Task-Specific Preconditioner for Cross-Domain Few-Shot Learning

Jungwon Park; Suhyun Kang; Wonjong Rhee; Wonseok Lee

REVIEW 3 major objections 5 minor 67 references

Task-Specific Preconditioner for Cross-Domain Few-Shot Learning

T0 review · 3 major / 5 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 task-specific preconditioner, built by combining meta-learned per-domain preconditioners, achieves state-of-the-art few-shot accuracy on Meta-Dataset across all 13 test domains.

desk verdict Solid empirical CDFSL paper whose central 'task-specific' claim needs an ablation; still deserves peer review. read the letter →

arxiv 2412.15483 v1 pith:LUTIMD3H submitted 2024-12-20 cs.LG cs.AIcs.CV

Suhyun Kang , Jungwon Park , Wonseok Lee , Wonjong Rhee This is my paper · ORCID

classification cs.LGcs.AIcs.CV

keywords cross-domainfew-shotlearningpreconditionedgradientdescentmeta-learningpositivedefinitematrixMeta-Datasettask-specificadaptationdatasetclassifier

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 proposes an adaptation mechanism, Task-Specific Preconditioned gradient descent (TSP), that replaces the fixed optimizer used to adapt task-specific parameters in cross-domain few-shot learning. TSP meta-learns one positive-definite preconditioner per meta-training domain, then at test time forms a task-specific preconditioner as a softmax-weighted combination, with weights from a dataset classifier that reads the support set. The authors show that with positive definiteness enforced, preconditioned gradient descent consistently converges on seen and unseen domains, and that applying TSP to existing methods TSA and TA2-Net gives the best accuracy on Meta-Dataset in multi-domain, single-domain, and low-shot settings. If correct, the work shows that adapting the optimization geometry itself, rather than just the parameters, is an effective route to cross-domain generalization.

What carries the argument

The central object is the Task-Specific Preconditioner $P_T^\ell = \sum_{k=1}^K p_{T,k} P_k^\ell$, a convex combination of Domain-Specific Preconditioners $P_k^\ell = M_k^{\ell T}M_k^\ell + I$, where $M_k^\ell$ are meta-learned matrices and $p_{T,k}$ are softmax coefficients produced by a dataset classifier trained with an auxiliary task loss in addition to cross-entropy. Because each $P_k^\ell$ is symmetric positive definite and the coefficients are nonnegative and sum to one, the combination is positive definite and hence a valid Riemannian metric; the meta-parameters are trained by bi-level optimization, using the domain label to select the DSP in the inner loop. This gives gradient descent an adaptive geometry that varies with the target task.

What would settle it

Take a trained TSP and run Meta-Dataset test tasks while replacing the dataset classifier's coefficients with random permutations of the same values. If accuracy remains at the reported level, the coefficient mechanism is not the cause of the gains; more directly, train the classifier on a subset of domains and test on a held-out domain deliberately chosen to be far from all training domains (e.g., a medical or satellite image domain absent from Meta-Dataset); a sharp drop in performance would indicate the method relies on seen-domain coverage rather than a generalizable preconditioning principle.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that meta-learning a set of domain-specific preconditioners—each of the form $P_k^\ell = M_k^{\ell T}M_k^\ell + I$—and combining them with task coefficients $p_{T,k}$ from a dataset classifier produces a preconditioned gradient descent that adapts better to unseen domains than the fixed optimization strategies used by prior CDFSL methods. The central empirical claim is that TSP applied to TA2-Net achieves the best accuracy on all 13 Meta-Dataset test sets in the multi-domain setting, with an average rank of 1.0, and that it also leads on 12 of 13 datasets in the single-domain setting. The paper further argues that the positive definiteness of the Task-Specific Preconditioner, guaranteed by construction and proven in Theorem 1, is what makes this reliable: without the PD constraint, preconditioners lose definiteness during meta-training and fail to converge on unseen domains.

Load-bearing premise

At test time, a task from an unseen domain is assigned coefficients by a dataset classifier trained only on the eight seen domains, and the method assumes that combining the seen-domain preconditioners with those coefficients improves optimization on the unseen domain.

