Pith. sign in

REVIEW 3 major objections 6 minor 66 references

Breaking Diversity Collapse in Spiking Pseudo-Ensembles for Efficient OOD Detection in Remote Sensing

T0 review · 3 major / 6 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 A frozen spiking backbone fitted with five lightweight heads, trained to agree on clean images and disagree on box-blurred ones, matches a five-model deep ensemble for remote-sensing out-of-distribution detection while using roughly 38%…

desk verdict Solid SNN pseudo-ensemble efficiency result, but the headline UCM comparison is partly tuned and needs a held-out validation split. read the letter →

arxiv 2608.01090 v1 pith:222CCCGL submitted 2026-08-02 cs.LG

Srinivas Anumasa , Rushi Shah , Qiran Zou , Dianbo Liu This is my paper · ORCID
classification cs.LG
keywords spikingneuralnetworksout-of-distributiondetectionpseudo-ensembledeepensemblediversitycollapseagree-disagreeobjectiveboxblurremotesensing
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

Spiking neural networks promise low-cost inference for satellites and edge devices, but they lack reliable out-of-distribution (OOD) detection. The paper proposes a cheap substitute for deep ensembles: freeze one pretrained SNN backbone, attach several lightweight classification heads, and train those heads so they agree on clean images but deliberately disagree on box-blurred versions of the same images. This agree-disagree training prevents diversity collapse, the tendency of heads trained independently with cross-entropy to converge to correlated predictions. With three backbones and five heads each on EuroSAT as in-distribution data, the resulting ADPE(3,5) matches or improves the OOD detection of a five-model deep ensemble on UCM and AID while using about 38% fewer parameters and 40% fewer backbone evaluations. If correct, this would give resource-constrained remote-sensing systems ensemble-quality uncertainty at a fraction of the deployment cost.

What carries the argument

The load-bearing object is the agree-disagree spiking pseudo-ensemble: a frozen SNN backbone (Spikformer or ResNet19-SNN) with $K_h$ lightweight spiking heads broadcast to the same temporal features, trained under $L_{\mathrm{AD}}=L_{\mathrm{CE}}-\lambda_{\mathrm{dis}}D_{\mathrm{JS}}$ with $\lambda_{\mathrm{dis}}=0.3$. The disagreement signal is the generalized Jensen-Shannon divergence $$D_{\mathrm{JS}} = \frac{1}{K_h|\mathcal{I}_b|} \sum_{i\in\mathcal{I}_b} \sum_{k=1}^{K_h} \mathrm{KL}(p_k(\tilde{x}_i) \| \bar{p}(\tilde{x}_i)),$$ computed only on box-blurred inputs $\tilde{x}_i$, with kernel size sampled uniformly from $\{5,7,9,11\}$ and $30\%$ of each minibatch blurred. Clean images contribute only cross-entropy, so every head keeps solving the original classification task; blurred images carry no labels and only push the heads toward different predictive distributions. A diagnostic with a five-model Spikformer ensemble selects box blur because medium and large blur kernels produce the strongest rise in ensemble entropy and mutual information while preserving coarse scene structure. Because the backbone is frozen and evaluated once per input, the added cost of more heads is small: five heads add roughly $4\%$ parameter overhead per backbone.

What would settle it

Re-run ADPE training with the disagreement term evaluated on additive Gaussian noise instead of box blur, holding all other hyperparameters fixed; if the reported UCM and AID AUROC advantages over CEPE disappear or reverse, then the structure-preserving property of box blur carries the claimed diversity transfer.

