PatchGen: Learning Soft Intra-Image Predictive Subsets for Visual Generalization

Weimiao Yu; Xi Yang; Zhaorui Tan

REVIEW 4 major objections 5 minor 237 references

PatchGen: Learning Soft Intra-Image Predictive Subsets for Visual Generalization

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read The paper claims each image contains an oracle intra-image predictive subset of patches sufficient for label prediction, and that a learned soft mask can serve as an effective task-driven proxy for it.

desk verdict A serious method paper whose theoretical bridge is honest but unverified; worth a full referee, not a desk reject. read the letter →

arxiv 2608.12766 v1 pith:KBO5MYGS submitted 2026-08-13 cs.CV

Zhaorui Tan , Weimiao Yu , Xi Yang This is my paper · ORCID

classification cs.CV

keywords intra-imagepredictivesubsetvisualgeneralizationdomaincontinualcategorydiscoveryall-shiftsoftpatchmasksufficiencyhistopathology

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

PatchGen sets out to establish that, for visual classification under distribution shifts, each image contains a sample-specific oracle intra-image predictive subset $C^\star(X)$ of patches that carries all label-relevant information; the remaining patches are complementary context that may correlate with the label but is conditionally redundant once $C^\star(X)$ is given. If true, this matters because existing generalization methods align whole-image representations and can be misled by spurious context that co-occurs with predictive evidence in the source data. The paper proves that restricting prediction to the oracle subset preserves the Bayes risk of the full-patch representation while its empirical Rademacher complexity is bounded by $\Lambda B\sqrt{s}/\sqrt{n}$ instead of $\Lambda B\sqrt{P}/\sqrt{n}$, a strictly tighter bound whenever $s

What carries the argument

The load-bearing object is the oracle intra-image predictive subset $C^\star(X)$, an unobserved sample-dependent partition of the patch index set into predictive and complementary parts. The identity that carries the argument is the masked concatenation $\Phi_m(X)$, the patch features weighted by a mask $m$ and concatenated: because each selected patch has norm at most $B$ and at most $s$ patches are selected, $\|\Phi_C(X)\|_2 \leq B\sqrt{s}$, which converts directly into the empirical Rademacher bound of Proposition 3.1. The mechanism that operationalizes this is PatchGen's soft mask $m_\phi(X)\in(0,1)^P$, computed as the sigmoid of the average outgoing cross-patch interaction score across attention heads and target patches; the selected and complementary aggregates are formed by a shared depthwise patch aggregator refined by a sample-dependent channel gate, and the mask is trained by the main task loss together with low-score suppression, confidence on the selected branch, and within-class similarity.

What would settle it

Build a synthetic classification task with a known oracle subset (for example, a digit pasted on procedurally varied backgrounds whose distribution shifts across domains), train PatchGen, and compare the observed risk gap between two domains with the Proposition 3.2 bound $LBP(\epsilon_d+\epsilon_{d'})$; if the gap consistently exceeds the bound, or if the Bayes risk using only the oracle subset drops below that of all patches, the stable predictive-sufficiency assumption fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that predictive sufficiency is an intra-image property: for every image there is an unobserved oracle subset $C^\star(X)$ such that the masked representation $\Phi_C$ is conditionally sufficient for the label, and the rest of the image contributes no label information once $\Phi_C$ is given. Under this assumption, Proposition 3.1 shows that the unrestricted Bayes risks based on $\Phi_C$ and $\Phi_{\mathrm{all}}$ coincide, while the empirical Rademacher complexity of the oracle-subset score class scales with $\sqrt{s}$ rather than $\sqrt{P}$. Proposition 3.2 adds that, for shifts that only change the complementary-context distribution, the risk gap between two domains is at most $LBP(\epsilon_d(\phi)+\epsilon_{d'}(\phi))$, where $\epsilon_d$ is the normalized mask-approximation error between the learned soft mask and the oracle mask; Proposition 3.3 shows that if the learned representation stays $\eta$-close to the oracle representation, distinct classes remain separated by $\gamma-2(r+\eta)$. The paper's positive thesis is that PatchGen's learned soft mask, derived from cross-patch interaction scores and trained with low-score suppression, selected-confidence regularization, and class-conditional alignment, is an effective task-driven proxy for $C^\star(X)$, with matched-backbone improvements in most evaluated configurations and better unknown-class accuracy in discovery tasks.

Load-bearing premise

The load-bearing premise is Assumption 3.2: once the oracle predictive subset's representation is given, the remaining patches are conditionally redundant and the label-predictive mechanism is the same in every domain; the paper does not directly measure whether that holds for real images.

Editorial extensions

If this is right

If the oracle subset exists, sample-adaptive patch selection is not a capacity trade-off: it attains the same Bayes risk as the full representation while shrinking the empirical Rademacher complexity bound from $\Lambda B\sqrt{P}/\sqrt{n}$ to $\Lambda B\sqrt{s}/\sqrt{n}$.
Under complementary-context shifts, the risk discrepancy across domains is bounded by the normalized per-patch mask error between the learned and oracle masks, so improving mask fidelity directly translates into more stable cross-domain accuracy.
When the learned mask is close to the oracle mask, class-conditional representation sets remain pairwise disjoint under the assumed clusterability condition, which is the paper's stated explanation for improved unknown-class accuracy in continual category discovery and the all-shift setting.
Empirically, PatchGen improves matched-backbone baselines in most evaluated configurations on natural and histopathology benchmarks, and its largest average gains appear on unknown classes in target-shift and all-shift experiments.

Reading between the lines

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

Extension: Proposition 3.2 gives a testable quantitative prediction—the cross-domain risk gap is at most a constant times the sum of per-image mask errors—so a synthetic benchmark with a known oracle subset could measure the mask approximation error directly and check whether the bound holds, which the paper's diagnostics do not do.
Extension: Because the mask is learned without text or category names, the same selector could serve as a weakly supervised localization signal in histopathology or other dense-image domains, turning a generalization module into an interpretability output.
Extension: The all-shift guarantee is componentwise, so a natural next step is to couple the mask approximation error to pseudo-label quality, testing whether proxy-mask errors compound with discovery errors in later continual sessions.

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.

Desk Editor's Note

read the letter

Two things you should know about PatchGen. First, it is a genuine new method, not a repackaging. The sample-dependent soft predictive-subset mask, jointly trained with mask suppression, confidence regularization, and class-conditional alignment, is not present in the cited token-pruning, attention-visualization, or causal-DG literature. Second, the theory is explicitly conditional and the authors admit the central bridge is unverified. That honesty is rare and worth crediting.

What the paper does well: the experiments are extensive. They cover multi-domain generalization, continual category discovery, and the combined mDG+GCD setting, on natural and histopathological images, with multiple backbones. The matched-backbone comparisons show positive average gains in most configurations. The ablation study is well designed, including an equal-parameter control (B+Mp) that isolates the proposed constraints from added capacity. The histopathology visualizations were reviewed by a pathologist and are presented as model-selection maps, not clinical ground truth. The reproducibility appendix documents hyperparameters, seeds, and data splits carefully, and the Appendix Table 10 sensitivity analysis on TerraIncognita shows low variance across seeds.

The soft spots are real but proportionate. The load-bearing assumption, stable predictive sufficiency (Assumption 3.2), is untestable with the current experiments. The diagnostics in Section 5.5 show that the learned mask is decision-relevant for the trained classifier, but they do not establish that the mask recovers the oracle subset or that the selected conditional law is stable across domains. The paper says exactly this in Sections 5.5 and 6, so it is not overclaiming. Still, Proposition 3.2's bound is vacuous unless the mask-approximation error is small, and no evidence establishes that. The main tables report single-seed results for most comparisons; the appendix addresses stability on one dataset only. Code release is deferred to publication, which is a minor reproducibility concern.

