Pith. sign in

REVIEW 3 major objections 4 minor 52 references

Predictive Uncertainty for Runtime Assurance of a Real-Time Computer Vision-Based Landing System

T0 review · 3 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read A real-time runway-keypoint detector can return calibrated, sub-pixel uncertainties, and an adapted GPS-style RAIM monitor can reject faulty predictions that are inconsistent with the known runway shape.

desk verdict Solid engineering integration of Soft Argmax + NLL + RAIM for runway landing; the RAIM independence caveat is real but the authors already own it. read the letter →

arxiv 2508.09732 v1 pith:GCKYOU7M submitted 2025-08-13 cs.CV cs.RO

Romeo Valentin , Sydney M. Katz , Artur B. Carneiro , Don Walker , Mykel J. Kochenderfer This is my paper · ORCID
classification cs.CVcs.RO
keywords predictiveuncertaintyruntimeassuranceRAIMsoftargmaxkeypointregressioncalibrationposeestimationautonomouslanding
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that a vision-only landing system can do more than estimate runway corners: it can also say how sure it is about each corner, in pixel units, and use that calibrated uncertainty to reject predictions that no physically plausible aircraft pose can explain. The authors build a keypoint regressor with a spatial Soft Argmax head that reads out sub-pixel coordinates from any standard CNN backbone, train it with a negative log-likelihood loss so the predicted per-keypoint variances are faithful to actual errors, and then wrap the pose solver in an adaptation of the RAIM integrity test used in GPS navigation. On the paper's runway-image dataset, the SAM head beats fully-connected regression heads on accuracy while running at 30-60 Hz, and the RAIM-style residual statistic follows its theoretical chi-squared distribution in nominal cases while separating clearly from a simulated 184 m far-endpoint misprediction. If this holds up, a perception system could provide an integrity flag comparable to satellite navigation integrity monitoring, exactly the kind of runtime assurance that aviation certification frameworks ask for.

What carries the argument

Three pieces carry the argument. The spatial Soft Argmax operator turns each keypoint's activation heatmap into a probability distribution and returns the expected grid coordinate, enabling differentiable sub-pixel regression with almost no added parameters. The negative log-likelihood loss with a diagonal Gaussian output head produces a predicted $\sigma_x$, $\sigma_y$ per keypoint; it is a proper scoring rule, so training it encourages calibrated uncertainties. The RAIM adaptation (Algorithm 1) estimates pose by weighted least squares, reprojects the known runway corners, forms the residual $r = (I - HH^\dagger)(y_{\text{reproject}} - \mu)$, and compares the variance-normalized norm to a $

What would settle it

Generate a suite of runway images where the four corner predictions are corrupted by errors drawn from a multivariate Gaussian with increasing off-diagonal covariance, and measure the RAIM rejection rate at a fixed false-alarm threshold. Under the paper's assumptions the statistic remains chi-squared-distributed in nominal cases and the monitor flags incompatible geometry; if the nominal statistic already departs from chi-squared, or if rejection stays high under perfectly correlated errors, the independence and Gaussianity premise would be falsified.

Watch

Extended reading notes

Core claim

The central claim is that a probabilistic keypoint regressor can support runtime integrity monitoring for landing: the Soft Argmax operator extracts expected coordinates from low-resolution feature maps with sub-pixel precision, the negative log-likelihood loss with a diagonal Gaussian covariance gives uncertainty estimates whose coverage matches observed errors, and Algorithm 1's residual-based RAIM check decides ACCEPT or REJECT by comparing a variance-normalized reprojection residual to a chi-squared distribution. In the nominal validation set, the test statistic matches the theoretical chi-squared density; when the far runway threshold is mispredicted 184 m too close, the statistic separ

Load-bearing premise

The whole fault-detection scheme rests on each keypoint's pixel error being an independent, Gaussian-distributed quantity with the variance the network predicts; if real errors are correlated or non-Gaussian, the chi-squared threshold no longer means what it claims.

Editorial extensions

If this is right

  • Across ResNet18, ResNet50, and EfficientNet, the SAM head lowers median pixel error from 1.46, 2.50, and 10.59 pixels to 0.80, 0.65, and 0.50 pixels, with a minimal parameter footprint and 30-60 Hz inference.
  • Calibration curves sit close to the identity, so the Gaussian error model is adequate; inflating the predicted standard deviations by about 20% would make the model fully calibrated on the validation set.
  • The RAIM residual statistic follows the theoretical chi-squared distribution in nominal runs, which is what makes a fixed threshold meaningful rather than ad hoc.
  • A far-threshold misprediction by 184 m produces a cleanly separated residual statistic, demonstrating that geometrically incompatible outputs can be rejected at runtime without ground-truth projections.