Editorial extensions

If this is right

Applying TSP to TA2-Net yields the best mean accuracy on all 13 Meta-Dataset test domains in the multi-domain setting, with an average rank of 1.0 versus 3.2 for the next-best method.
TSP improves unseen-domain accuracy in particular: in the single-domain setting, TSP on TA2-Net raises average unseen accuracy from 75.0 to 76.9, and in Five-Way One-Shot from 61.3 to 64.1.
The positive definiteness constraint is load-bearing: without it, averaged non-PD rates across DSPs reach 0.29 and performance on unseen domains drops from 79.8 to 73.8.
TSP is inference-efficient compared to the prior PGD method GAP, completing a task in about 1.1 seconds versus 14.2 seconds.
The identity term in the DSP design behaves as a regularizer toward plain gradient descent; it helps when shots are plentiful and hurts slightly in the five-shot regime.

Reading between the lines

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

If the dataset classifier's softmax coefficients truly capture similarity to seen domains, TSP could extend to any new domain without retraining, making the approach a drop-in optimizer for open-set few-shot benchmarks.
Nothing in the method ties the preconditioner to a specific backbone or task-specific parameter type; the same recipe could be applied to other fine-tuning procedures, such as full-network adaptation or prompt tuning, whenever a few-shot task arrives.
The auxiliary loss on the dataset classifier, which couples coefficient learning to query accuracy, is the component most responsible for the benefit; a testable prediction is that removing it (λ=0) should hurt unseen domains more than seen ones, as the ablation already hints (78.8 vs 79.8 average unseen).
Because the coefficients are probabilistic, the variance of the preconditioner across tasks from the same domain could be used as a measure of domain ambiguity, potentially informing when the model should fall back on plain gradient descent.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Colleague,

Here's my read of Kang et al.'s TSP paper. The headline: this is a solid, reproducible-looking empirical contribution to cross-domain few-shot learning. The main thing it does well is show that per-domain preconditioners combined by a dataset classifier improve two existing SOTA methods (TSA, TA2-Net) on Meta-Dataset across several settings. The gains are consistent, the average ranks are strong, and the evaluation is a genuine held-out test. The PD-constraint story is well motivated and backed by an ablation showing that dropping it hurts, especially on unseen domains.

What's actually new is modest but real: using domain-specific preconditioners plus softmax coefficients from a dataset classifier is a natural extension of prior PGD work like GAP, and the authors demonstrate it on a large benchmark. The theory (Theorem 1) is elementary—a convex combination of PD matrices is PD—and the paper does not oversell it, though the "steepest descent" language is a bit loose. The auxiliary loss on the classifier (lambda ablation in the appendix) is a nice touch.

The soft spot is the one the stress-test flags: no ablation separates the contribution of the learned task coefficients from simpler alternatives. Uniform mixing (p_k = 1/K), a random simplex vector, or a per-domain oracle would tell you whether the classifier's softmax weighting is actually doing the work, or whether the gains come from having K per-domain preconditioners (extra capacity) or an averaging effect. This matters because the method's name and framing rest on the coefficients being task-specific. Appendix D.2 shows consistent coefficient patterns but no quantitative comparison to a control. I would want to see that before accepting the central interpretability claim. It is a fixable gap, not a fatal flaw: the empirical improvements over TSA and TA2-Net are still there even if uniform mixing matches them, but the "task-specific" story would weaken considerably.

Minor issues: no code released, some per-dataset differences overlap with 95% CIs (though the average-rank evidence is convincing), and the paper could be clearer about the classifier being trained on seen domains only and extrapolating to unseen ones.

Bottom line: this deserves a serious referee. The empirical claims are strong enough to warrant publication, but the missing ablation should be addressed. I would bring it to a reading group as a case study in how to (and how not to) validate a claimed mechanism.

My recommendation: send it to review, with a request for the uniform-mixing control.

