Pith. sign in

REVIEW 5 major objections 7 minor 37 references

Composite Gaussian Processes Flows for Learning Discontinuous Multimodal Policies

T0 review · 5 major / 7 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read CGP-Flows: using a multimodal Gaussian-process base lets continuous normalizing flows model discontinuous robot policies more accurately.

desk verdict Sensible, genuinely new GP-flow integration with solid real-robot results, but the headline simulation statistics are internally inconsistent and need correction before the significance claim is citable. read the letter →

arxiv 2502.01913 v1 pith:NCRRZZ7T submitted 2025-02-04 cs.RO cs.LG

Shu-yuan Wang , Hikaru Sasaki , Takamitsu Matsubara This is my paper · ORCID
classification cs.ROcs.LG
keywords multimodaldistributionslocaldiscontinuitiesnormalizingflowsGaussianprocessesbehaviorcloningNeuralODEroboticpolicylearningoverlappingmixtureof
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

This paper tries to establish that a normalizing flow used as a robotic control policy works better when its base distribution already contains the policy's multimodality. The authors combine an overlapping mixture of Gaussian processes (a sparse OMGP) with a continuous normalizing flow, so the flow only has to adjust an already multimodal distribution instead of creating multiple modes from a single Gaussian. In behavior-cloning experiments, the resulting CGP-Flows raise success from 56.3% to 81.0% on a simulated ball-shooting task whose mode count changes with the state, and from 50% to 85–90% on a real robot-grasping task; chi-square statistics are reported for the differences. If the claim is right, where a generative policy model places its multimodality matters as much as how expressive its flow is.

What carries the argument

The load-bearing object is the sparse OMGP base distribution feeding a Neural-ODE transformation. The sparse OMGP models the latent variable's density as a mixture of $M$ Gaussian-process experts built on pseudo-inputs, so the base density is already multimodal and state-dependent; the Neural-ODE integrates an ordinary differential equation to transport latent samples to actions and corrects the density by the trace of the Jacobian. The paper's mechanism is that a smaller disparity between base and target distribution lowers the ODE solver's workload, and the flow's remaining job is to refine boundaries and create modes the base lacks, which is why CGP-Flows can handle local discontinuities that a fixed mixture cannot represent.

What would settle it

Run the ball-shooting behavior-cloning experiment with the sparse-OMGP learning rate set equal to the Neural-ODE's learning rate and with pseudo-input counts $P$ varied from small to large: if the 81.0% success rate and low NFE persist, the sensitivity limitation is not load-bearing, while a large success-rate drop would show the reported advantage depends on that hand-tuned ratio rather than on the multimodal base itself.

Watch

Extended reading notes

Core claim

The central claim is that the expressiveness of the base distribution, not just the flow network, determines how well a conditional continuous normalizing flow models multimodal, locally discontinuous policies. CGP-Flows replace the unimodal GP base of NGGPs with a sparse overlapping mixture of Gaussian processes: $M$ GP experts each carry their own kernel, and each action is assigned to one expert through a latent indicator variable. A Neural-ODE then maps the latent variable to the action space, with the log-likelihood corrected by the instantaneous change-of-variables formula. Because the base is already multimodal, the ODE transformation has less work to do, which shows up as fewer solver function evaluations and sharper mode separation. The paper reports an 81.0% ball-shooting success rate versus 56.3% for NGGP and 38.2% for OMGP, and 85–90% success on a real grasping task versus 50% for NGGP and 55–67.5% for OMGPs.

Load-bearing premise

The claimed gains rest on the sparse OMGP base staying stable and expressive under a hand-tuned learning-rate ratio (the OMGP is updated at 10% of the Neural-ODE's rate) and on an appropriately chosen pseudo-input count, which the paper does not report; its own discussion says the OMGP is much more sensitive than the Neural-ODE, so a mis-set ratio could make the multimodal base overfit or collapse.