Watch

Extended reading notes

Core claim

The paper's central claim is that shared-backbone pseudo-ensemble failure is not inevitable. A frozen spiking backbone's time-resolved features can be broadcast to several independently initialized spiking heads, but when those heads are trained with cross-entropy alone they suffer diversity collapse: every head sees the same representation, labels, and objective, so their predictions become correlated. The paper replaces that training with an agree-disagree objective, $$L_{\mathrm{AD}} = L_{\mathrm{CE}} - \lambda_{\mathrm{dis}} D_{\mathrm{JS}}, \quad \lambda_{\mathrm{dis}} = 0.3,$$ where $L_{\mathrm{CE}}$ applies cross-entropy to clean in-distribution images and $D_{\mathrm{JS}}$ is a generalized Jensen-Shannon divergence among head predictions on box-blurred copies of the same images. Because blurred inputs are generated from the training set, the diversity signal requires no external outlier data or teacher ensemble. The paper reports that ADPE(3,5) matches or exceeds a five-model deep ensemble on UCM and AID with roughly 38% fewer parameters and 40% fewer backbone evaluations, and that it consistently improves MSP (maximum-softmax-probability)- and mutual-information-based OOD detection over cross-entropy pseudo-ensembles for both Spikformer and ResNet19-SNN.

Load-bearing premise

The load-bearing premise, acknowledged in the paper when selecting disagreement inputs, is that disagreement induced on box-blurred in-distribution images transfers to real distribution shifts; the authors state that blurred samples 'are not assumed to follow the true test-time OOD distribution.'

Editorial extensions

If this is right

  • A single frozen backbone with five heads already improves MSP-based OOD AUROC over one deterministic model (for Spikformer, from 89.53% to 94.82% on UCM), so within-backbone diversity is itself a usable source of uncertainty.
  • ADPE(3,5) needs three backbone evaluations and about 3.12 full-model parameter equivalents instead of five, so ensemble-like OOD detection is available on memory- and energy-limited hardware.
  • The disagreement objective improves over cross-entropy-trained pseudo-ensembles for both a Transformer-based and a convolutional SNN backbone, indicating the mechanism is not tied to one architecture.
  • Detection remains harder for semantically close shifts (Global-Near) than for far shifts (Global-Far), so the method improves but does not remove the difficulty of near-OOD acceptance.

Reading between the lines

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

  • The paper does not test shifts that alter only high-level semantics while preserving texture statistics, such as rearranged land-cover configurations; on such shifts the blur-induced disagreement may be uninformative.
  • Because the loss needs only predictions and blurred inputs, the same agree-disagree recipe could be attached to non-spiking backbones or to larger head banks, but that extension is untested.
  • The reported FPR@95 is less monotonic than AUROC, so converting the AUROC gains into a reliable fixed operating threshold may require calibration or adaptive disagreement inputs.
  • The blur choice was made once from a diagnostic study; automatically selecting disagreement transformations per dataset could yield stronger or more robust diversity.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes Agree–Disagree Spiking Pseudo-Ensembles (ADPE) for out-of-distribution detection in remote sensing. A frozen SNN backbone is equipped with multiple lightweight spiking heads; the heads are trained to agree on clean in-distribution images and to disagree on box-blurred versions of the same images, via a Jensen–Shannon divergence term. The authors evaluate Spikformer and ResNet19-SNN backbones on EuroSAT as in-distribution data and UCM, AID, and two Sentinel-2 Global LULC subsets as OOD sets, comparing against deep ensembles, MC dropout, last-layer Laplace, Mahalanobis distance, and cross-entropy-trained pseudo-ensembles (CEPE). The headline result is that ADPE(3,5) (three backbones, five heads each) matches or improves a five-model deep ensemble on UCM and AID while using about 38% fewer parameters and 40% fewer backbone evaluations. The paper also reports consistent AUROC improvements over CEPE across both backbones and all four OOD datasets.