Referee Report

3 major / 5 minor

Summary. The paper presents TSP (Task-Specific Preconditioned gradient descent), an adaptation mechanism for cross-domain few-shot learning. TSP meta-learns a positive-definite Domain-Specific Preconditioner (DSP) for each of K meta-training domains. At meta-test time, a dataset classifier maps the support set to a softmax vector, and the task-specific preconditioner is the weighted sum of DSPs (Eq. (15)). This preconditioner is used in the inner-loop gradient updates of the task-specific parameters. The method is instantiated on top of TSA and TA2-Net and evaluated on Meta-Dataset in four settings (multi-domain, single-domain, varying-way five-shot, five-way one-shot). The paper reports state-of-the-art average accuracy and average rank in most settings, with ablations showing the importance of the PD constraint and the DSP design.

Significance. If the central mechanism is validated, TSP is a conceptually simple and effective addition to CDFSL: it adapts the optimization geometry per task using a convex combination of domain-specific preconditioners. The paper ships a correct proof of positive definiteness (Theorem 1) and a thorough empirical evaluation with confidence intervals on a standard benchmark. The reported average-rank improvements are consistent across settings. However, the experiments do not isolate the contribution of the learned task-coefficients from uniform mixing, added per-domain capacity, or a single global preconditioner, so the specific claim that the dataset classifier forms an effective task-specific preconditioner is not yet load-bearing. The missing control experiments are straightforward and should be added.

major comments (3)

[§4.3, Eq. (15); §5.3; Appendix D.2] The paper never tests whether the learned task-coefficients from the dataset classifier are responsible for the reported gains. Because Theorem 1 guarantees positive definiteness for any coefficient vector in the simplex, the PD property does not validate the coefficients. A control experiment comparing Eq. (15) with the learned coefficients against (i) uniform coefficients p_k = 1/K, (ii) random simplex vectors, and (iii) oracle one-hot selection on seen domains is required. Without such an ablation, the SOTA results in Tables 1 and 2 could equally be explained by the extra expressivity of K per-domain preconditioners, a beneficial averaging effect, or a task-dependent learning-rate rescaling. The qualitative coefficient visualizations in Appendix D.2 do not substitute for this quantitative comparison.
[§5.2, Table 1; §5.3] The paper does not compare against a single global preconditioner trained with the same base method and the same bi-level objective. TSP applied to TSA or TA2-Net has K preconditioners, whereas TSA/TA2-Net uses ordinary gradient descent; a K=1 variant (one shared preconditioner) under the same update rule would isolate the benefit of domain-specificity. Such a control is necessary to rule out the possibility that the observed improvements come mainly from added parameters or from preconditioning per se rather than from the domain-specific combination.
[§4.2; §6; Appendix D.2] The paper assumes that softmax coefficients from a dataset classifier trained on the eight seen domains produce useful preconditioners for the unseen test domains. This premise is never validated: the classifier's accuracy or calibration on unseen domains is not reported, and no analysis shows how performance degrades when the classifier is wrong. The heatmaps in Appendix D.2 show consistent patterns within domains but provide no quantitative evidence that these patterns outperform simpler alternatives on unseen domains. The authors should report the dataset classifier's prediction quality on unseen domains and include a control with coefficients fixed to the uniform distribution on the unseen test split.

minor comments (5)

[Figure 1 caption] The word 'baed' should be 'based'.
[Algorithm 2, line 8] The notation d_{T,k} is ambiguous; the text should state that d_T is a one-hot encoding of the integer domain label d_T.
[§4.2] The single-domain setting uses a sigmoid because the dataset classifier output dimension is one; please clarify how the DSPs and task-coefficients are defined when K=1, since the standard Meta-Dataset single-domain protocol has only one meta-training domain.
[Table 2] Several entries for ALFA+Proto-MAML and GAP+Proto-MAML are reported as '-' (MNIST, CIFAR-10, CIFAR-100); please add a footnote explaining the missing values or complete the table.
[General] The paper does not state whether the code is publicly available; a reproducibility statement would be helpful.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: TSP's construction, PD proof, and held-out Meta-Dataset evaluation are independent; self-citations to GAP are not load-bearing.