Editorial extensions

If this is right

  • On behavior-cloning tasks with state-varying multimodality, CGP-Flows should beat both fixed-mixture policies and single-GP-flow policies by success rate, not just by likelihood.
  • Because the ODE component is shared with NGGPs, the reduction in function evaluations transfers directly to training time: in the object-swiping experiment, training time at the strictest tolerance drops from 1230 seconds for NGGP to 355 seconds for CGP-Flows.
  • The complexity comparison $O(MNP^2 + (D+H)UE)$ versus $O(N^3 + (D+H)UE)$ implies CGP-Flows become relatively more attractive as demonstration datasets grow.
  • When the base distribution has more modes than the target, the advantage diminishes and solver cost rises, so the mixture count $M$ should stay close to the true mode count.
  • The model occupies a practical middle ground between fast-but-rigid GP mixtures and expressive-but-slow deep policies, which the paper notes can require over 12 hours to train.

Reading between the lines

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

  • Beyond the paper, the same 'start from a multimodal non-parametric base' recipe should be testable in flow-matching and diffusion policies, whose priors are typically simple; if the mechanism is general, those models would need fewer denoising or integration steps when the prior already carries the target modes.
  • Beyond the paper, the hand-tuned 10% learning-rate ratio for the OMGP parameters points to a need for two-timescale or preconditioned optimization; robustifying this step would make the method usable without its current sensitivity tuning.
  • Beyond the paper, the ball-shooting result shows that a two-mode base can still produce a three-mode target, suggesting the flow can manufacture missing modes cheaply; an adaptive mode-counting prior, which the paper floats as future work, is therefore a concrete and testable next step.
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

5 major / 7 minor

Summary. The paper proposes Composite Gaussian Processes Flows (CGP-Flows), a policy model that combines a sparse Overlapping Mixture of Gaussian Processes (OMGP) as a multimodal base distribution with a Continuous Normalizing Flow (CNF) implemented as a Neural-ODE. The authors argue that this richer base distribution reduces the complexity of the ODE transformation, improving both accuracy and computational efficiency relative to Non-Gaussian Gaussian Processes (NGGPs) and plain OMGPs. The method is evaluated by behavior cloning on a simulated ball-shooting task, a simulated object-swiping task, and a real-robot grasping task. The central empirical claim is that CGP-Flows achieve significantly higher success rates than the baselines, with chi-square tests reported for the ball-shooting comparison.

Significance. If the empirical claims are reproducible, the paper makes a useful contribution by showing that a carefully chosen non-Gaussian base distribution can reduce the computational burden of Neural-ODE-based policy models and improve accuracy on multimodal, discontinuous policies. The model itself is a coherent and natural extension of NGGPs, and the complexity analysis in Section 4.4 is a reasonable contribution. The paper does not provide code or data, and several reporting inconsistencies currently prevent verification of the main quantitative claims; the significance of the contribution is therefore conditional on those issues being resolved.