Significance. If the results hold, the method provides a practically useful efficiency–uncertainty trade-off for SNN deployment and introduces a clean, OOD-data-free diversity-promoting objective. The paper is unusually thorough on the experimental side: two architectures, four OOD datasets, multiple uncertainty metrics, and detailed ablations. The authors are also explicit about the key limitation—that blurred inputs are not assumed to match the true test-time OOD distribution—and they provide honest discussion of the mixed FPR@95 results. However, the central empirical claims rest on a few experimental-design choices that currently weaken the internal validity of the headline comparisons; these are fixable and do not invalidate the overall idea.

major comments (3)
  1. [Supplementary, Tables 8 and 9; Table 2] The hyperparameters Kh=5 and λ_dis=0.3 are selected by sweeping Kh and λ_dis on UCM FPR@95 (Tables 8 and 9, using ResNet19-SNN backbones with seeds 0,1,2). The main experiments, including the UCM rows of Table 2, then use this configuration for both backbones. Because UCM is one of the two datasets in the headline claim 'matches or improves a five-model deep ensemble on UCM and AID,' the UCM leg of that claim is optimistically biased and is not a clean out-of-sample evaluation. The reported flatness of the ablation surface mitigates the effect, but it does not remove the selection bias. I request that the authors either (a) perform model selection on a held-out validation split (e.g., a subset of EuroSAT with synthetic corruptions, or a separate OOD dataset such as one of the Global-LULC subsets) and then report the chosen configuration on UCM, or (b) explicitly quantify the selection bias by reporting how the selected configuration performs when the selection is repeated on different OOD targets.
  2. [Supplementary, 'Backbone and Pseudo-Ensemble Training'; Section 'Effect of Agree–Disagree Training'] The main text states that the only difference between CEPE and ADPE is the disagreement term on blurred inputs, but the supplementary reports that CEPE heads receive minibatches from their own independently shuffled data loaders, whereas ADPE uses a shared minibatch. This is a real confound: the two methods differ not only in the objective but also in the per-head data stochasticity. The shared-minibatch requirement is inherent to the JSD objective, but the authors should either implement a CEPE variant with a shared minibatch as an additional control, or explicitly discuss how the different data loading schemes could affect head diversity. Without this, the attribution of the observed improvements to the disagreement objective is not fully supported.
  3. [Table 2; Section 'Overall OOD Detection Performance'] The deep-ensemble baseline DE(5,1) is reported with a standard deviation of zero because only one subset of five backbones exists. Meanwhile, ADPE(3,5) reports means and standard deviations over ten backbone subsets. The claimed improvements over the deep ensemble on ResNet19-SNN are small (e.g., UCM AUROC 92.10 vs. 91.84; AID AUROC 93.12 vs. 92.79), and without a measure of the deep ensemble's variance across independent training seeds or a statistical test, it is unclear whether these differences are meaningful. The authors should provide either confidence intervals for the deep-ensemble baseline (e.g., by retraining multiple sets of five backbones with different seed sets) or a statistical comparison that accounts for the fact that only one ensemble was evaluated.
minor comments (6)
  1. [Table 2 caption] The caption says 'MSP-based OOD detection,' but the Mahalanobis row reports results from its native feature-distance score rather than MSP. Please clarify this in the caption so the row is not misinterpreted.
  2. [Table 1 and Eq. (10)] The main text uses Blur-S, Blur-M, and Blur-L in Table 1 without specifying the kernel sizes; the supplementary explains these are 5, 9, and 11. Please define the kernel sizes in the main text when the abbreviations are first introduced.
  3. [Section 'Agree–Disagree Training', Eq. (12)] The value λ_dis=0.3 is presented as a fixed design choice, but the supplementary shows it is the result of an ablation. Please state explicitly in the main text that this value was selected via the ablation and, if possible, cite the ablation tables there.
  4. [Section 'Inference and Computational Efficiency'] The parameter counts P_b≈12.5M and P_h≈0.1M are presented without a derivation; a brief note on how these numbers were obtained (e.g., from the architecture configurations in the supplementary) would aid reproducibility.
  5. [Supplementary, 'Backbone and Pseudo-Ensemble Training'] The supplementary describes CEPE as using separate optimizers and separate shuffling, but does not specify the random seed that controls the sharding; please clarify the seed handling for these per-head loaders.
  6. [Throughout] The paper does not provide a code-availability statement. Given the Reproducibility section's detail, a public code link or a statement of intention to release code would strengthen the manuscript.