These issues do not sink the paper. The method works as a practical heuristic, the theory is a motivation rather than a load-bearing proof, and the empirical evidence is fairly presented. A revision with error bars in the main tables and a clearer statement that the oracle interpretation is a hypothesis, not a verified mechanism, would strengthen it. But even as is, this is a solid contribution to visual generalization that deserves a serious referee.

My recommendation: accept for peer review. The authors have engaged honestly with the literature and their own limitations, and the work is novel enough that a good referee could help shape a useful final version.

Referee Report

4 major / 5 minor

Summary. The paper proposes the structural hypothesis that each image contains a sample-adaptive oracle intra-image predictive subset C*(X) whose patch representations are sufficient for label prediction, with the complementary context conditionally redundant given the subset. The theoretical section (Section 3) states Assumptions 3.1-3.3 and derives Proposition 3.1 (Bayes-risk equality plus oracle-subset-size-dependent Rademacher complexity), Proposition 3.2 (risk-discrepancy bound for complementary-context shifts controlled by mask-approximation error), Proposition 3.3 (class-separation preservation under target shifts), and Corollary 3.1 (componentwise implications for all shifts). The proposed method PatchGen (Section 4) learns a soft per-patch mask via attention-derived interaction scores, jointly optimized with a main task objective and three auxiliary losses (low-score mask suppression Lms, selected-confidence regularization Lconf, and class-conditional feature alignment Lsim). Experiments cover multi-domain generalization (mDG) on natural and histopathological images, continual category discovery (CCD), and multi-domain generalization with generalized category discovery (mDG+GCD). The paper reports average gains over matched-backbone baselines in most evaluated configurations, competitive results against vision-language methods without text supervision, and qualitative diagnostics on patch-selection maps.

Significance. The paper is valuable for its clear formulation of an intra-image predictive-sufficiency view of generalization and for the honest, well-scoped presentation of its assumptions and limitations. The Rademacher complexity bound in Eq. (2) is a correct and useful illustration of how restricting prediction to a small predictive subset can tighten capacity bounds, and the empirical study is broad, covering three shift settings and including matched-backbone comparisons, ablation controls, and robustness checks. The method is text-free, which is a practical advantage for category-discovery settings. However, the significance is limited by the unverified bridge between the unobserved oracle mask and the learned soft mask: neither the existence of C*(X) nor the smallness of the mask-approximation error epsilon_d(phi) is established for the evaluated benchmarks, so the theoretical results currently motivate rather than explain the observed empirical gains.

major comments (4)

[Section 3, Proposition 3.1] Proposition 3.1's first claim, that the unrestricted Bayes risk based on Phi_C equals that based on Phi_all, is a direct restatement of the first equality in Assumption 3.2 rather than a derived consequence. As stated, the proposition proves the equality by invoking the assumption in the proof (Appendix C.3). The genuinely new content is the Rademacher complexity bound in Eq. (2), which follows from Assumption 3.3. I recommend rephrasing Proposition 3.1 to distinguish the assumed Bayes-risk equality from the derived complexity bound, so that the 'oracle sufficiency' language does not imply the assumption has been established.
[Section 5.5 and Section 6] The central link between the theory and PatchGen is not established. Section 5.5 explicitly states that the diagnostics 'do not validate the oracle structural assumption or identify C*(X)', and Section 6 states that neither the training objectives nor the diagnostic experiments guarantee recovery of m*_C(X). This means the quantity epsilon_d(phi) in Proposition 3.2 (Eq. 4) is never measured or bounded, so the risk-discrepancy bound is vacuous unless mask-approximation error is small, and no evidence is provided that it is. The observed average gains could plausibly arise from generic attention-like feature weighting or regularization rather than from recovering a predictive-sufficient subset. I ask the authors to provide a concrete test of the oracle-subset hypothesis, for example by constructing a synthetic benchmark where C*(X) is known and measuring the learned mask's approximation error, or by developing an indirect falsifiable check of Assumption 3.2 on real data. At a minimum, the paper should explicitly state that the theoretical results are conditional on an untested structural assumption and do not by themselves explain the empirical improvements.
[Section 5.1, Tables 1 and 10] The headline comparisons are based on a single seed (seed 1), and across-seed stability is reported only for TerraIncognita (Table 10). Several per-dataset entries show negative gains relative to the LP+LN baseline (e.g., OfficeHome with DINOv2 ViT-B/14: 83.7 vs. 86.1 in Table 1), while the average gain is positive. This makes the claim 'improves average performance in most evaluated configurations' sensitive to the particular seed and to the choice of datasets. I recommend reporting mean and standard deviation over multiple seeds for all matched comparisons, or at least providing a leave-one-dataset-out analysis, to support the average-effect claim.
[Section 5.4, mDG+GCD protocol] The mDG+GCD setting follows the L-Reg protocol in which the unlabeled pool contains only samples from unknown classes. This is a non-standard generalized category discovery setup, and it means the unknown-class accuracy numbers are not directly comparable to mixed known/unknown GCD protocols. The paper should make this protocol distinction more prominent when claiming 'generalization to unknown classes', since the practical difficulty of mixed unlabeled pools is not addressed.

minor comments (5)

[Abstract and Introduction] The abstract mentions 'softhe P patches' in the Introduction (near Eq. 1); this appears to be a typo for 'subset of the P patches' and should be corrected.
[Section 4, Eq. (3)] The definition of the patch-selection score in Eq. (3) divides by H*P, where H is the number of heads and P the number of patches. It may be clearer to explicitly state that the normalization is over both heads and target patches; the current notation H/P is slightly ambiguous.
[Section 5.2, HISTOPANTUM results] In Table 3, PatchGen with CLIP ViT-B/16 without SWAD is 88.5 average, slightly below the LP+LN baseline at 88.8. The text says 'improves the matched baselines in average performance across most backbone and evaluation settings'; this is accurate, but the exception should be acknowledged in the text for full transparency.
[Section 5.5, diagnostics] The patch perturbation test uses a fixed threshold of 0.5 to define the selected set. Since the learned mask is soft, a sensitivity analysis over the threshold would strengthen the claim that the diagnostic conclusions are not artifacts of the threshold choice.
[Appendix C.5] The proof of Proposition 3.3 relies on an approximation assumption for every sample (eta). The paper mentions that the expected error epsilon_d(phi) does not imply the uniform condition, which is good; this caveat should be referenced in the main text when discussing target-shift preservation.

Circularity Check

1 steps flagged · score 4.0 of 10

The oracle-sufficiency half of Proposition 3.1 restates Assumption 3.2 as a theorem; the method and experiments are otherwise self-contained.

self definitional [Section 3, Assumption 3.2 and Proposition 3.1; proof in Appendix C.3]
"Assumption 3.2 (Stable predictive sufficiency). For every domain d∈D, P_d(Y|Φ_all)=P_d(Y|Φ_C)=P(Y|Φ_C). ... Proposition 3.1 ... Under Assumptions 3.2 and 3.3, for every domain d∈D, the unrestricted Bayes risk based on ΦC equals that based on Φall."