major comments (5)
  1. [§5.1.3 and §5.1.4, Table 2] The success rates in Table 2 are inconsistent with the stated evaluation protocol. Section 5.1.3 says each policy is evaluated by executing the task 200 times, yet the reported rates OMGP 38.2%, NGGP 56.3%, and CGP-Flow 81.0% imply non-integer success counts (76.4, 112.6, and 162.0). More seriously, the reported chi-square statistics do not match n=200. For equal group sizes, the comparison 81.0% vs. 56.3% with χ²=14.59 implies approximately 103 trials per condition, while the comparison 81.0% vs. 38.2% with χ²=35.02 implies approximately 92 trials per condition. The exact success counts, trial counts, and the formula used for the chi-square tests must be reported; as written, the abstract's 'significantly different' claim is unverifiable.
  2. [§4.2, Algorithm 1] Lines 5 and 6 of Algorithm 1 reference 'Eq. ??' for the updates of q({f̄^(m)}) and q(Z). The E-step updates are therefore not specified in the main text; the reader is left to infer them from Appendix B. The algorithm is incomplete as printed and must be fixed by citing the actual equations (for example, Eq. (23) and Eq. (24)).
  3. [§5.1.3, §5.2.3, §6.2.1] The number of pseudo-inputs P is never reported in any of the three experimental sections, even though Section 4.4 makes training complexity O(MNP²) and Appendix B's variational updates depend on P through the kernel matrices. Without P, the computational claims cannot be checked and the sparse OMGP base distribution cannot be reproduced. Please report P for every setting, together with the initialization and optimization of the pseudo-inputs.
  4. [§7, Discussion] The paper acknowledges that the sparse OMGP is 'much more sensitive to parameter updates' than the Neural-ODE and that the learning rate of the OMGP was set to 10% of the Neural-ODE learning rate as an ad hoc fix. Since this ratio is a central stabilization mechanism for the proposed training procedure, the exact learning-rate schedule used in each experiment should be reported, and ideally a sensitivity analysis around this ratio should be included. As written, the risk is that the reported success rates depend on a single hand-tuned value whose robustness is unknown.
  5. [§6.1 vs. §6.2.1] The real-robot state is defined in Section 6.1 as two-dimensional, s_n = [rchair_n, robs_n], but Section 6.2.1 states that 'the state input is only one-dimensional' and defines the condition extractor as h(s_n)=s_n. These statements are mutually incompatible. Please clarify the actual state and condition used in the real-robot experiment, since this determines whether the obstacle rotation enters the policy or not.
minor comments (7)
  1. [Abstract] The abbreviation 'GCP-Flow' appears in the abstract and should be 'CGP-Flow'.
  2. [§2.1] The word 'addressig' should be 'addressing'.
  3. [§5.1.3] The activation function is written as 'SiLu'; the standard spelling is 'SiLU'.
  4. [§5.2.3] The phrase 'The leand policy's performance' should read 'The learned policy's performance'.
  5. [§5.2.4] The text refers to 'Table 6' when discussing training time cost, but the table in question is labeled 'Table 3'. Please renumber the references consistently.
  6. [§6.3 and Fig. 17] The text says that 'all 20 predictive outputs correctly located the correct grasping positions' for CGP-Flow with M=2, but Fig. 17(a) reports 'Success: 16 of 20' for that model. This discrepancy should be reconciled.
  7. [Appendix C] The table headers say 'Spesification'; this should be 'Specification'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the central empirical comparison is self-contained, though the ball-shooting statistics have an internal reporting inconsistency.

full rationale

The derivation chain is not circular. The proposed CGP-Flow is a composition of two previously published components: the sparse OMGP base distribution (Eqs. 15-16, from the authors' prior work [16]) and the cCNF/Neural-ODE transformation (Eqs. 9-12, from external work [8,10]). The predictive distribution (Eq. 20) is the standard instantaneous-change-of-variables identity applied to an OMGP base; it is a construction, not a result that is later 'predicted' from fitted values. The central claims—higher success rates, reduced NFE, and lower training time—are measured empirically against external baselines (NGGP, OMGP) in simulation and on a real robot, with hyperparameters (M, tolerances, network widths) reported. The sparse-OMGP complexity O(MNP^2) is attributed to [16], a self-citation, but the same quantities are re-derived in Appendix B, so the citation is corroborated by the paper's own equations and is not load-bearing for the empirical comparison. The only anomaly is an internal reporting inconsistency: Table 2 success rates (38.2%, 56.3%, 81.0%) over 200 executions imply non-integer success counts, and the chi-square values (14.59, 35.02) imply inconsistent sample sizes. This is a statistical-reporting/validity issue, not a circular derivation; it should be corrected by the authors, but it does not make any claimed result equivalent to its inputs by construction.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The central claim rests on the sparse OMGP base distribution being a faithful multimodal density for the transformed latent space, on the correctness of the inherited variational updates from [16], and on the validity of the empirical success-rate comparisons. The unstated pseudo-input count P, the hand-tuned learning-rate ratio, and the inconsistent success-rate counts are the most fragile pieces.