full rationale

TSP's derivation is self-contained. The method defines each Domain-Specific Preconditioner as P_k = M_k^T M_k + I (Eq. 8), meta-learns the M_k through bi-level optimization (Eqs. 9-10), obtains task-coefficients from a separately trained dataset classifier (Eqs. 11-14), and forms the Task-Specific Preconditioner as a convex combination (Eq. 15). Theorem 1's positive-definiteness guarantee is proved directly from Lemma 1 using elementary algebra and does not assume the claimed result. The SOTA claim is supported by held-out Meta-Dataset tasks, so the benchmark accuracy is an output of the method, not an input that is fitted and renamed as a prediction. Self-citations to GAP (Kang et al. 2023) appear in related work and as background motivation for the positive-definite constraint, but the paper's own Figure 1b and Table 4 establish the empirical motivation, so the self-citation is not the load-bearing evidence. The main validation gap identified by the reader's take—the absence of an ablation comparing the learned task-coefficients against uniform mixing or other simple alternatives, which is not present in Section 5.3 or Appendix D.2—is an empirical-support concern rather than a circularity: no equation in the paper reduces the reported gains to a fitted parameter, to the benchmark target, or to a self-citation chain. Hence no circular step is present.

Assumptions & free parameters 10 free parameters · 4 assumptions · 0 invented entities

The method introduces no new physical or conceptual entities beyond standard learned model parameters. The free parameters are hyperparameters common in meta-learning pipelines; the axioms capture the main empirical assumptions: per-domain sufficiency, classifier reliability on unseen domains, and the practical benefit of the PD constraint.

free parameters (10)

lambda (dataset classifier loss weight) = 0.1
Tuned on Meta-Dataset validation; Table 7 in Appendix D.1 shows 0.1 is best among {10, 1, 0.1, 0.01}.
alpha_in (inner learning rate) = 0.1
Reported in Table 8; used for all inner steps during meta-training.
alpha_out (outer learning rate) = 0.1
Reported in Table 8.
beta_test seen domains = (0.05, 0.30) or (0.05, 0.20)
Per-adapter and pre-classifier test-time learning rates for seen domains; Table 8.
beta_test unseen domains = (0.25, 0.05)
Per-adapter and pre-classifier test-time learning rates for unseen domains; Table 8.
M_initialization_scale = 0.1 * I
Initial scale for meta-parameters M_k; chosen by hand, Table 8.
num_inner_steps_training = 5
Number of inner gradient steps during meta-training (Table 8).
num_inner_steps_testing = 40
Number of inner steps at meta-testing (Table 8).
weight_decay = 0.0007
Used for both DSP and dataset classifier training; Table 8 and Table 9.
batch_size = 16
Used for all training; Table 8 and Table 9.

assumptions (4)

domain assumption For each seen domain, a single positive definite DSP P_k = M_k^T M_k + I captures the geometric characteristics of the parameter space for tasks from that domain.
Core representational assumption of Section 4.1, Eq. (8). No proof or analysis that a single matrix per domain is sufficient.
domain assumption The dataset classifier's softmax outputs, trained on seen domains, produce good combination weights for unseen domains.
Section 4.2 and Eq. (15): at meta-testing, tasks from unseen domains rely on coefficients from a classifier that never saw those domain labels.
domain assumption Positive definiteness of the preconditioner is sufficient for reliable and beneficial adaptation in CDFSL.
Section 4.4: based on prior Riemannian geometry results (Amari et al.), not proven for this setting.
standard math A linear combination of DSPs with convex coefficients remains a valid and effective preconditioner.
Theorem 1 proves positive definiteness of the combination, but does not prove optimality or efficacy.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Task-Specific Preconditioner for Cross-Domain Few-Shot Learning." pith.science (2026). https://pith.science/paper/LUTIMD3H

@misc{pith2026241215483,
  author       = {Pith},
  title        = {Pith review of: Task-Specific Preconditioner for Cross-Domain Few-Shot Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LUTIMD3H}},
  note         = {Machine review of arXiv:2412.15483}
}

read the original abstract

Cross-Domain Few-Shot Learning~(CDFSL) methods typically parameterize models with task-agnostic and task-specific parameters. To adapt task-specific parameters, recent approaches have utilized fixed optimization strategies, despite their potential sub-optimality across varying domains or target tasks. To address this issue, we propose a novel adaptation mechanism called Task-Specific Preconditioned gradient descent~(TSP). Our method first meta-learns Domain-Specific Preconditioners~(DSPs) that capture the characteristics of each meta-training domain, which are then linearly combined using task-coefficients to form the Task-Specific Preconditioner. The preconditioner is applied to gradient descent, making the optimization adaptive to the target task. We constrain our preconditioners to be positive definite, guiding the preconditioned gradient toward the direction of steepest descent. Empirical evaluations on the Meta-Dataset show that TSP achieves state-of-the-art performance across diverse experimental scenarios.