The first clause of Proposition 3.1 is a verbatim restatement of Assumption 3.2: if the conditional label law under Φ_all equals the conditional label law under Φ_C, then any risk that is a functional of that law is identical by construction. The proof in Appendix C.3 confirms this: 'Stable predictive sufficiency gives P(Y|Φ_all)=P(Y|Φ_C). Hence the conditional label law, and therefore the Bayes decision, depends only on ΦC.' The 'oracle-sufficiency' result is therefore not derived from weaker evidence; it is the defining content of the assumption. Only the Rademacher bound in the same proposition is an independent derivation.

full rationale

One genuine circular step exists: the Bayes-risk-preservation half of Proposition 3.1 is exactly Assumption 3.2, so the paper's 'theoretical analysis shows... preserves the Bayes risk' is an assumption restated as a conclusion. This is a self-definitional move, but it is limited to the oracle-sufficiency framing. The complexity bound in Proposition 3.1 and the bounds in Propositions 3.2 and 3.3 are valid derivations from their stated assumptions; Proposition 3.2 is a Lipschitz/triangle-inequality translation of mask-approximation error into risk discrepancy, which is a routine implication rather than a disguised fit. The paper does not claim to have tested Assumption 3.2; Section 5.5 states the diagnostics 'do not validate the oracle structural assumption or identify C*(X)', and the Limitations section concedes that no training objective guarantees recovery of m*_C(X). Those are honest gaps between theory and method, not circular reductions. The empirical evaluation is self-contained against matched-backbone baselines, and the self-citations to GMDG and L-Reg are used as baselines or host pseudo-label protocols, not as load-bearing justification for the core derivation. Overall, the circularity is confined to one definitional theorem, so a moderate score of 4 is appropriate.

Assumptions & free parameters 3 free parameters · 6 assumptions · 1 invented entities

The ledger shows that the theory rests on strong, unverified domain assumptions about an unobserved oracle subset, while the empirical method introduces only standard hyperparameters. No new physical entity or external benchmark is imported into the central derivation.

free parameters (3)

lambda_ms (mask suppression weight) = 0.001 (selected from {0.01, 0.001, 0.0001} on source-domain validation splits)
Controls the strength of the low-score mask suppression loss; chosen by validation, not derived.
lambda_sim (class-conditional alignment weight) = 0.001 (selected similarly from {0.01, 0.001, 0.0001})
Controls the class-conditional similarity loss; chosen by source-domain validation.
tau (low-score threshold) = 0.25
Fixed threshold used in the mask-suppression loss to identify weak mask entries; set by hand and used in all experiments.

assumptions (6)

domain assumption Assumption 3.1: the oracle mask m*_C(X) is a deterministic function of Phi_all(X).
Ensures the oracle mask introduces no side information beyond the full-patch representation; used throughout the theory.
domain assumption Assumption 3.2: P_d(Y|Phi_all)=P_d(Y|Phi_C)=P(Y|Phi_C) for every domain d.
The central premise of the paper; implies complementary context is conditionally redundant and the predictive mechanism is domain-invariant. If false, the Bayes-risk equality collapses.
domain assumption Assumption 3.3: |C*(X)| <= s and ||z_p(X)|| <= B for every image and patch.
Needed for the Rademacher complexity bound in Proposition 3.1.
domain assumption Proposition 3.3 assumes an explicit clusterability condition: class centers, intra-class radius r, inter-class separation gamma, and representation approximation error eta.
Required for the class-separation result under target shift; the paper states that predictive sufficiency alone does not imply unknown-class discovery.
standard math The loss u -> ell(h(u), y) is L-Lipschitz and the aggregation map G is non-expansive.
Regularity conditions used in Proposition 3.2 and Lemma C.1.
standard math Standard Rademacher complexity and generalization bound arguments are used in Proposition 3.1 and Corollary C.1.
Background learning theory that the proofs rely on.

invented entities (1)

Oracle intra-image predictive subset C*(X) and its mask m*_C(X)
purpose: Posited sample-adaptive set of patches sufficient for label prediction; used to motivate the learned soft mask proxy.
Unobserved and unverifiable. The paper explicitly states in Section 6 that neither the training objectives nor the diagnostic experiments guarantee recovery of m*_C(X).

how reviews work

0 comments

Cite this review

Pith. "Pith review of PatchGen: Learning Soft Intra-Image Predictive Subsets for Visual Generalization." pith.science (2026). https://pith.science/paper/KBO5MYGS

@misc{pith2026260812766,
  author       = {Pith},
  title        = {Pith review of: PatchGen: Learning Soft Intra-Image Predictive Subsets for Visual Generalization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KBO5MYGS}},
  note         = {Machine review of arXiv:2608.12766}
}

read the original abstract

Visual classifiers are expected to generalize under data shifts, target shifts, and their combinations, yet most existing methods focus on domain invariance while failing to address intra-image predictive sufficiency. We investigate the structural hypothesis that each image contains a sample-adaptive oracle intra-image predictive subset sufficient for label prediction, while the remaining patches form non-essential complementary context that may correlate with the label. The theoretical analysis shows that restricting prediction to this oracle subset preserves the Bayes risk achievable by the full-patch representation while admitting a complexity bound that tightens with the oracle-subset size. Based on this view, we propose PatchGen, a text-free module that learns a sample-dependent soft predictive-subset mask as a task-driven proxy for the unobserved oracle subset mask. Specifically, histopathology visualizations suggest that PatchGen assigns higher scores to tumor-consistent regions than to some frequently co-occurring inflammatory context. Extensive experiments on natural and histopathological image benchmarks spanning all three shift settings show that PatchGen improves average performance over matched-backbone baselines in most evaluated configurations, enhances generalization to unknown classes, and remains competitive with vision-language methods without text supervision.

Figures

Figures reproduced from arXiv: 2608.12766 by the authors.

**Figure 1.** Qualitative histopathology visualization on an unseen-domain tumor image: PatchGen’s learned soft predictive-subset mask compared with the corresponding attention maps of previous methods. evidence from a complementary context that repeatedly cooccurs with predictive evidence in the source data but becomes unreliable under domain or label-space shifts. For example, a model may exploit inflammation, which may usuall… view at source ↗

**Figure 2.** Overview of PatchGen. (a) The structural hypothesis. (b) PatchGen learns a soft mask as a task-driven proxy for the oracle intra-image predictive subset. ∥G(u) − G(v)∥2 ≤ ∥u − v∥2. Let h be a fixed classifier operating on R dg , and let ℓ denote its loss. Define the learned selected representation and the oracle representation as ze + ϕ (X) = G [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 4.** [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗

Figures from the paper (8 more)

**Figure 3.** Figure 3: Main MDG results for natural images: Main comparison between PatchGen and previous MDG methods, including VLM-based and VM-based methods. Detailed results for all 36 methods are in Appendix [PITH_FULL_IMAGE:figures/full_fig_p005_3.png]

**Figure 6.** Figure 6: CCD results (A/K/U = all/known/unknown classes): Unknown-class accuracy across CCD sessions and average A/K/U gains of PatchGen over PromptCCD for DINO/DINOv2. Detailed results for each dataset, backbone, and session are in Appendix [PITH_FULL_IMAGE:figures/full_fig_…

**Figure 5.** Figure 5: MDG results for histopathological images: Comparison between the baselines and PatchGen with average and per-domain results across different backbones on the histopathological datasets (HISTOPANTUM and HISTOCOLON). Detailed results are in Appendix [PITH_FULL_IMAGE:f…

**Figure 8.** Figure 8: Patch-selection diagnostics on VLCS (SUN09 unseen) with DINOv2-ViT-S/14. (A) Patch perturbation: keeping selected patches is robust, while replacing selected patches causes a sharp accuracy drop. (B) Insertion/deletion AUC curves. (C) UMAP of baseline vs. PatchGen fea…

**Figure 7.** Figure 7: MDG+GCD results: Stacked per-dataset comparison across prior methods, matched baselines, and PatchGen under simultaneous unseen-domain and unknown-class shifts. See numeric results in Appendix Tables 6&7. {p : mϕ,p(X) ≥ 0.5} denote the thresholded selected patch set, …