free parameters (4)
  • Mixture number M for CGP-Flow base distribution = M=2 in ball-shooting and real robot; M=2..5 in object-swiping sweep
    Chosen by hand per task; the true number of modes varies with state (2 or 3 in ball-shooting; up to 5 in object-swiping). No automatic selection mechanism is provided; comparisons set M differently for CGP-Flows and baselines (e.g., M=2 vs M=3).
  • Number of pseudo-inputs P = Not reported
    P controls the sparse GP approximation quality and cost O(M N P^2). The experiments never state P, so the method cannot be reproduced and the approximation error is uncontrolled.
  • ODE solver tolerances (atol, rtol) = 1e-5 in ball-shooting and real robot; swept from 1e-6 to 1e-2 in object-swiping
    Tolerances affect NFE and accuracy; the reported NFE comparisons depend directly on this choice.
  • Learning-rate ratio between sparse OMGP and Neural-ODE parameters = OMGP learning rate set to 10% of the Neural-ODE learning rate
    Introduced ad hoc in Discussion as a fix for the sensitivity mismatch between components; the central results could depend on this hand-tuned stabilization.
assumptions (6)
  • standard math Standard Gaussian process regression and sparse GP approximations are valid for modeling the latent variable L.
    Section 3.1 and Section 4.1 use GP regression and sparse GP formulas from [21] and [16] without proof.
  • standard math The conditional continuous normalizing flow is an invertible transformation and the instantaneous change-of-variables formula (Eq. 12) holds.
    Section 3.3 and Eq. 17 rely on the Neural-ODE framework from [10], treated as a black-box.
  • domain assumption The sparse OMGP variational ELBO J_base (Eq. 19) and its analytical updates (Eqs. 23-25) are correct and inherited from [16].
    The paper cites [16] for the sparse OMGP; two of the present authors are co-authors of [16]. No independent verification is provided.
  • domain assumption Smaller divergence between base and target distributions reduces Neural-ODE NFE and training time, as argued in [11].
    This prior observation motivates the hypothesis and is used to interpret the NFE reductions in Fig. 12.
  • ad hoc to paper The hand-designed condition extractor h(.) that maps the object-swiping state to a subset index 1..31 preserves the task's learning signal.
    Section 5.2.3 defines h(s) = integer encoding of which objects remain, discarding all geometric positions. This makes the policy artificially multimodal and is a task-specific engineering choice.
  • standard math The success-rate measurements and chi-square tests are computed from independent, identically distributed trials with fixed counts.
    Chi-square tests in Section 5.1.4 require integer counts; the reported rates (56.3%, 38.2%) are inconsistent with the stated 200 trials, so this assumption is violated or the repetition count is misreported.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Composite Gaussian Processes Flows for Learning Discontinuous Multimodal Policies." pith.science (2026). https://pith.science/paper/NCRRZZ7T

@misc{pith2026250201913,
  author       = {Pith},
  title        = {Pith review of: Composite Gaussian Processes Flows for Learning Discontinuous Multimodal Policies},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NCRRZZ7T}},
  note         = {Machine review of arXiv:2502.01913}
}
read the original abstract

Learning control policies for real-world robotic tasks often involve challenges such as multimodality, local discontinuities, and the need for computational efficiency. These challenges arise from the complexity of robotic environments, where multiple solutions may coexist. To address these issues, we propose Composite Gaussian Processes Flows (CGP-Flows), a novel semi-parametric model for robotic policy. CGP-Flows integrate Overlapping Mixtures of Gaussian Processes (OMGPs) with the Continuous Normalizing Flows (CNFs), enabling them to model complex policies addressing multimodality and local discontinuities. This hybrid approach retains the computational efficiency of OMGPs while incorporating the flexibility of CNFs. Experiments conducted in both simulated and real-world robotic tasks demonstrate that CGP-flows significantly improve performance in modeling control policies. In a simulation task, we confirmed that CGP-Flows had a higher success rate compared to the baseline method, and the success rate of GCP-Flow was significantly different from the success rate of other baselines in chi-square tests.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 36 canonical work pages