Circularity Check

1 steps flagged · score 4.0 of 10

The ADPE objective and AID/Global-LULC evidence are self-contained, but the headline UCM comparison is contaminated because Kh and lambda_dis were selected using UCM FPR@95 and then reported on the same UCM set.

  1. fitted input called prediction [Supplementary Material, Hyperparameter Ablation, Tables 8 and 9; main paper Table 2]
    "We vary the number of heads Kh ∈ {2,3,4,5,6,7} and the disagreement weight λdis∈{0.1,0.2,0.3,0.4,0.5} for ResNet19-SNN. Every entry is the mean and sample standard deviation over deterministic-backbone seeds {0,1,2}. The setting used in the main experiments, Kh = 5 and λdis = 0.3, is shown in bold... Table 8: Ablation of the number of heads Kh and disagreement weight λdis on UCM using MSP-based FPR@95."

    The main claim that ADPE(3,5) matches or improves a five-model deep ensemble on UCM rests on the UCM rows of Table 2, including FPR@95. The supplementary sweep used exactly that UCM MSP-based FPR@95 metric to choose Kh=5 and λdis=0.3, which are then labeled 'the setting used in the main experiments.' Thus the reported UCM results are not an independent out-of-sample evaluation of the selected configuration: the same OOD set and metric used for model selection are later presented as a predicted outcome.

full rationale

The agree-disagree objective (Eq. 12) is not circular by construction: it maximizes generalized Jensen-Shannon divergence on box-blurred EuroSAT images, while the claimed OOD improvements are measured on UCM, AID, and Global-LULC, which are natural distribution shifts that never enter the training loss. The paper explicitly states that blurred samples 'are not assumed to follow the true test-time OOD distribution,' so the disagreement signal is an input-design heuristic rather than an encoding of the target evaluation on real OOD data. The diagnostic selection of blur as the transformation is performed with an independently trained deep ensemble and is not a parameter fitted to the OOD test labels. Self-citations (Anumasa et al. 2024; Bojkovic et al. 2024) are background references for direct SNN training and are not load-bearing, and no uniqueness theorem or ansatz is imported from the authors' prior work. The one substantive circularity-adjacent defect is the UCM-based hyperparameter selection described above, which partially compromises the UCM leg of the headline comparison but leaves the core derivation and the AID/Global-LULC results independent. A score of 4 reflects that partial contamination without claiming that any central equation reduces to its own input.

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

The central claim depends on the choice of blur as a disagreement input and on the weights/probabilities around that choice. These are data-dependent design choices, but they are not derived from first principles. No new physical entities are introduced. The ablation on UCM is a key reason the method is not fully held-out.

free parameters (4)
  • disagreement weight λ_dis = 0.3
    Weight on the generalized Jensen-Shannon divergence term in Eq. (12). Selected via hyperparameter ablation on UCM (Tables 8/9); the paper notes differences are within ~1.2 pp of best FPR@95.
  • blur probability ρ = 0.3
    Probability of applying box blur to each training image in the ADPE objective; chosen by hand (Section 'Agree-Disagree Training').
  • blur kernel sizes k = 5, 7, 9, 11
    Kernel sizes sampled uniformly for each blur-applied image (Eq. 10); selected based on diagnostic study (Table 1) showing medium/large blur induce high uncertainty. The set is a hand-chosen design decision.
  • number of heads Kh = 5
    Number of lightweight heads per backbone. Tested over {2,...,7} in ablation (Tables 8/9) with UCM FPR@95; chosen as a robust intermediate setting.
assumptions (4)
  • domain assumption The frozen backbone's temporal features Fθ(x) (Eq. 5) retain sufficient information for lightweight spiking heads to express functionally diverse predictive distributions.
    The whole ADPE architecture rests on this; if the backbone representation is exhausted by a single linear readout, multiple heads would have limited room for diversity.
  • ad hoc to paper Agreement on clean ID samples preserves ID classification accuracy while disagreement on box-blurred inputs transfers to improved OOD detection on real distribution shifts.
    The method's core hypothesis, explicitly stated in 'Selection of Structured Disagreement Inputs': blurred samples are 'not assumed to follow the true test-time OOD distribution' yet are used as the diversity signal. This is a modeling assumption tailored to the method.
  • standard math Surrogate-gradient training with T=2 simulation steps yields well-calibrated temporal features (standard in SNN literature, but assumed here).
    The paper relies on surrogate gradient training (Neftci et al. 2019) and Spikformer/ResNet19-SNN as backbones without re-deriving them.
  • domain assumption The OOD test distributions (UCM, AID, Global-LULC) are representative of distribution shifts encountered in remote sensing and are not used to fit the model (except the hyperparameter ablation on UCM).
    The evaluation assumes these datasets measure OOD detection ability. The hyperparameter ablation on UCM partially undermines this, hence this assumption is flagged.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Breaking Diversity Collapse in Spiking Pseudo-Ensembles for Efficient OOD Detection in Remote Sensing." pith.science (2026). https://pith.science/paper/222CCCGL