**Figure 9.** Figure 9: of [PITH_FULL_IMAGE:figures/full_fig_p014_9.png]

**Figure 10.** Figure 10: Ablation results: Average-accuracy trend as components are added, plus per-dataset gains of the full model over the baseline. The detailed results of each dataset are in Appendix [PITH_FULL_IMAGE:figures/full_fig_p015_10.png]

**Figure 11.** Figure 11: Additional pathology tumor examples under the mDG setting. Highlighted regions correspond to regions visually [PITH_FULL_IMAGE:figures/full_fig_p016_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

237 extracted references · 33 canonical work pages

[1]

International Journal of Computer Vision , volume=

Incremental model enhancement via memory-based contrastive learning , author=. International Journal of Computer Vision , volume=. 2025 , publisher=

2025
[2]

International Journal of Computer Vision , volume=

Domain-specific bias filtering for single labeled domain generalization , author=. International Journal of Computer Vision , volume=. 2023 , publisher=

2023
[3]

Neural computation , volume=

Long short-term memory , author=. Neural computation , volume=. 1997 , publisher=

1997
[4]

Advances in Neural Information Processing Systems , volume=

Interpret your decision: Logical reasoning regularization for generalization in visual classification , author=. Advances in Neural Information Processing Systems , volume=
[5]

Journal of Machine Learning Research , volume=

Generalization on the unseen, logic reasoning and degree curriculum , author=. Journal of Machine Learning Research , volume=
[6]

2022 , publisher=

Universal algebraic logic: dedicated to the unity of science , author=. 2022 , publisher=

2022
[7]

Artificial Intelligence , volume=

Conditional logic and the principle of entropy , author=. Artificial Intelligence , volume=. 2000 , publisher=

2000
[8]

proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

Rethinking multi-domain generalization with a general learning objective , author=. proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

Show all 237 references

[9]

Proceedings of the IEEE/CVF international conference on computer vision , pages=

Parametric information maximization for generalized category discovery , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=
[10]

Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

Generalized category discovery , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=
[11]

Advances in neural information processing systems , volume=

Sinkhorn distances: Lightspeed computation of optimal transport , author=. Advances in neural information processing systems , volume=
[12]

IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

Autonovel: Automatically discovering and learning novel visual categories , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2021 , publisher=

2021
[13]

Proceedings of the IEEE/CVF international conference on computer vision , pages=

A unified objective for novel class discovery , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=
[14]

arXiv preprint arXiv:2102.03526 , year=

Open-world semi-supervised learning , author=. arXiv preprint arXiv:2102.03526 , year=

arXiv
[15]

Advances in neural information processing systems , volume=

Discriminative clustering by regularized information maximization , author=. Advances in neural information processing systems , volume=
[16]

Advances in Neural Information Processing Systems , volume=

Information maximization for few-shot learning , author=. Advances in Neural Information Processing Systems , volume=
[17]

Proceedings of the IEEE/CVF international conference on computer vision , pages=

Emerging properties in self-supervised vision transformers , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=
[18]

Proceedings ofthe 5th Berkeley symposium on mathematical statisticsand probability , volume=

Multivariate observations , author=. Proceedings ofthe 5th Berkeley symposium on mathematical statisticsand probability , volume=. 1967 , organization=

1967
[19]

2009 IEEE conference on computer vision and pattern recognition , pages=

Imagenet: A large-scale hierarchical image database , author=. 2009 IEEE conference on computer vision and pattern recognition , pages=. 2009 , organization=

2009
[20]

2011 , institution=

The caltech-ucsd birds-200-2011 dataset , author=. 2011 , institution=

2011
[21]

Proceedings of the IEEE international conference on computer vision workshops , pages=

3d object representations for fine-grained categorization , author=. Proceedings of the IEEE international conference on computer vision workshops , pages=
[22]

arXiv preprint arXiv:1906.05372 , year=

The herbarium challenge 2019 dataset , author=. arXiv preprint arXiv:1906.05372 , year=

2019 arXiv
[23]

Proceedings of the IEEE international conference on computer vision , pages=

Delving deep into rectifiers: Surpassing human-level performance on imagenet classification , author=. Proceedings of the IEEE international conference on computer vision , pages=
[24]

Proceedings of the IEEE/CVF international conference on computer vision , pages=

Learning to discover novel visual categories via deep transfer clustering , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=
[25]

Computer , volume=

Self-organization in a perceptual network , author=. Computer , volume=. 1988 , publisher=

1988
[26]

IEEE transactions on pattern analysis and machine intelligence , volume=

Deep clustering: On the link between discriminative models and k-means , author=. IEEE transactions on pattern analysis and machine intelligence , volume=. 2019 , publisher=

2019
[27]

International conference on machine learning , pages=

Learning discrete representations via information maximizing self-augmented training , author=. International conference on machine learning , pages=. 2017 , organization=

2017
[28]

Advances in Neural Information Processing Systems , volume=

Realistic evaluation of transductive few-shot learning , author=. Advances in Neural Information Processing Systems , volume=
[29]

Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

Few-shot segmentation without meta-learning: A good transductive inference is all you need? , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=
[30]

pairwise losses , author=

A unifying mutual information view of metric learning: cross-entropy vs. pairwise losses , author=. European conference on computer vision , pages=. 2020 , organization=

2020
[31]

International conference on machine learning , pages=

Better mixing via deep representations , author=. International conference on machine learning , pages=. 2013 , organization=

2013
[32]

Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

Deep feature interpolation for image content changes , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=
[33]

IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

Regularizing deep networks with semantic data augmentation , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2021 , publisher=

2021
[34]

Proceedings of the AAAI Conference on Artificial Intelligence , volume=

Semantic-aware data augmentation for text-to-image synthesis , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=
[35]

arXiv preprint arXiv:1907.13625 , year=

On mutual information maximization for representation learning , author=. arXiv preprint arXiv:1907.13625 , year=

1907 arXiv
[36]

arXiv preprint arXiv:1808.06670 , year=

Learning deep representations by mutual information estimation and maximization , author=. arXiv preprint arXiv:1808.06670 , year=

arXiv
[37]

European conference on computer vision , pages=

Info3d: Representation learning on 3d objects using mutual information maximization and contrastive learning , author=. European conference on computer vision , pages=. 2020 , organization=

2020
[38]

IEEE Transactions on Information theory , volume=

Divergence measures based on the Shannon entropy , author=. IEEE Transactions on Information theory , volume=. 2002 , publisher=

2002
[39]

European conference on computer vision , pages=

Domain generalization by mutual-information regularization with pre-trained models , author=. European conference on computer vision , pages=. 2022 , organization=

2022
[40]

The Eleventh International Conference on Learning Representations , year=

Simple: Specialized model-sample matching for domain generalization , author=. The Eleventh International Conference on Learning Representations , year=
[41]

International conference on machine learning , pages=

Domain generalization via invariant feature representation , author=. International conference on machine learning , pages=. 2013 , organization=

2013
[42]

Proceedings of the AAAI conference on artificial intelligence , volume=

Domain generalization via conditional invariant representations , author=. Proceedings of the AAAI conference on artificial intelligence , volume=
[43]

Prentice Hall , volume=

Statistical signal processing: estimation theory , author=. Prentice Hall , volume=
[44]

Advances in Neural Information Processing Systems , volume=

Cooperative distribution alignment via jsd upper bound , author=. Advances in Neural Information Processing Systems , volume=
[45]

arXiv preprint arXiv:2007.01434 , year=

In search of lost domain generalization , author=. arXiv preprint arXiv:2007.01434 , year=

2007 arXiv
[46]

IEEE transactions on knowledge and data engineering , volume=

Generalizing to unseen domains: A survey on domain generalization , author=. IEEE transactions on knowledge and data engineering , volume=. 2022 , publisher=

2022
[47]