  1. [1]

    IEEE Transactions on Pattern Analysis and Machine Intelligence 37(2), 408–423 (2013)

    Deisenroth, M.P., Fox, D., Rasmussen, C.E.: Gaussian processes for data-efficient learning in robotics and control. IEEE Transactions on Pattern Analysis and Machine Intelligence 37(2), 408–423 (2013)

    work page 2013

  2. [2]

    In: IEEE International Conference on Robotics and Automation (ICRA), pp

    Schreiter, J., Englert, P., Nguyen-Tuong, D., Toussaint, M.: Sparse gaussian pro- cess regression for compliant, real-time robot control. In: IEEE International Conference on Robotics and Automation (ICRA), pp. 2586–2591 (2015)

    work page 2015

  3. [3]

    In: International Conference on Robotics and Automation (ICRA), pp

    Sasaki, H., Hirabayashi, T., Kawabata, K., Matsubara, T.: Gaussian process self-triggered policy search in weakly observable environments. In: International Conference on Robotics and Automation (ICRA), pp. 5946–5952 (2022)

    work page 2022

  4. [4]

    Robotics and Autonomous Systems 112, 72–83 (2019)

    Tsurumine, Y., Cui, Y., Uchibe, E., Matsubara, T.: Deep reinforcement learning with smooth policy update: Application to robotic cloth manipulation. Robotics and Autonomous Systems 112, 72–83 (2019)

    work page 2019

  5. [5]

    In: International Conference on Machine Learning (ICML), pp

    Shen, Q., Li, Y., Jiang, H., Wang, Z., Zhao, T.: Deep reinforcement learning with robust and smooth policy. In: International Conference on Machine Learning (ICML), pp. 8707–8718 (2020)

    work page 2020

  6. [6]

    Journal of Mathematical Psychology 85, 1–16 (2018)

    Schulz, E., Speekenbrink, M., Krause, A.: A tutorial on gaussian process regres- sion: Modelling, exploring, and exploiting functions. Journal of Mathematical Psychology 85, 1–16 (2018)

    work page 2018

  7. [7]

    Pattern Recognition45(4), 1386–1395 (2012)

    L´ azaro-Gredilla, M., Van Vaerenbergh, S., Lawrence, N.D.: Overlapping mixtures of gaussian processes for the data association problem. Pattern Recognition45(4), 1386–1395 (2012)

    work page 2012

  8. [8]

    In: Neural Information Processing Systems (NeurIPS), vol

    Sendera, M., Tabor, J., Nowak, A., Bedychaj, A., Patacchiola, M., Trzcinski, T., Spurek, P., Zieba, M.: Non-gaussian gaussian processes for few-shot regression. In: Neural Information Processing Systems (NeurIPS), vol. 34, pp. 10285–10298 (2021)

    work page 2021

Show all 37 references
  1. [9]

    Journal of Machine Learning Research 22(57), 1–64 (2021)

    Papamakarios, G., Nalisnick, E., Rezende, D.J., Mohamed, S., Lakshmi- narayanan, B.: Normalizing flows for probabilistic modeling and inference. Journal of Machine Learning Research 22(57), 1–64 (2021)

    2021

  2. [10]

    In: Neural Information Processing Systems (NeurIPS), vol

    Chen, R.T., Rubanova, Y., Bettencourt, J., Duvenaud, D.K.: Neural ordinary differential equations. In: Neural Information Processing Systems (NeurIPS), vol. 31, pp. 6572–6583 (2018)

    2018

  3. [11]

    In: International Conference on Learning Representations (ICLR) (2019) 32

    Grathwohl, W., Chen, R.T., Bettencourt, J., Sutskever, I., Duvenaud, D.: FFJORD: Free-form continuous dynamics for scalable reversible generative models. In: International Conference on Learning Representations (ICLR) (2019) 32