Figures

Figures reproduced from arXiv: 2412.15483 by the authors.

**Figure 3.** (a) PGD with Domain-Specific Preconditioner (DSP) in the inner-level optimization. During metatraining, for a train task T , DSP is chosen based on the domain label dT , and each task-specific parameter θ l are optimized using PGD with the selected DSP Pl dT . (b) PGD with Task-Specific Preconditioner. During meta-testing, for a test task, each Task-Specific Preconditioner Pl T is contructed using DSPs and task… view at source ↗

**Figure 4.** Learning curves of PGD with and without the PD [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗

Figures from the paper (2 more)

**Figure 5.** Figure 5: Task-coefficient values used in the construction of Task-Specific Preconditioner. Columns represent DSPs trained on [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]

**Figure 6.** Figure 6: TSP applied on TSA and TA2 -Net. (a) PGD with Domain-Specific Preconditioner (DSP) applied on TSA during meta-training. (b) PGD with Task-Specific Preconditioner applied on TSA during meta-testing. (c) PGD with Domain-Specific Preconditioner (DSP) applied on TA2 -Net d…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

67 extracted references · 42 canonical work pages

[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]

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]

Amari, S. 1967. A theory of adaptive pattern classifiers. IEEE Transactions on Electronic Computers, (3): 299--307

work page 1967
[4]

Amari, S.-i. 1996. Neural learning in structured parameter spaces-natural Riemannian gradient. Advances in neural information processing systems, 9

work page 1996
[5]

Amari, S.-I. 1998. Natural gradient works efficiently in learning. Neural computation, 10(2): 251--276

work page 1998
[6]

Amari, S.-i.; Ba, J.; Grosse, R.; Li, X.; Nitanda, A.; Suzuki, T.; Wu, D.; and Xu, J. 2020. When does preconditioning help or hurt generalization? arXiv preprint arXiv:2006.10732

arXiv 2020
[7]

Amari, S.-I.; and Douglas, S. C. 1998. Why natural gradient? In Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP'98 (Cat. No. 98CH36181), volume 2, 1213--1216. IEEE

work page 1998
[8]

Baik, S.; Choi, M.; Choi, J.; Kim, H.; and Lee, K. M. 2023. Learning to learn task-adaptive hyperparameters for few-shot learning. IEEE Transactions on Pattern Analysis and Machine Intelligence

work page 2023

Show all 67 references

[9]

Bateni, P.; Barber, J.; Van de Meent, J.-W.; and Wood, F. 2022. Enhancing few-shot image classification with unlabelled examples. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2796--2805

2022
[10]

Bateni, P.; Goyal, R.; Masrani, V.; Wood, F.; and Sigal, L. 2020. Improved few-shot visual classification. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 14493--14502

2020
[11]

F.; and Huang, J.-B

Chen, W.-Y.; Liu, Y.-C.; Kira, Z.; Wang, Y.-C. F.; and Huang, J.-B. 2019. A closer look at few-shot classification. arXiv preprint arXiv:1904.04232

2019 arXiv
[12]

Cimpoi, M.; Maji, S.; Kokkinos, I.; Mohamed, S.; and Vedaldi, A. 2014. Describing textures in the wild. In Proceedings of the IEEE conference on computer vision and pattern recognition, 3606--3613

2014
[13]

Duchi, J.; Hazan, E.; and Singer, Y. 2011. Adaptive subgradient methods for online learning and stochastic optimization. Journal of machine learning research, 12(7)

2011
[14]

Dvornik, N.; Schmid, C.; and Mairal, J. 2020. Selecting relevant features from a multi-domain representation for few-shot classification. In Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part X 16, 769--786. Springer

2020
[15]