Advances in neural information processing systems , volume=

Generalizing from several related classification tasks to a new unlabeled sample , author=. Advances in neural information processing systems , volume=
[48]

Uncertainty in artificial intelligence , pages=

Domain generalization via multidomain discriminant analysis , author=. Uncertainty in artificial intelligence , pages=. 2020 , organization=

2020
[49]

arXiv preprint arXiv:1206.6471 , year=

On causal and anticausal learning , author=. arXiv preprint arXiv:1206.6471 , year=

arXiv
[50]

Journal of machine learning research , volume=

Domain-adversarial training of neural networks , author=. Journal of machine learning research , volume=
[51]

Proceedings of the AAAI conference on artificial intelligence , volume=

Domain generalization using a mixture of multiple latent domains , author=. Proceedings of the AAAI conference on artificial intelligence , volume=
[52]

European conference on computer vision , pages=

Deep coral: Correlation alignment for deep domain adaptation , author=. European conference on computer vision , pages=. 2016 , organization=

2016
[53]

Proceedings of the European conference on computer vision (ECCV) , pages=

Deep domain generalization via conditional invariant adversarial networks , author=. Proceedings of the European conference on computer vision (ECCV) , pages=
[54]

Advances in Neural Information Processing Systems , volume=

Exploiting domain-specific features to enhance domain generalization , author=. Advances in Neural Information Processing Systems , volume=
[55]

European Conference on Computer Vision , pages=

Learning to balance specificity and invariance for in and out of domain generalization , author=. European Conference on Computer Vision , pages=. 2020 , organization=

2020
[56]

Proceedings of the IEEE international conference on computer vision , pages=

Deeper, broader and artier domain generalization , author=. Proceedings of the IEEE international conference on computer vision , pages=
[57]

Proceedings of the IEEE international conference on computer vision , pages=

Unbiased metric learning: On the utilization of multiple datasets and web images for softening bias , author=. Proceedings of the IEEE international conference on computer vision , pages=
[58]

Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

Deep hashing network for unsupervised domain adaptation , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=
[59]

Proceedings of the European conference on computer vision (ECCV) , pages=

Recognition in terra incognita , author=. Proceedings of the European conference on computer vision (ECCV) , pages=
[60]

Proceedings of the IEEE/CVF international conference on computer vision , pages=

Moment matching for multi-source domain adaptation , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=
[61]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Revisiting weakly supervised pre-training of visual perception models , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[62]

Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

Deep residual learning for image recognition , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=
[63]

Advances in Neural Information Processing Systems , volume=

Swad: Domain generalization by seeking flat minima , author=. Advances in Neural Information Processing Systems , volume=
[64]

Proceedings of the IEEE/CVF conference on Computer Vision and Pattern Recognition , pages=

Adversarially adaptive normalization for single domain generalization , author=. Proceedings of the IEEE/CVF conference on Computer Vision and Pattern Recognition , pages=
[65]

Journal of statistical planning and inference , volume=

Improving predictive inference under covariate shift by weighting the log-likelihood function , author=. Journal of statistical planning and inference , volume=. 2000 , publisher=

2000
[66]

Proceedings of the AAAI Conference on Artificial Intelligence , volume=

Multi-source domain adaptation: A causal view , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=
[67]

International conference on machine learning , pages=

Domain adaptation with conditional transferable components , author=. International conference on machine learning , pages=. 2016 , organization=

2016
[68]

arXiv preprint arXiv:1907.02893 , year=

Invariant risk minimization , author=. arXiv preprint arXiv:1907.02893 , year=

1907 arXiv
[69]

IEEE transactions on neural networks , volume=

An overview of statistical learning theory , author=. IEEE transactions on neural networks , volume=. 1999 , publisher=

1999
[70]

Proceedings of the IEEE/CVF international conference on computer vision , pages=

Selfreg: Self-supervised contrastive regularization for domain generalization , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=
[71]

Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

Reducing domain gap by reducing style bias , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=
[72]

arXiv preprint arXiv:2104.09937 , year=

Gradient matching for domain generalization , author=. arXiv preprint arXiv:2104.09937 , year=

arXiv
[73]

Proceedings of the AAAI conference on artificial intelligence , volume=

Learning to generalize: Meta-learning for domain generalization , author=. Proceedings of the AAAI conference on artificial intelligence , volume=
[74]

Journal of machine learning research , volume=

Domain generalization by marginal transfer learning , author=. Journal of machine learning research , volume=
[75]

European conference on computer vision , pages=

Self-challenging improves cross-domain generalization , author=. European conference on computer vision , pages=. 2020 , organization=

2020
[76]

International conference on machine learning , pages=

Out-of-distribution generalization via risk extrapolation (rex) , author=. International conference on machine learning , pages=. 2021 , organization=

2021
[77]

Advances in neural information processing systems , volume=

Adaptive risk minimization: Learning to adapt to domain shift , author=. Advances in neural information processing systems , volume=
[78]

arXiv preprint arXiv:1911.08731 , year=

Distributionally robust neural networks for group shifts: On the importance of regularization for worst-case generalization , author=. arXiv preprint arXiv:1911.08731 , year=

1911 arXiv
[79]

arXiv preprint arXiv:2104.02008 , year=

Domain generalization with mixstyle , author=. arXiv preprint arXiv:2104.02008 , year=

arXiv
[80]

Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

Domain generalization with adversarial feature learning , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=
[81]

IEEE transactions on pattern analysis and machine intelligence , volume=

Domain generalization: A survey , author=. IEEE transactions on pattern analysis and machine intelligence , volume=. 2022 , publisher=

2022
[82]

Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

Robustnet: Improving domain generalization in urban-scene segmentation via instance selective whitening , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=
[83]

Proceedings of the European conference on computer vision (ECCV) , pages=

Encoder-decoder with atrous separable convolution for semantic image segmentation , author=. Proceedings of the European conference on computer vision (ECCV) , pages=
[84]

Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

The cityscapes dataset for semantic urban scene understanding , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=
[85]

Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

Bdd100k: A diverse driving dataset for heterogeneous multitask learning , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=
[86]

Proceedings of the IEEE international conference on computer vision , pages=

The mapillary vistas dataset for semantic understanding of street scenes , author=. Proceedings of the IEEE international conference on computer vision , pages=
[87]

European conference on computer vision , pages=

Playing for data: Ground truth from computer games , author=. European conference on computer vision , pages=. 2016 , organization=

2016
[88]

Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

The synthia dataset: A large collection of synthetic images for semantic segmentation of urban scenes , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=
[89]

2022 IEEE International Symposium on Information Theory (ISIT) , pages=

Empirical risk minimization with relative entropy regularization: Optimality and sensitivity analysis , author=. 2022 IEEE International Symposium on Information Theory (ISIT) , pages=. 2022 , organization=

2022
[90]

International conference on machine learning , pages=

Domain adaptation under target and conditional shift , author=. International conference on machine learning , pages=. 2013 , organization=

2013
[91]

Proceedings of the european conference on computer vision (ECCV) , pages=

Two at once: Enhancing learning and generalization capacities via ibn-net , author=. Proceedings of the european conference on computer vision (ECCV) , pages=
[92]

arXiv preprint arXiv:2302.06556 , year=

Va-depthnet: A variational approach to single image depth prediction , author=. arXiv preprint arXiv:2302.06556 , year=

arXiv
[93]

European conference on computer vision , pages=

Indoor segmentation and support inference from rgbd images , author=. European conference on computer vision , pages=. 2012 , organization=

2012
[94]

Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

Style neophile: Constantly seeking novel styles for domain generalization , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=
[95]

Proceedings of the IEEE/CVF international conference on computer vision , pages=

Swin transformer: Hierarchical vision transformer using shifted windows , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=
[96]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Semantic-aware domain generalized segmentation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[97]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