    2019

  4. [12]

    Foundations and Trends in Robotics 7(1-2), 1–179 (2018)

    Osa, T., Pajarinen, J., Neumann, G., Bagnell, J.A., Abbeel, P., Peters, J.: An algo- rithmic perspective on imitation learning. Foundations and Trends in Robotics 7(1-2), 1–179 (2018)

    2018

  5. [13]

    In: Proceedings of the 4th Robot Learning Workshop at NeurIPS 2021 (2021)

    Chang, W., Higuera, J.C.G., Fujimoto, S., Meger, D., Dudek, G.: IL-flOw: Imita- tion Learning from Observation using Normalizing Flows. In: Proceedings of the 4th Robot Learning Workshop at NeurIPS 2021 (2021)

    2021

  6. [14]

    Conference on Robot Learning (CoRL) (2024)

    Chisari, E., Heppert, N., Argus, M., Welschehold, T., Brox, T., Valada, A.: Learning robotic manipulation policies from point clouds with conditional flow matching. Conference on Robot Learning (CoRL) (2024)

    2024

  7. [15]

    AStA Advances in Statistical Analysis 107(1), 51–72 (2023)

    Fadel, S.G., Mair, S., Silva Torres, R., Brefeld, U.: Contextual movement models based on normalizing flows. AStA Advances in Statistical Analysis 107(1), 51–72 (2023)

    2023

  8. [16]

    Neural Networks 143, 291–302 (2021)

    Sasaki, H., Matsubara, T.: Variational policy search using sparse gaussian process priors for learning multimodal optimal actions. Neural Networks 143, 291–302 (2021)

    2021

  9. [17]

    In: Neural Information Processing Systems (NeurIPS), pp

    Hausman, K., Chebotar, Y., Schaal, S., Sukhatme, G., Lim, J.J.: Multi-modal imitation learning from unstructured demonstrations using generative adversar- ial nets. In: Neural Information Processing Systems (NeurIPS), pp. 1235–1245 (2017)

    2017

  10. [18]

    IEEE Signal Processing Magazine 35(1), 53–65 (2018)

    Creswell, A., White, T., Dumoulin, V., Arulkumaran, K., Sengupta, B., Bharath, A.A.: Generative adversarial networks: An overview. IEEE Signal Processing Magazine 35(1), 53–65 (2018)

    2018

  11. [19]

    In: Conference on Robot Learning (CoRL), pp

    Florence, P., Lynch, C., Zeng, A., Ramirez, O.A., Wahid, A., Downs, L., Wong, A., Lee, J., Mordatch, I., Tompson, J.: Implicit behavioral cloning. In: Conference on Robot Learning (CoRL), pp. 158–168 (2022)

    2022

  12. [20]

    The International Journal of Robotics Research (2024)

    Chi, C., Xu, Z., Feng, S., Cousineau, E., Du, Y., Burchfiel, B., Tedrake, R., Song, S.: Diffusion policy: Visuomotor policy learning via action diffusion. The International Journal of Robotics Research (2024)

    2024

  13. [21]

    Adaptive Computation and Machine Learning

    Rasmussen, C.E., Williams, C.K.I.: Gaussian Processes for Machine Learning. Adaptive Computation and Machine Learning. MIT Press, Cambridge, MA (2006)

    2006

  14. [22]

    In: 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp

    Todorov, E., Erez, T., Tassa, Y.: Mujoco: A physics engine for model-based control. In: 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 5026–5033 (2012)

    2012

  15. [23]

    Neural Networks 107, 33 3–11 (2018)

    Elfwing, S., Uchibe, E., Doya, K.: Sigmoid-weighted linear units for neural net- work function approximation in reinforcement learning. Neural Networks 107, 33 3–11 (2018)

    2018

  16. [24]

    American journal of evaluation 33(3), 448–458 (2012)