Finn, C.; Abbeel, P.; and Levine, S. 2017. Model-agnostic meta-learning for fast adaptation of deep networks. In International conference on machine learning, 1126--1135. PMLR

2017
[16]

Garcia, V.; and Bruna, J. 2017. Few-shot learning with graph neural networks. arXiv preprint arXiv:1711.04043

2017 arXiv
[17]

W.; Rezende, D.; and Eslami, S

Garnelo, M.; Rosenbaum, D.; Maddison, C.; Ramalho, T.; Saxton, D.; Shanahan, M.; Teh, Y. W.; Rezende, D.; and Eslami, S. A. 2018. Conditional neural processes. In International conference on machine learning, 1704--1713. PMLR

2018
[18]

Guo, Y.; Du, R.; Dong, Y.; Hospedales, T.; Song, Y.-Z.; and Ma, Z. 2023. Task-aware Adaptive Learning for Cross-domain Few-shot Learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 1590--1599

2023
[19]

Ha, D.; and Eck, D. 2017. A neural representation of sketch drawings. arXiv preprint arXiv:1704.03477

2017 arXiv
[20]

M.; et al

Himmelblau, D. M.; et al. 2018. Applied nonlinear programming. McGraw-Hill

2018
[21]

A.; and Johnson, C

Horn, R. A.; and Johnson, C. R. 2012. Matrix analysis. Cambridge university press

2012
[22]

Houben, S.; Stallkamp, J.; Salmen, J.; Schlipsing, M.; and Igel, C. 2013. Detection of traffic signs in real-world images: The German Traffic Sign Detection Benchmark. In The 2013 international joint conference on neural networks (IJCNN), 1--8. Ieee

2013
[23]

Kakade, S. M. 2001. A natural policy gradient. Advances in neural information processing systems, 14

2001
[24]

Kang, S.; Hwang, D.; Eo, M.; Kim, T.; and Rhee, W. 2023. Meta-Learning with a Geometry-Adaptive Preconditioner. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 16080--16090

2023
[25]

P.; and Ba, J

Kingma, D. P.; and Ba, J. 2014. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980

2014 arXiv
[26]

Krizhevsky, A.; Hinton, G.; et al. 2009. Learning multiple layers of features from tiny images

2009
[27]

Lake, B.; Salakhutdinov, R.; Gross, J.; and Tenenbaum, J. 2011. One shot learning of simple visual concepts. In Proceedings of the annual meeting of the cognitive science society, volume 33

2011
[28]

M.; Salakhutdinov, R.; and Tenenbaum, J

Lake, B. M.; Salakhutdinov, R.; and Tenenbaum, J. B. 2015. Human-level concept learning through probabilistic program induction. Science, 350(6266): 1332--1338

2015
[29]

LeCun, Y.; Bottou, L.; Bengio, Y.; and Haffner, P. 1998. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11): 2278--2324

1998
[30]

B.; and M \"u ller, K.-R

LeCun, Y.; Bottou, L.; Orr, G. B.; and M \"u ller, K.-R. 2002. Efficient backprop. In Neural networks: Tricks of the trade, 9--50. Springer

2002
[31]

Lee, Y.; and Choi, S. 2018. Gradient-based meta-learning with learned layerwise metric and subspace. In International Conference on Machine Learning, 2927--2936. PMLR

2018
[32]

Li, W.-H.; Liu, X.; and Bilen, H. 2021. Universal representation learning from multiple domains for few-shot classification. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 9526--9535

2021
[33]

Li, W.-H.; Liu, X.; and Bilen, H. 2022. Cross-domain few-shot learning with task-specific adapters. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 7161--7170

2022
[34]

Li, X.-L. 2017. Preconditioned stochastic gradient descent. IEEE transactions on neural networks and learning systems, 29(5): 1454--1466

2017
[35]

Li, Z.; Zhou, F.; Chen, F.; and Li, H. 2017. Meta-sgd: Learning to learn quickly for few-shot learning. arXiv preprint arXiv:1707.09835

2017 arXiv
[36]

Lin, T.-Y.; Maire, M.; Belongie, S.; Hays, J.; Perona, P.; Ramanan, D.; Doll \'a r, P.; and Zitnick, C. L. 2014. Microsoft coco: Common objects in context. In Computer Vision--ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part V 1...