Reading between the lines

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

  • The same geometric-compatibility test transfers to any perception task with known 3D landmarks, such as aircraft docking, satellite inspection, or warehouse fiducials, so the paper's integrity monitor is a general runtime safety layer rather than a landing-specific fix.
  • Correlated keypoint errors are the natural stress test: because the paper concedes that perfectly correlated errors are undetectable, a synthetic benchmark injecting spatially coherent shifts would quantify how much correlation the monitor tolerates before detection power collapses.
  • Because the nominal calibration is validated only in distribution, a deployment-time calibration audit would be needed to distinguish a genuine fault from a distribution shift that makes the network overconfident.
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 / 4 minor

Summary. The paper proposes a landing-scene pose-estimation pipeline for runway keypoints, combining a Soft Argmax (SAM) regression head with a heteroscedastic Gaussian negative-log-likelihood loss, and adds a RAIM-inspired geometric integrity check. The authors claim the SAM head outperforms fully-connected baselines in accuracy, the predicted uncertainties are well calibrated with sub-pixel precision, and the RAIM adaptation can detect and reject faulty keypoint predictions at runtime. Experiments use the LARD dataset with three CNN backbones and evaluate accuracy, calibration, sharpness, and nominal/off-nominal residual statistics.

Significance. If the claims hold, the paper offers a practical, real-time-compatible template for coupling per-keypoint predictive uncertainty with a geometric consistency test that does not require ground-truth annotations at runtime. The use of calibration curves and the explicit comparison to the theoretical chi-squared residual distribution are appropriate and transparent. The SAM + Gaussian-head design is simple and likely reusable across keypoint-based pose estimation tasks. However, the current strength of the claims is limited by three load-bearing gaps: the validation protocol does not use the official test split and may leak video-sequence information; the Algorithm 1 residual statistic is not the standard weighted RAIM residual and likely does not follow the claimed chi-squared distribution; and the fault-detection capability is demonstrated only on a single artificial failure mode despite the paper's own concession that correlated errors defeat the method.

major comments (3)
  1. [Section V.D] The validation protocol is a random 80/20 split of the LARD training data; the provided 1261-image validation set is not used. If LARD contains video sequences or repeated airport approaches, a random frame split can place near-duplicate frames in both training and validation, inflating all reported metrics (Table I, Fig. 4, Fig. 5). Furthermore, the 'approximately 20% standard-deviation increase' for full calibration is fit to this same split, so the claimed calibration improvement is not evaluated on an independent set. Please report results on the official validation/test split or a sequence-disjoint split, and re-derive the calibration adjustment on that split.
  2. [Algorithm 1, steps 5-6] For weighted least squares with covariance Σ and W=Σ^{-1}, the residual after pose estimation is e = y - h(θhat), and the weighted residual norm ||W^{1/2}e||² follows approximately χ²_{n-m} under the stated Gaussian model. Algorithm 1 instead computes r = (I - H H†)(yreproject - µ) with H† the unweighted pseudo-inverse of the Jacobian H. Since yreproject - µ = -e, this is -(I - H H†)e. When W is not a multiple of the identity, the projection (I - H H†) does not preserve the weighted residual distribution; the quadratic form ||L^{-1} r||² is not the standard RAIM test statistic and need not have the claimed χ²_{n-m} distribution. This undermines the threshold decision in steps 7-10. Please either define the residual as e = y - h(θhat) and use ||L^{-1}e||², or use the weighted projection P_W = H(H^TWH)^{-1}H^TW and derive the distribution of the resulting statistic explicitly.
  3. [Section VI and Fig. 5] The fault-detection claim is central ('automatically detect and flag predictions that are incompatible with the known runway shape'), but the only off-nominal experiment is a single artificial 184 m shift of the far threshold. The paper's conclusion concedes that correlated pixel errors degrade RAIM and that perfectly correlated errors make faults undetectable. A common-mode error such as a uniform translation of all four corners is geometrically absorbed by a small pose perturbation and leaves tiny reprojection residuals, so the proposed test cannot detect it. Please add synthetic fault-injection experiments with controlled correlation structures (e.g., common-mode translation, scaled runway, single-keypoint outliers) and report detection probability versus false-alarm rate, or explicitly scope the runtime-assurance claim to uncorrelated/independent keypoint errors.
minor comments (4)
  1. [Table I] The caption says 'Median Performance Results' but the columns are labeled 'NLL' and 'Mean Pixel Error.' The reported NLL values are negative, which is possible only if they are medians and the typical predicted uncertainty is well below one pixel. Please clarify which aggregate (mean/median) each column reports and include confidence intervals or per-image distributions.
  2. [Algorithm 1] The Jacobian is described as H ∈ R^{K×6}; since each projection contributes two coordinates, H should be R^{2K×6}. As written, the dimensions of H, r, and Σ do not align in steps 5-6.
  3. [Section V.C] The statement that nominal and off-nominal cases are 'clearly separated' would be stronger with quantitative separation metrics (e.g., AUC, minimum margin, or false-alarm rate at a chosen threshold).
  4. [General] There are several typographical issues: 'the simplicity of training which and model architecture', 'the validate the Gaussian error model assumption', and the 'Bogo crop' terminology. Please proofread.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's claims are supported by independent empirical validation against ground truth, and the self-citations are pointers to companion work, not load-bearing premises.