    Franke, T.M., Ho, T., Christie, C.A.: The chi-square test: Often used and more often misinterpreted. American journal of evaluation 33(3), 448–458 (2012)

    2012

  17. [25]

    In: International Conference on Learning Representations (ICLR) (2018)

    Ramachandran, P., Zoph, B., Le, Q.V.: Searching for activation functions. In: International Conference on Learning Representations (ICLR) (2018)

    2018

  18. [26]

    In: Dasgupta, S., McAllester, D

    Ross, J., Dy, J.: Nonparametric mixture of gaussian processes with constraints. In: Dasgupta, S., McAllester, D. (eds.) International Conference on Machine Learning (ICML), vol. 28, pp. 1346–1354 (2013)

    2013

  19. [27]

    Neural Networks 158, 42–58 (2023)

    Oh, H., Sasaki, H., Michael, B., Matsubara, T.: Bayesian disturbance injection: Robust imitation learning of flexible policies for robot manipulation. Neural Networks 158, 42–58 (2023)

    2023

  20. [28]

    Journal of Machine Learning Research 12(99), 3227–3257 (2011)

    Jyl¨ anki, P., Vanhatalo, J., Vehtari, A.: Robust gaussian process regression with a student-t likelihood. Journal of Machine Learning Research 12(99), 3227–3257 (2011)

    2011

  21. [29]

    The Visual Computer 36, 39–52 (2020)

    Li, X., Huang, H., Zhao, H., Wang, Y., Hu, M.: Learning a convolutional neural network for propagation-based stereo image segmentation. The Visual Computer 36, 39–52 (2020)

    2020

  22. [30]

    IEEE Transactions on Cybernetics 52(7), 5682–5694 (2021)

    Zhang, L., Su, G., Yin, J., Li, Y., Lin, Q., Zhang, X., Shao, L.: Bioinspired scene classification by deep active learning with remote sensing applications. IEEE Transactions on Cybernetics 52(7), 5682–5694 (2021)

    2021

  23. [31]

    EURASIP Journal on Wireless Communi- cations and Networking 2019, 1–7 (2019)

    Jin, X., He, T., Lin, Y.: Multi-objective model selection algorithm for online sequential ultimate learning machine. EURASIP Journal on Wireless Communi- cations and Networking 2019, 1–7 (2019)

    2019

  24. [32]

    Information Sciences 509, 150–163 (2020)

    Chen, L., Su, W., Feng, Y., Wu, M., She, J., Hirota, K.: Two-layer fuzzy mul- tiple random forest for speech emotion recognition in human-robot interaction. Information Sciences 509, 150–163 (2020)

    2020

  25. [33]

    Journal of Computational and Nonlinear Dynamics 12(6), 061016 (2017)

    Dallas, S., Machairas, K., Papadopoulos, E.: A comparison of ordinary differential equation solvers for dynamical systems with impacts. Journal of Computational and Nonlinear Dynamics 12(6), 061016 (2017)

    2017

  26. [34]

    Wanner, G., Hairer, E.: Solving Ordinary Differential Equations II vol. 375. Springer, Heidelberg (1996)

    1996

  27. [35]

    Hu, J., Tang, H.: Numerical Methods for Differential Equations, pp. 6–9. City University, Hong Kong (2003)

    2003

  28. [36]

    Gerhard Wanner: Solving Ordinary Differential Equations I: Nonstiff Problems, pp

    Ernst Hairer, S.P.N. Gerhard Wanner: Solving Ordinary Differential Equations I: Nonstiff Problems, pp. 129–353. Springer, Heidelberg (1993) 34

    1993

  29. [37]

    In: 1st Symposium on Advances in Approximate Bayesian Inference, pp

    Grathwohl, W., Chen, R.T., Bettencourt, J., Duvenaud, D.: Scalable reversible generative models with free-form continuous dynamics. In: 1st Symposium on Advances in Approximate Bayesian Inference, pp. 1–14 (2018) 35

    2018

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.