@misc{pith2026260801090,
  author       = {Pith},
  title        = {Pith review of: Breaking Diversity Collapse in Spiking Pseudo-Ensembles for Efficient OOD Detection in Remote Sensing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/222CCCGL}},
  note         = {Machine review of arXiv:2608.01090}
}
read the original abstract

Spiking Neural Networks (SNNs) are attractive for resource-constrained remote-sensing systems, but reliable out-of-distribution (OOD) detection remains challenging. Deep ensembles provide strong predictive uncertainty, yet require multiple complete models and backbone evaluations. We propose an efficient spiking pseudo-ensemble that attaches multiple lightweight classification heads to a frozen SNN backbone. Naively training these heads with cross-entropy can lead to diversity collapse, where independently parameterized heads may produce correlated predictions. To address this, we introduce an agree--disagree objective that preserves correct predictions on clean in-distribution samples while encouraging diversity on structured, uncertainty-inducing transformations of the same inputs. This provides a diversity-promoting training signal without requiring external OOD data. Experiments with Spikformer and ResNet19-SNN on EuroSAT demonstrate consistent improvements over conventionally trained pseudo-ensembles. Using three backbones with five heads each matches or improves upon a five-model deep ensemble on UCM and AID, while requiring approximately 38% fewer parameters and 40% fewer backbone evaluations. These results show that explicit diversity promotion can recover useful ensemble-style uncertainty at substantially lower deployment cost.

Figures

Figures reproduced from arXiv: 2608.01090 by the authors.

Figure 1
Figure 1. Multihead SNN OOD detection. A shared spik￾ing ResNet19 backbone produces temporal features F ∈ R B×T ×512 which are broadcast to Kh spike-consistent heads. Head predictions are aggregated to form the ensemble predictive distribution, from which MSP, predictive entropy, and mutual information (MI) are computed. content. A classifier may nevertheless produce a highly con￾fident prediction for such unfamiliar inputs. … view at source ↗

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

66 extracted references · 46 canonical work pages

  1. [1]

    Zur Elektrodynamik bewegter Körper

    Albert Einstein. Zur Elektrodynamik bewegter Körper. Annalen der Physik. 1905

    1905

  2. [2]

    The \ Companion

    Michel Goossens and Frank Mittelbach and Alexander Samarin. The \ Companion. 1993

    1993

  3. [3]

    Scaling Learning Algorithms Towards

    Bengio, Yoshua and LeCun, Yann , booktitle =. Scaling Learning Algorithms Towards

  4. [4]

    and Osindero, Simon and Teh, Yee Whye , journal =

    Hinton, Geoffrey E. and Osindero, Simon and Teh, Yee Whye , journal =. A Fast Learning Algorithm for Deep Belief Nets , volume =

  5. [5]

    2016 , publisher =

    Deep Learning , author =. 2016 , publisher =

    2016

  6. [6]

    IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , volume=

    Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification , author=. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , volume=. 2019 , publisher=