WildNet: Learning domain generalized semantic segmentation from the wild , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[98]

Advances in Neural Information Processing Systems , volume=

Diverse weight averaging for out-of-distribution generalization , author=. Advances in Neural Information Processing Systems , volume=
[99]

arXiv preprint arXiv:2004.08697 , year=

Causalvae: Structured causal disentanglement in variational autoencoder , author=. arXiv preprint arXiv:2004.08697 , year=

2004 arXiv
[100]

arXiv preprint arXiv:2207.03162 , year=

Harnessing out-of-distribution examples via augmenting content and style , author=. arXiv preprint arXiv:2207.03162 , year=

arXiv
[101]

arXiv preprint arXiv:2310.09753 , year=

When can transformers reason with abstract symbols? , author=. arXiv preprint arXiv:2310.09753 , year=

arXiv
[102]

Entropy , volume=

Seeing is believing: Brain-inspired modular training for mechanistic interpretability , author=. Entropy , volume=. 2023 , publisher=

2023
[103]

Proceedings of the AAAI Conference on Artificial Intelligence , volume=

Entropy-based logic explanations of neural networks , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=
[104]

arXiv preprint arXiv:2402.04875 , year=

On Provable Length and Compositional Generalization , author=. arXiv preprint arXiv:2402.04875 , year=

arXiv
[105]

Advances in Neural Information Processing Systems , volume=

Provable advantage of curriculum learning on parity targets with mixed inputs , author=. Advances in Neural Information Processing Systems , volume=
[106]

arXiv preprint arXiv:2404.00560 , year=

A Theory for Length Generalization in Learning to Reason , author=. arXiv preprint arXiv:2404.00560 , year=

arXiv
[107]

Advances in Neural Information Processing Systems , volume=

Neuro-symbolic Learning Yielding Logical Constraints , author=. Advances in Neural Information Processing Systems , volume=
[108]

Proceedings of the AAAI conference on artificial intelligence , volume=

Beyond sparsity: Tree regularization of deep models for interpretability , author=. Proceedings of the AAAI conference on artificial intelligence , volume=
[109]

arXiv preprint arXiv:2211.08686 , year=

Improving Interpretability via Regularization of Neural Activation Sensitivity , author=. arXiv preprint arXiv:2211.08686 , year=

arXiv
[110]

IEEE/ACM Transactions on Audio, Speech, and Language Processing , volume=

Improving interpretability and regularization in deep learning , author=. IEEE/ACM Transactions on Audio, Speech, and Language Processing , volume=. 2017 , publisher=

2017
[111]

Proceedings of the AAAI conference on artificial intelligence , volume=

Regional tree regularization for interpretability in deep neural networks , author=. Proceedings of the AAAI conference on artificial intelligence , volume=
[112]

IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems , volume=

Circuitnet: An open-source dataset for machine learning in vlsi cad applications with improved domain-specific evaluation metric and learning strategies , author=. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems , volume=. 2023 , publisher=

2023
[113]

Proceedings of the eleventh annual conference on Computational learning theory , pages=

Learning atomic formulas with prescribed properties , author=. Proceedings of the eleventh annual conference on Computational learning theory , pages=
[114]

Advances in Neural Information Processing Systems , volume=

Circuit as set of points , author=. Advances in Neural Information Processing Systems , volume=
[115]

Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

Sparse linear concept discovery models , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=
[116]

IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

Class-incremental learning: survey and performance evaluation on image classification , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2022 , publisher=

2022
[117]

NeurIPS Continual Learning Workshop , volume=

Three continual learning scenarios , author=. NeurIPS Continual Learning Workshop , volume=
[118]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Prototype augmentation and self-supervision for incremental learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[119]

Advances in Neural Information Processing Systems , volume=

Class-incremental learning via dual augmentation , author=. Advances in Neural Information Processing Systems , volume=
[120]

Advances in neural information processing systems , volume=

Imagenet classification with deep convolutional neural networks , author=. Advances in neural information processing systems , volume=
[121]

Psychology of learning and motivation , volume=

Catastrophic interference in connectionist networks: The sequential learning problem , author=. Psychology of learning and motivation , volume=. 1989 , publisher=

1989
[122]

arXiv preprint arXiv:1312.6211 , year=

An empirical investigation of catastrophic forgetting in gradient-based neural networks , author=. arXiv preprint arXiv:1312.6211 , year=

arXiv
[123]

Information fusion , volume=

Continual learning for robotics: Definition, framework, learning strategies, opportunities and challenges , author=. Information fusion , volume=. 2020 , publisher=

2020
[124]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Few-shot class-incremental learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[125]

arXiv preprint arXiv:2304.08130 , year=

A survey on few-shot class-incremental learning , author=. arXiv preprint arXiv:2304.08130 , year=

arXiv
[126]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Few-shot incremental learning with continually evolved classifiers , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[127]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Semantic-aware knowledge distillation for few-shot class-incremental learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[128]

Advances in neural information processing systems , volume=

Margin-based few-shot class-incremental learning with class-level overfitting mitigation , author=. Advances in neural information processing systems , volume=
[129]

IEEE Transactions on Pattern Analysis and Machine Intelligence , year=

Few-shot class-incremental learning by sampling multi-phase tasks , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , year=
[130]

arXiv preprint arXiv:2302.03004 , year=

Neural collapse inspired feature-classifier alignment for few-shot class incremental learning , author=. arXiv preprint arXiv:2302.03004 , year=

arXiv
[131]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Constrained few-shot class-incremental learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[132]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Forward compatible few-shot class-incremental learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[133]

Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=

icarl: Incremental classifier and representation learning , author=. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=
[134]

Advances in Neural Information Processing Systems , volume=

Experience replay for continual learning , author=. Advances in Neural Information Processing Systems , volume=
[135]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Class-incremental learning by knowledge distillation with adaptive feature consolidation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[136]

The Eleventh International Conference on Learning Representations , year=

Warping the space: Weight space rotation for class-incremental few-shot learning , author=. The Eleventh International Conference on Learning Representations , year=
[137]

International Conference on Learning Representations (ICLR) 2023 , year=

On the Soft-Subnetwork for Few-Shot Class Incremental Learning , author=. International Conference on Learning Representations (ICLR) 2023 , year=

2023
[138]

The Eleventh International Conference on Learning Representations , year=

A Model or 603 Exemplars: Towards Memory-Efficient Class-Incremental Learning , author=. The Eleventh International Conference on Learning Representations , year=
[139]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Der: Dynamically expandable representation for class incremental learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[140]

IEEE transactions on pattern analysis and machine intelligence , volume=

Learning without forgetting , author=. IEEE transactions on pattern analysis and machine intelligence , volume=. 2017 , publisher=

2017
[141]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Large scale incremental learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[142]

Proceedings of the European conference on computer vision (ECCV) , pages=

End-to-end incremental learning , author=. Proceedings of the European conference on computer vision (ECCV) , pages=
[143]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Learning a unified classifier incrementally via rebalancing , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[144]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Metafscil: A meta-learning approach for few-shot class incremental learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[145]

International conference on machine learning , pages=

Model-agnostic meta-learning for fast adaptation of deep networks , author=. International conference on machine learning , pages=. 2017 , organization=

2017
[146]

Advances in Neural Information Processing Systems , volume=

Implicit semantic data augmentation for deep networks , author=. Advances in Neural Information Processing Systems , volume=
[147]

Proceedings of the AAAI Conference on Artificial Intelligence , volume=

Imagine by reasoning: A reasoning-based implicit semantic data augmentation for long-tailed classification , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=
[148]

IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

Bridging the gap between few-shot and many-shot learning via distribution calibration , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2021 , publisher=