2014
[37]

Liu, L.; Hamilton, W.; Long, G.; Jiang, J.; and Larochelle, H. 2020. A universal representation transformer layer for few-shot image classification. arXiv preprint arXiv:2006.11702

2020 arXiv
[38]

Liu, Y.; Lee, J.; Zhu, L.; Chen, L.; Shi, H.; and Yang, Y. 2021. A multi-mode modulator for multi-domain few-shot classification. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 8453--8462

2021
[39]

Maji, S.; Rahtu, E.; Kannala, J.; Blaschko, M.; and Vedaldi, A. 2013. Fine-grained visual classification of aircraft. arXiv preprint arXiv:1306.5151

2013 arXiv
[40]

Mishra, N.; Rohaninejad, M.; Chen, X.; and Abbeel, P. 2017. A simple neural attentive meta-learner. arXiv preprint arXiv:1707.03141

2017 arXiv
[41]

Munkhdalai, T.; and Yu, H. 2017. Meta networks. In International conference on machine learning, 2554--2563. PMLR

2017
[42]

Nilsback, M.-E.; and Zisserman, A. 2008. Automated flower classification over a large number of classes. In 2008 Sixth Indian conference on computer vision, graphics & image processing, 722--729. IEEE

2008
[43]

Nocedal, J.; and Wright, S. J. 1999. Numerical optimization. Springer

1999
[44]

Oreshkin, B.; Rodr \' guez L \'o pez, P.; and Lacoste, A. 2018. Tadam: Task dependent adaptive metric for improved few-shot learning. Advances in neural information processing systems, 31

2018
[45]

Park, E.; and Oliva, J. B. 2019. Meta-curvature. Advances in Neural Information Processing Systems, 32

2019
[46]

S.; and Shah, M

Rajasegaran, J.; Khan, S.; Hayat, M.; Khan, F. S.; and Shah, M. 2020. Meta-learning the learning trends shared across tasks. arXiv preprint arXiv:2010.09291

2020 arXiv
[47]

M.; and Levine, S

Rajeswaran, A.; Finn, C.; Kakade, S. M.; and Levine, S. 2019. Meta-learning with implicit gradients. Advances in neural information processing systems, 32

2019
[48]

Ravi, S.; and Larochelle, H. 2016. Optimization as a model for few-shot learning. In International conference on learning representations

2016
[49]

Requeima, J.; Gordon, J.; Bronskill, J.; Nowozin, S.; and Turner, R. E. 2019. Fast and flexible multi-task classification using conditional neural adaptive processes. Advances in Neural Information Processing Systems, 32

2019
[50]

Roy, O.; and Vetterli, M. 2007. The effective rank: A measure of effective dimensionality. In 2007 15th European signal processing conference, 606--610. IEEE

2007
[51]

Russakovsky, O.; Deng, J.; Su, H.; Krause, J.; Satheesh, S.; Ma, S.; Huang, Z.; Karpathy, A.; Khosla, A.; Bernstein, M.; et al. 2015. Imagenet large scale visual recognition challenge. International journal of computer vision, 115: 211--252

2015
[52]

Saad, Y. 2003. Iterative methods for sparse linear systems. SIAM

2003
[53]

Saikia, T.; Brox, T.; and Schmid, C. 2020. Optimized generic feature learning for few-shot classification across domains. arXiv preprint arXiv:2001.07926

2020 arXiv
[54]

Santoro, A.; Bartunov, S.; Botvinick, M.; Wierstra, D.; and Lillicrap, T. 2016. Meta-learning with memory-augmented neural networks. In International conference on machine learning, 1842--1850. PMLR

2016
[55]

Schroeder, B.; and Cui, Y. 2018. Fgvcx fungi classification challenge 2018. Available online: github. com/visipedia/fgvcx\_fungi\_comp (accessed on 14 July 2021)

2018
[56]

Simon, C.; Koniusz, P.; Nock, R.; and Harandi, M. 2020. On modulating the gradient for meta-learning. In Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part VIII 16, 556--572. Springer

2020
[57]