    2019

  7. [7]

    arXiv preprint arXiv:2202.11946 , year=

    Temporal efficient training of spiking neural network via gradient re-weighting , author=. arXiv preprint arXiv:2202.11946 , year=

    arXiv

  8. [8]

    Proceedings of the 18th SIGSPATIAL international conference on advances in geographic information systems , pages=

    Bag-of-visual-words and spatial extensions for land-use classification , author=. Proceedings of the 18th SIGSPATIAL international conference on advances in geographic information systems , pages=

Show all 66 references
  1. [9]

    Geomatics , year =

    Remote Sensing Image Scene Classification: Advances and Open Challenges , author =. Geomatics , year =

  2. [10]

    arXiv preprint arXiv:1610.02136 , year =

    A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks , author =. arXiv preprint arXiv:1610.02136 , year =

    arXiv

  3. [11]

    International Conference on Learning Representations (ICLR) , year =

    Enhancing the Reliability of Out-of-distribution Image Detection in Neural Networks , author =. International Conference on Learning Representations (ICLR) , year =

  4. [12]

    Advances in Neural Information Processing Systems (NeurIPS) , year =

    Energy-based Out-of-distribution Detection , author =. Advances in Neural Information Processing Systems (NeurIPS) , year =

  5. [13]

    International Conference on Machine Learning (ICML) , year =

    Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning , author =. International Conference on Machine Learning (ICML) , year =

  6. [14]

    Advances in Neural Information Processing Systems (NeurIPS) , year =

    Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles , author =. Advances in Neural Information Processing Systems (NeurIPS) , year =

  7. [15]

    Science , year =

    A million spiking-neuron integrated circuit with a scalable communication network and interface , author =. Science , year =

  8. [16]

    IEEE Micro , year =

    Loihi: A Neuromorphic Manycore Processor with On-Chip Learning , author =. IEEE Micro , year =

  9. [17]

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

    TrueNorth: Design and Tool Flow of a 65 mW 1 Million Neuron Programmable Neurosynaptic Chip , author =. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems , year =

  10. [18]

    Neural Networks , volume=

    Networks of spiking neurons: The third generation of neural network models , author=. Neural Networks , volume=

  11. [19]

    Science , volume=

    A million spiking-neuron integrated circuit with a scalable communication network and interface , author=. Science , volume=

  12. [20]

    International Conference on Artificial Intelligence and Statistics , pages=

    Data driven threshold and potential initialization for spiking neural networks , author=. International Conference on Artificial Intelligence and Statistics , pages=. 2024 , organization=

    2024

  13. [21]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Enhancing training of spiking neural network with stochastic latency , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

  14. [22]

    and Galluppi, Francesco and Temple, Steve and Plana, Luis A

    Furber, Steve B. and Galluppi, Francesco and Temple, Steve and Plana, Luis A. , journal=. The

  15. [23]

    Pehle, Christian and Billaudelle, Sebastian and Cramer, Benjamin and others , journal=. The

  16. [24]

    Neuronal Dynamics: From Single Neurons to Networks and Models of Cognition , author=

  17. [25]

    IEEE Signal Processing Magazine , volume=

    Surrogate Gradient Learning in Spiking Neural Networks: Bringing the Power of Gradient-Based Optimization to Spiking Neural Networks , author=. IEEE Signal Processing Magazine , volume=

  18. [26]

    Advances in Neural Information Processing Systems (NeurIPS) , year=

    SLAYER: Spike Layer Error Reassignment in Time , author=. Advances in Neural Information Processing Systems (NeurIPS) , year=

  19. [27]

    Advances in Neural Information Processing Systems (NeurIPS) , year=

    Deep Residual Learning in Spiking Neural Networks , author=. Advances in Neural Information Processing Systems (NeurIPS) , year=

  20. [28]

    International Joint Conference on Neural Networks (IJCNN) , year=

    Fast-classifying, high-accuracy spiking deep networks through weight and threshold balancing , author=. International Joint Conference on Neural Networks (IJCNN) , year=