full rationale

The paper's central claims are (i) the SAM-based architecture improves keypoint accuracy, (ii) NLL-trained predictive uncertainties are well-calibrated and sharp, and (iii) a RAIM-inspired residual test can detect geometrically inconsistent predictions. None of these claims is equivalent to an input by construction. The accuracy and calibration results are evaluated against ground-truth keypoint annotations on a held-out validation set (Sec. V-A, V-B), not fitted into the architecture or loss. The RAIM statistic (Algorithm 1) is derived from the assumed Gaussian error model and the known runway geometry; its chi-squared distribution is a theoretical consequence of that assumption, and the paper independently validates it by comparing the empirical histogram of nominal statistics to the theoretical density (Fig. 5). The off-nominal detection test uses synthetic faults and shows separation, which is an empirical demonstration rather than a tautology. The only post-hoc adjustment is the explicitly optional 20% standard-deviation inflation mentioned in Sec. V-B for a safety-critical setting; it is not used to produce the main reported results and is presented as a recalibration step, not as a prediction. The self-citations to Valentin et al. [6] are references for downstream pose estimation and probabilistic pose uncertainty, not premises of the paper's own derivations. The conclusion's admission that correlated pixel errors degrade RAIM performance is a stated limitation of the Gaussian/independence assumption, not a hidden circular dependency. The derivation chain is therefore self-contained; no fitted parameter is renamed as a prediction, and no uniqueness theorem or ansatz is imported from the authors' prior work to force the chosen approach.

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

The paper introduces no new entities. The central assumptions are the Gaussian error model, the chi-squared residual distribution from RAIM theory, and the independence of keypoint errors. The only data-fitted free parameter is the 20% recalibration factor for full calibration.

free parameters (3)
  • Recalibration factor on predicted standard deviations = ~1.2 multiplier
    Section V-B: 'we may want to apply post-hoc recalibration by increasing each predicted standard deviation by approximately 20%'. This is a post-hoc fit to the validation set and affects the full-calibration claim.
  • Bogo-crop margin = Unspecified
    Section V-D: the crop must contain the entire runway 'with a specified margin'; the margin is chosen by hand and affects which pixels are trained on.
  • RAIM rejection threshold tau = Set from target false-alarm probability, value not reported
    Algorithm 1 takes a rejection threshold tau; it is a design parameter that sets the false-alarm rate, not fitted to data, but it determines the operational trade-off. Listed for completeness.
assumptions (5)
  • domain assumption Gaussian error model for keypoint prediction errors
    Section IV-B assumes the predictive distribution is Gaussian with diagonal covariance, justified by CLT and tractability. The calibration results support it on the validation set, but it is an assumption that may fail off-nominal.
  • standard math Chi-squared distribution of the corrected residual norm
    Algorithm 1 step 7 uses pdf(chi-squared_{n-m}, stat), relying on the theory from Joerger et al. [37]. This requires uncorrelated Gaussian errors and local linearity of the projection function, which is not derived in this paper.
  • domain assumption Independence of keypoint errors
    The diagonal covariance in Eq. (6) and the RAIM statistic assume independence across keypoints. The authors acknowledge in the conclusion that correlated errors degrade the method and can make faults undetectable.
  • domain assumption Known 3D runway correspondence points and camera model
    Section III-A assumes accurate known world points and camera parameters; errors in these would corrupt the residual and RAIM statistic.
  • domain assumption Local linearity of the projection function for Jacobian propagation
    Algorithm 1 step 4 linearizes the projection at the estimated pose; this is valid for small residuals and typical approach angles, but may degrade at low altitude (the paper notes in Section III-A that the inverse problem becomes sensitive at low altitudes).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Predictive Uncertainty for Runtime Assurance of a Real-Time Computer Vision-Based Landing System." pith.science (2026). https://pith.science/paper/GCKYOU7M

