of academic social networks,

    J. Tang, J. Zhang, L. Yao, J. Li, L. Zhang, and Z. Su, “Arnetminer: extraction and mining of academic social networks,” in ACM SIGKDD Conference on Knowledge Discovery and Data Mining , 2008, pp. 990– 998

  34. [42]

    Deep gaussian embedding of graphs: Unsupervised inductive learning via ranking,

    A. Bojchevski and S. G ¨unnemann, “Deep gaussian embedding of graphs: Unsupervised inductive learning via ranking,” arXiv preprint arXiv:1707.03815, 2017

  35. [43]

    Graph attention networks,

    P. Veli ˇckovi´c, G. Cucurull, A. Casanova, A. Romero, P. Lio, and Y . Ben- gio, “Graph attention networks,” arXiv preprint arXiv:1710.10903, 2017

  36. [44]

    Rich feature hierarchies for accurate object detection and semantic segmentation,

    R. Girshick, J. Donahue, T. Darrell, and J. Malik, “Rich feature hierarchies for accurate object detection and semantic segmentation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2014, pp. 580–587

  37. [45]

    Ftf-er: Feature-topology fusion-based experience replay method for continual graph learning,

    J. Pang, C. Lin, X. Hao, R. Yin, Z. Wang, Z. Zhang, J. He, and H. Tai Sheng, “Ftf-er: Feature-topology fusion-based experience replay method for continual graph learning,” in Proceedings of the 32nd ACM International Conference on Multimedia , 2024, pp. 8336–8344

  38. [46]

    Memory aware synapses: Learning what (not) to forget,

    R. Aljundi, F. Babiloni, M. Elhoseiny, M. Rohrbach, and T. Tuytelaars, “Memory aware synapses: Learning what (not) to forget,” inProceedings of the European Conference on Computer Vision , 2018, pp. 139–154

  39. [47]

    Overcoming catastrophic forgetting in graph neural networks,

    H. Liu, Y . Yang, and X. Wang, “Overcoming catastrophic forgetting in graph neural networks,” in Proceedings of the AAAI Conference on Artificial Intelligence, 2021, pp. 8653–8661

  40. [48]

    An efficient memory module for graph few-shot class-incremental learning,

    D. Li, A. Zhang, J. Gao, and B. Qi, “An efficient memory module for graph few-shot class-incremental learning,” Advances in Neural Information Processing Systems , 2024

  41. [49]

    Visualizing data using t-sne

    L. Van der Maaten and G. Hinton, “Visualizing data using t-sne.” Journal of Machine Learning Research , 2008

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.