  21. [29]

    Frontiers in Neuroscience , volume=

    Conversion of Continuous-Valued Deep Networks to Efficient Event-Driven Networks for Image Classification , author=. Frontiers in Neuroscience , volume=

  22. [30]

    International Conference on Learning Representations (ICLR) , year=

    Optimal Conversion of Conventional Artificial Neural Networks to Spiking Neural Networks , author=. International Conference on Learning Representations (ICLR) , year=

  23. [31]

    Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence (IJCAI) , year=

    Efficient and Accurate Conversion of Spiking Neural Network with Burst Spikes , author=. Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence (IJCAI) , year=

  24. [32]

    Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence (IJCAI) , year=

    A New ANN-SNN Conversion Method with High Accuracy, Low Latency and Good Robustness , author=. Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence (IJCAI) , year=

  25. [33]

    2024 , url=

    Jiang, Haotian and others , booktitle=. 2024 , url=

    2024

  26. [34]

    IEEE Transactions on Geoscience and Remote Sensing , volume=

    An advanced Dirichlet prior network for out-of-distribution detection in remote sensing , author=. IEEE Transactions on Geoscience and Remote Sensing , volume=. 2022 , publisher=

    2022

  27. [35]

    Remote Sensing , volume=

    Evaluation of ten deep-learning-based out-of-distribution detection methods for remote sensing image scene classification , author=. Remote Sensing , volume=. 2024 , publisher=

    2024

  28. [36]

    Advances in Neural Information Processing Systems , volume =

    Deep Residual Learning in Spiking Neural Networks , author =. Advances in Neural Information Processing Systems , volume =

  29. [37]

    arXiv preprint arXiv:2209.15425 , year=

    Spikformer: When spiking neural network meets transformer , author=. arXiv preprint arXiv:2209.15425 , year=

    arXiv

  30. [38]

    International Conference on Learning Representations , year =

    A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks , author =. International Conference on Learning Representations , year =

  31. [39]

    International Conference on Learning Representations , year =

    Enhancing the Reliability of Out-of-Distribution Image Detection in Neural Networks , author =. International Conference on Learning Representations , year =

  32. [40]

    Advances in Neural Information Processing Systems , volume =

    Energy-Based Out-of-Distribution Detection , author =. Advances in Neural Information Processing Systems , volume =

  33. [41]

    Advances in Neural Information Processing Systems , volume =

    A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attacks , author =. Advances in Neural Information Processing Systems , volume =

  34. [42]

    Advances in Neural Information Processing Systems , volume =

    Predictive Uncertainty Estimation via Prior Networks , author =. Advances in Neural Information Processing Systems , volume =

  35. [43]

    Advances in neural information processing systems , volume=

    Simple and scalable predictive uncertainty estimation using deep ensembles , author=. Advances in neural information processing systems , volume=

  36. [44]

    Information Fusion , volume=

    A novel Out-of-Distribution detection approach for Spiking Neural Networks: Design, fusion, performance evaluation and explainability , author=. Information Fusion , volume=. 2023 , publisher=

    2023

  37. [45]

    International Conference on Learning Representations , year =

    BatchEnsemble: An Alternative Approach to Efficient Ensemble and Lifelong Learning , author =. International Conference on Learning Representations , year =

  38. [46]

    International Conference on Learning Representations , year =

    Training Independent Subnetworks for Robust Prediction , author =. International Conference on Learning Representations , year =

  39. [47]

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

    Masksembles for Uncertainty Estimation , author =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages =

  40. [48]

    International Conference on Learning Representations , year =

    Packed-Ensembles for Efficient Uncertainty Estimation , author =. International Conference on Learning Representations , year =

  41. [49]

    arXiv preprint arXiv:2001.04694 , year =

    Hydra: Preserving Ensemble Diversity for Model Distillation , author =. arXiv preprint arXiv:2001.04694 , year =