2021
[149]

Advances in Neural Information Processing Systems , volume=

Adaptive distribution calibration for few-shot learning with hierarchical optimal transport , author=. Advances in Neural Information Processing Systems , volume=
[150]

arXiv preprint arXiv:2110.07059 , year=

Subspace regularizers for few-shot class incremental learning , author=. arXiv preprint arXiv:2110.07059 , year=

arXiv
[151]

Advances in Neural Information Processing Systems , volume=

A theoretical study on solving continual learning , author=. Advances in Neural Information Processing Systems , volume=
[152]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Learning with Fantasy: Semantic-Aware Virtual Contrastive Constraint for Few-Shot Class-Incremental Learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[153]

International journal of computer vision , volume=

Imagenet large scale visual recognition challenge , author=. International journal of computer vision , volume=. 2015 , publisher=

2015
[154]

European Conference on Computer Vision , pages=

Few-shot class-incremental learning from an open-set perspective , author=. European Conference on Computer Vision , pages=. 2022 , organization=

2022
[155]

European Conference on Computer Vision , pages=

Few-shot class-incremental learning via entropy-regularized data-free replay , author=. European Conference on Computer Vision , pages=. 2022 , organization=

2022
[156]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Self-promoted prototype refinement for few-shot class-incremental learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[157]

Advances in neural information processing systems , volume=

Matching networks for one shot learning , author=. Advances in neural information processing systems , volume=
[158]

arXiv e-prints , pages=

Do we really need a learnable classifier at the end of deep neural network? , author=. arXiv e-prints , pages=
[159]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Few-Shot Class-Incremental Learning via Class-Aware Bilateral Distillation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[160]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Deepemd: Few-shot image classification with differentiable earth mover's distance and structured classifiers , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[161]

fast framework for few-shot class-incremental learning , author=

Mgsvf: Multi-grained slow vs. fast framework for few-shot class-incremental learning , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , year=
[162]

IEEE Transactions on Pattern Analysis and Machine Intelligence , year=

Learnable Distribution Calibration for Few-Shot Class-Incremental Learning , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , year=
[163]

Proceedings of the 30th ACM International Conference on Multimedia , pages=

Semantics-Driven Generative Replay for Few-Shot Class Incremental Learning , author=. Proceedings of the 30th ACM International Conference on Multimedia , pages=
[164]

Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

Generalized and incremental few-shot learning by explicit learning and calibration without forgetting , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=
[165]

Proceedings of ICML Workshop on Unsupervised and Transfer Learning , pages=

One-shot learning with a hierarchical nonparametric bayesian model , author=. Proceedings of ICML Workshop on Unsupervised and Transfer Learning , pages=. 2012 , organization=

2012
[166]

Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

Synthesized feature based few-shot class-incremental learning on a mixture of subspaces , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=
[167]

Advances in neural information processing systems , volume=

Supervised contrastive learning , author=. Advances in neural information processing systems , volume=
[168]

Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XIV 16 , pages=

Rethinking few-shot image classification: a good embedding is all you need? , author=. Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XIV 16 , pages=. 2020 , organization=

2020
[169]

Proceedings of the AAAI conference on artificial intelligence , volume=

Hybrid graph neural networks for few-shot learning , author=. Proceedings of the AAAI conference on artificial intelligence , volume=
[170]

Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=

Dynamic few-shot visual learning without forgetting , author=. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=
[171]

Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=

Low-shot learning with imprinted weights , author=. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=
[172]

arXiv preprint arXiv:2209.07383 , year=

Visual recognition with deep nearest centroids , author=. arXiv preprint arXiv:2209.07383 , year=

arXiv
[173]

IEEE Transactions on Circuits and Systems for Video Technology , volume=

Meta-learning-based incremental few-shot object detection , author=. IEEE Transactions on Circuits and Systems for Video Technology , volume=. 2021 , publisher=

2021
[174]

arXiv preprint arXiv:2101.06395 , year=

Free lunch for few-shot learning: Distribution calibration , author=. arXiv preprint arXiv:2101.06395 , year=

arXiv
[175]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Deep representation learning on long-tailed data: A learnable embedding augmentation perspective , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[176]

Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part IV 16 , pages=

Negative margin matters: Understanding margin in few-shot classification , author=. Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part IV 16 , pages=. 2020 , organization=

2020
[177]

Advances in Neural Information Processing Systems , volume=

Few-shot class-incremental learning via training-free prototype calibration , author=. Advances in Neural Information Processing Systems , volume=
[178]

Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=

Cosface: Large margin cosine loss for deep face recognition , author=. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=
[179]

IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

Convolutional prototype network for open set recognition , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2020 , publisher=

2020
[180]

Advances in neural information processing systems , volume=

Prototypical networks for few-shot learning , author=. Advances in neural information processing systems , volume=
[181]

Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

On the robustness of open-world test-time training: Self-training with dynamic prototype expansion , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=
[182]

arXiv preprint arXiv:2102.02038 , year=

Isometric propagation network for generalized zero-shot learning , author=. arXiv preprint arXiv:2102.02038 , year=

arXiv
[183]

Advances in Neural Information Processing Systems , volume=

Attribute prototype network for zero-shot learning , author=. Advances in Neural Information Processing Systems , volume=
[184]

Advances in neural information processing systems , volume=

Attention is all you need , author=. Advances in neural information processing systems , volume=
[185]

arXiv preprint arXiv:1312.6114 , year=

Auto-encoding variational bayes , author=. arXiv preprint arXiv:1312.6114 , year=

arXiv
[186]

Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

Variational few-shot learning , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=
[187]

International conference on machine learning , pages=

Manifold mixup: Better representations by interpolating hidden states , author=. International conference on machine learning , pages=. 2019 , organization=

2019
[188]

TMLR , year=

Variational Classification , author=. TMLR , year=
[189]

Advances in neural information processing systems , volume=

Semi-supervised learning with deep generative models , author=. Advances in neural information processing systems , volume=
[190]

Pattern Recognition , volume=

SaliencyCut: Augmenting plausible anomalies for anomaly detection , author=. Pattern Recognition , volume=. 2024 , publisher=

2024
[191]

Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , pages=

Robust Feature Learning and Global Variance-Driven Classifier Alignment for Long-Tail Class Incremental Learning , author=. Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , pages=
[192]

European Conference on Computer Vision , pages=

Long-tailed class incremental learning , author=. European Conference on Computer Vision , pages=. 2022 , organization=

2022
[193]

IEEE Transactions on Pattern Analysis and Machine Intelligence , year=

Class-incremental learning: A survey , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , year=
[194]

arXiv preprint arXiv:2310.02255 , year=

Mathvista: Evaluating mathematical reasoning of foundation models in visual contexts , author=. arXiv preprint arXiv:2310.02255 , year=

arXiv
[195]

arXiv preprint arXiv:2408.04591 , year=

Hilo: A learning framework for generalized category discovery robust to domain shifts , author=. arXiv preprint arXiv:2408.04591 , year=

arXiv
[196]

European conference on computer vision , pages=

Promptccd: Learning gaussian mixture prompt pool for continual category discovery , author=. European conference on computer vision , pages=. 2024 , organization=

2024
[197]

Advances in Neural Information Processing Systems , volume=

Grow and merge: A unified framework for continuous categories discovery , author=. Advances in Neural Information Processing Systems , volume=
[198]

Proceedings of the IEEE/CVF international conference on computer vision , pages=

Proxy anchor-based unsupervised learning for continuous generalized category discovery , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=
[199]

Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

Metagcd: Learning to continually learn in generalized category discovery , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=
[200]