Snell, J.; Swersky, K.; and Zemel, R. 2017. Prototypical networks for few-shot learning. Advances in neural information processing systems, 30

2017
[58]

H.; and Hospedales, T

Sung, F.; Yang, Y.; Zhang, L.; Xiang, T.; Torr, P. H.; and Hospedales, T. M. 2018. Learning to compare: Relation network for few-shot learning. In Proceedings of the IEEE conference on computer vision and pattern recognition, 1199--1208

2018
[59]

Tian, H.; Liu, F.; Liu, T.; Du, B.; Cheung, Y.-m.; and Han, B. 2024. MOKD: Cross-domain Finetuning for Few-shot Classification via Maximizing Optimized Kernel Dependence. arXiv preprint arXiv:2405.18786

2024 arXiv
[60]

B.; and Isola, P

Tian, Y.; Wang, Y.; Krishnan, D.; Tenenbaum, J. B.; and Isola, P. 2020. Rethinking few-shot image classification: a good embedding is all you need? In Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XIV 16, 266--282. Springer

2020
[61]

Triantafillou, E.; Larochelle, H.; Zemel, R.; and Dumoulin, V. 2021. Learning a universal template for few-shot dataset generalization. In International Conference on Machine Learning, 10424--10433. PMLR

2021
[62]

Triantafillou, E.; Zhu, T.; Dumoulin, V.; Lamblin, P.; Evci, U.; Xu, K.; Goroshin, R.; Gelada, C.; Swersky, K.; Manzagol, P.-A.; et al. 2019. Meta-dataset: A dataset of datasets for learning to learn from few examples. arXiv preprint arXiv:1903.03096

2019 arXiv
[63]

Von Oswald, J.; Zhao, D.; Kobayashi, S.; Schug, S.; Caccia, M.; Zucchet, N.; and Sacramento, J. 2021. Learning where to learn: Gradient sparsity in meta and continual learning. Advances in Neural Information Processing Systems, 34: 5250--5263

2021
[64]

Wah, C.; Branson, S.; Welinder, P.; Perona, P.; and Belongie, S. 2011. The caltech-ucsd birds-200-2011 dataset

2011
[65]

Yoon, J.; Kim, T.; Dia, O.; Kim, S.; Bengio, Y.; and Ahn, S. 2018. Bayesian model-agnostic meta-learning. Advances in neural information processing systems, 31

2018
[66]

R.; and Smola, A

Zaheer, M.; Kottur, S.; Ravanbakhsh, S.; Poczos, B.; Salakhutdinov, R. R.; and Smola, A. J. 2017. Deep sets. Advances in neural information processing systems, 30

2017
[67]

Zhao, D.; Kobayashi, S.; Sacramento, J.; and von Oswald, J. 2020. Meta-learning via hypernetworks. In 4th Workshop on Meta-Learning at NeurIPS 2020 (MetaLearn 2020). NeurIPS

2020

Pith tools

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

[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]

Amari, S. 1967. A theory of adaptive pattern classifiers. IEEE Transactions on Electronic Computers, (3): 299--307

work page 1967

[4] [4]

Amari, S.-i. 1996. Neural learning in structured parameter spaces-natural Riemannian gradient. Advances in neural information processing systems, 9

work page 1996

[5] [5]

Amari, S.-I. 1998. Natural gradient works efficiently in learning. Neural computation, 10(2): 251--276

work page 1998

[6] [6]

Amari, S.-i.; Ba, J.; Grosse, R.; Li, X.; Nitanda, A.; Suzuki, T.; Wu, D.; and Xu, J. 2020. When does preconditioning help or hurt generalization? arXiv preprint arXiv:2006.10732

arXiv 2020

[7] [7]

Amari, S.-I.; and Douglas, S. C. 1998. Why natural gradient? In Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP'98 (Cat. No. 98CH36181), volume 2, 1213--1216. IEEE

work page 1998

[8] [8]

Baik, S.; Choi, M.; Choi, J.; Kim, H.; and Lee, K. M. 2023. Learning to learn task-adaptive hyperparameters for few-shot learning. IEEE Transactions on Pattern Analysis and Machine Intelligence

work page 2023