    2001 arXiv

  42. [50]

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

    Sub-Ensembles for Fast Uncertainty Estimation in Neural Networks , author =. Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops , pages =

  43. [51]

    and Dong, Zhen and Wang, Haofan and Okuno, Tomoyuki and Nakata, Yohei and Keutzer, Kurt and Zhang, Shanghang , booktitle =

    Chen, Anthony and Yang, Huanrui and Gan, Yulu and Gudovskiy, Denis A. and Dong, Zhen and Wang, Haofan and Okuno, Tomoyuki and Nakata, Yohei and Keutzer, Kurt and Zhang, Shanghang , booktitle =. Split-Ensemble: Efficient. 2024 , publisher =

    2024

  44. [52]

    International Conference on Learning Representations , year =

    Agree to Disagree: Diversity through Disagreement for Better Transferability , author =. International Conference on Learning Representations , year =

  45. [53]

    Proceedings of the Thirty-Eighth Conference on Uncertainty in Artificial Intelligence , pages =

    Semi-Supervised Novelty Detection Using Ensembles with Regularized Disagreement , author =. Proceedings of the Thirty-Eighth Conference on Uncertainty in Artificial Intelligence , pages =. 2022 , editor =

    2022

  46. [54]

    International Conference on Learning Representations , year =

    Deep Anomaly Detection with Outlier Exposure , author =. International Conference on Learning Representations , year =

  47. [55]

    Proceedings of the 25th International Conference on Artificial Intelligence and Statistics , series =

    Embedded Ensembles: Infinite Width Limit and Operating Regimes , author =. Proceedings of the 25th International Conference on Artificial Intelligence and Statistics , series =. 2022 , publisher =

    2022

  48. [56]

    ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (ACM GIS) , year =

    Bag-of-Visual-Words and Spatial Extensions for Land-Use Classification , author =. ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (ACM GIS) , year =

  49. [57]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume =

    Going Deeper With Directly-Trained Larger Spiking Neural Networks , author =. Proceedings of the AAAI Conference on Artificial Intelligence , volume =. 2021 , doi =

    2021

  50. [58]

    Science Advances , volume=

    Spikingjelly: An open-source machine learning infrastructure platform for spike-based intelligence , author=. Science Advances , volume=. 2023 , publisher=

    2023

  51. [59]

    Proceedings of the 39th International Conference on Machine Learning , pages =

    Out-of-Distribution Detection with Deep Nearest Neighbors , author =. Proceedings of the 39th International Conference on Machine Learning , pages =. 2022 , volume =

    2022

  52. [60]

    Advances in neural information processing systems , volume=

    Laplace redux-effortless bayesian deep learning , author=. Advances in neural information processing systems , volume=

  53. [61]

    Horowitz, Mark , booktitle =. 1.1. 2014 , month =

    2014

  54. [62]

    arXiv preprint arXiv:2010.06610 , year=

    Training independent subnetworks for robust prediction , author=. arXiv preprint arXiv:2010.06610 , year=

    2010 arXiv

  55. [63]

    2025 , note =

    Proceedings of the. 2025 , note =

    2025

  56. [64]

    Proceedings of the 18th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems , pages =

    Yang, Yi and Newsam, Shawn , title =. Proceedings of the 18th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems , pages =. 2010 , publisher =

    2010

  57. [65]

    IEEE Transactions on Geoscience and Remote Sensing , volume =

    Xia, Gui-Song and Hu, Jingwen and Hu, Fan and Shi, Baoguang and Bai, Xiang and Zhong, Yanfei and Zhang, Liangpei and Lu, Xiaoqiang , title =. IEEE Transactions on Geoscience and Remote Sensing , volume =. 2017 , doi =

    2017

  58. [66]

    Scientific Data , volume =

    Benhammou, Yassir and Alcaraz-Segura, Domingo and Guirado, Emilio and Khaldi, Rohaifa and Achchab, Boujem. Scientific Data , volume =. 2022 , doi =

    2022

Pith tools

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