The Twelfth International Conference on Learning Representations , year=

SPTNet: An Efficient Alternative Framework for Generalized Category Discovery with Spatial Prompt Tuning , author=. The Twelfth International Conference on Learning Representations , year=
[201]

arXiv preprint arXiv:2304.07193 , year=

Dinov2: Learning robust visual features without supervision , author=. arXiv preprint arXiv:2304.07193 , year=

arXiv
[202]

2021 , publisher=

Open-set recognition: A good closed-set classifier is all you need? , author=. 2021 , publisher=

2021
[203]

European Conference on Computer Vision , pages=

Selex: Self-expertise in fine-grained generalized category discovery , author=. European Conference on Computer Vision , pages=. 2024 , organization=

2024
[204]

Advances in Neural Information Processing Systems , volume=

Flipped classroom: Aligning teacher attention with student in generalized category discovery , author=. Advances in Neural Information Processing Systems , volume=
[205]

2021 , publisher=

PyTorch library for CAM methods , author=. 2021 , publisher=

2021
[206]

Proceedings of the IEEE international conference on computer vision , pages=

Grad-cam: Visual explanations from deep networks via gradient-based localization , author=. Proceedings of the IEEE international conference on computer vision , pages=
[207]

Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

Masked autoencoders are scalable vision learners , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=
[208]

arXiv preprint arXiv:2409.17063 , year=

Benchmarking domain generalization algorithms in computational pathology , author=. arXiv preprint arXiv:2409.17063 , year=
[209]

Scientific reports , volume=

Multi-class texture analysis in colorectal cancer histology , author=. Scientific reports , volume=. 2016 , publisher=

2016
[210]

PLoS medicine , volume=

Predicting survival from colorectal cancer histology slides using deep learning: A retrospective multicenter study , author=. PLoS medicine , volume=. 2019 , publisher=

2019
[211]

Medical image analysis , volume=

Cellular community detection for tissue phenotyping in colorectal cancer histology images , author=. Medical image analysis , volume=. 2020 , publisher=

2020
[212]

International conference on machine learning , pages=

Learning transferable visual models from natural language supervision , author=. International conference on machine learning , pages=. 2021 , organization=

2021
[213]

2009 , publisher=

Learning multiple layers of features from tiny images , author=. 2009 , publisher=

2009
[214]

Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

Gradient estimation for unseen domain risk minimization with pre-trained models , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=
[215]

Learning a Cross-Modal Schr

Zheng, Hao and Yi, Jingjun and Bi, Qi and Huang, Huimin and Zhan, Haolan and Huang, Yawen and Li, Yuexiang and Wu, Xian and Zheng, Yefeng , booktitle=. Learning a Cross-Modal Schr
[216]

Advances in Neural Information Processing Systems , volume=

Ensemble of averages: Improving model selection and boosting performance in domain generalization , author=. Advances in Neural Information Processing Systems , volume=
[217]

arXiv preprint arXiv:2209.14926 , year=

Domain-unified prompt representations for source-free domain generalization , author=. arXiv preprint arXiv:2209.14926 , year=

arXiv
[218]

International journal of computer vision , volume=

Learning to prompt for vision-language models , author=. International journal of computer vision , volume=. 2022 , publisher=

2022
[219]

arXiv preprint arXiv:2203.04600 , year=

Domain generalization using pretrained models without fine-tuning , author=. arXiv preprint arXiv:2203.04600 , year=

arXiv
[220]

Transactions of the Japanese Society for Artificial Intelligence , volume=

Domain prompt learning for efficiently adapting clip to unseen domains , author=. Transactions of the Japanese Society for Artificial Intelligence , volume=. 2023 , publisher=

2023
[221]

International conference on machine learning , pages=

Clipood: Generalizing clip to out-of-distributions , author=. International conference on machine learning , pages=. 2023 , organization=

2023
[222]

Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

Promptstyler: Prompt-driven style generation for source-free domain generalization , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=
[223]

2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) , pages=

Domain generalization using large pretrained models with mixture-of-adapters , author=. 2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) , pages=. 2025 , organization=

2025
[224]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Disentangled prompt representation for domain generalization , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
[225]

Advances in Neural Information Processing Systems , volume=

Clipceil: Domain generalization through clip via channel refinement and image-text alignment , author=. Advances in Neural Information Processing Systems , volume=
[226]

International Journal of Computer Vision , volume=

A Causal Intervention Method for Domain Generalization with a Self-Supervised Auxiliary Task , author=. International Journal of Computer Vision , volume=. 2025 , publisher=

2025
[227]

International conference on machine learning , pages=

Training data-efficient image transformers & distillation through attention , author=. International conference on machine learning , pages=. 2021 , organization=

2021
[228]

Proceedings of the 33rd ACM International Conference on Multimedia , pages=

CauRDG: Enhancing Domain Generalization with Causal-Driven Semantic Consistency Reasoning , author=. Proceedings of the 33rd ACM International Conference on Multimedia , pages=
[229]

2024 IEEE International Conference on Multimedia and Expo (ICME) , pages=

Mix background and foreground separately: Transformer-based Augmentation Strategies for Domain Generalization , author=. 2024 IEEE International Conference on Multimedia and Expo (ICME) , pages=. 2024 , organization=

2024
[230]

IEEE Transactions on Image Processing , year=

Exploring invariance matters for domain generalization , author=. IEEE Transactions on Image Processing , year=
[231]

arXiv preprint arXiv:2508.07577 , year=

Exploiting Layer Normalization Fine-tuning in Visual Transformer Foundation Models for Classification , author=. arXiv preprint arXiv:2508.07577 , year=

arXiv
[232]

2023 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW) , pages=

On the effectiveness of layernorm tuning for continual learning in vision transformers , author=. 2023 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW) , pages=. 2023 , organization=

2023
[233]

IEEE Transactions on Information Theory , volume=

Rademacher penalties and structural risk minimization , author=. IEEE Transactions on Information Theory , volume=. 2002 , publisher=

2002
[234]

Causal image modeling for efficient visual understanding , author=
[235]

Advances in neural information processing systems , volume=

Dynamicvit: Efficient vision transformers with dynamic token sparsification , author=. Advances in neural information processing systems , volume=
[236]

arXiv preprint arXiv:2202.07800 , year=

Not all patches are what you need: Expediting vision transformers via token reorganizations , author=. arXiv preprint arXiv:2202.07800 , year=

arXiv
[237]

Proceedings of the 58th annual meeting of the association for computational linguistics , pages=

Quantifying attention flow in transformers , author=. Proceedings of the 58th annual meeting of the association for computational linguistics , pages=

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

International Journal of Computer Vision , volume=

Incremental model enhancement via memory-based contrastive learning , author=. International Journal of Computer Vision , volume=. 2025 , publisher=

2025

[2] [2]

International Journal of Computer Vision , volume=

Domain-specific bias filtering for single labeled domain generalization , author=. International Journal of Computer Vision , volume=. 2023 , publisher=

2023

[3] [3]

Neural computation , volume=

Long short-term memory , author=. Neural computation , volume=. 1997 , publisher=

1997

[4] [4]

Advances in Neural Information Processing Systems , volume=

Interpret your decision: Logical reasoning regularization for generalization in visual classification , author=. Advances in Neural Information Processing Systems , volume=

[5] [5]

Journal of Machine Learning Research , volume=

Generalization on the unseen, logic reasoning and degree curriculum , author=. Journal of Machine Learning Research , volume=

[6] [6]

2022 , publisher=

Universal algebraic logic: dedicated to the unity of science , author=. 2022 , publisher=

2022

[7] [7]

Artificial Intelligence , volume=

Conditional logic and the principle of entropy , author=. Artificial Intelligence , volume=. 2000 , publisher=

2000

[8] [8]

proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

Rethinking multi-domain generalization with a general learning objective , author=. proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=