@misc{pith2026250809732,
  author       = {Pith},
  title        = {Pith review of: Predictive Uncertainty for Runtime Assurance of a Real-Time Computer Vision-Based Landing System},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GCKYOU7M}},
  note         = {Machine review of arXiv:2508.09732}
}
read the original abstract

Recent advances in data-driven computer vision have enabled robust autonomous navigation capabilities for civil aviation, including automated landing and runway detection. However, ensuring that these systems meet the robustness and safety requirements for aviation applications remains a major challenge. In this work, we present a practical vision-based pipeline for aircraft pose estimation from runway images that represents a step toward the ability to certify these systems for use in safety-critical aviation applications. Our approach features three key innovations: (i) an efficient, flexible neural architecture based on a spatial Soft Argmax operator for probabilistic keypoint regression, supporting diverse vision backbones with real-time inference; (ii) a principled loss function producing calibrated predictive uncertainties, which are evaluated via sharpness and calibration metrics; and (iii) an adaptation of Residual-based Receiver Autonomous Integrity Monitoring (RAIM), enabling runtime detection and rejection of faulty model outputs. We implement and evaluate our pose estimation pipeline on a dataset of runway images. We show that our model outperforms baseline architectures in terms of accuracy while also producing well-calibrated uncertainty estimates with sub-pixel precision that can be used downstream for fault detection.

Figures

Figures reproduced from arXiv: 2508.09732 by the authors.

Figure 1
Figure 1. Known world points are projected onto a camera, where they are measured as under the influence of noise. We wish to determine uncertainties about our measurements and use those to compute our pose and reject outliers. algorithm for runtime assurance. We illustrate this application in [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Prediction examples with increasing uncertainty. Predicted runway (yellow) overlaid over the true runway (white) and two standard deviations of predicted uncertainty (orange). Each element is contrasted by its opposing color for visual clarity. we represent as a standard deviation σx and σy for each coordinate. Therefore, for each keypoint, instead of outputting a point estimate, we predict a Gaussian distribution w… view at source ↗
Figure 3
Figure 3. Rejection example of a common failure mode: true runway corners (white), corner predictions (yellow), repro￾jected corners (green) and predicted uncertainty (orange, two standard deviations). C. Runtime Integrity Monitoring through RAIM Adaptation To ensure the operational reliability and safety of our pipeline, particularly in safety-critical landing scenarios, it is insufficient to merely produce pose estimates; w… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Probabilistic metrics for ResNet18 + SAM. A. Predictive Accuracy To evaluate the performance of our architecture, we con￾struct models with three different standard CNN backbones: ResNet18, ResNet50 [51] and EfficientNet [52]. For each backbone, we train a baseline mod…
Figure 5
Figure 5. Figure 5: Runtime separation of nominal cases and off-nominal cases with mispredicted far threshold. reliably detect when a prediction occurs that is incompatible with the geometric constraints of the known correspondence points: no pose can be found such that the reprojections …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 46 canonical work pages

  1. [1]

    A Golden Decade of Deep Learning: Computing Systems & Applications,

    J. Dean, “A Golden Decade of Deep Learning: Computing Systems & Applications,” Daedalus, vol. 151, no. 2, pp. 58–74, 2022

    work page 2022

  2. [2]

    of Automotive Engineers Aerospace, Guidelines for Development of Civil Aircraft and Systems: SAE ARP 4754 Rev

    S. of Automotive Engineers Aerospace, Guidelines for Development of Civil Aircraft and Systems: SAE ARP 4754 Rev. B. SAE, 2023

    work page 2023

  3. [3]

    Framework for Certification of AI-Based Systems,

    M. Gariel, B. Shimanuki, R. Timpe, and E. Wilson, “Framework for Certification of AI-Based Systems,” arXiv, arXiv: 2302 . 11049 (cs), 2023

    work page 2023

  4. [4]

    Concepts of Design Assurance for Neural Networks (CoDANN),

    EASA and Daedalean AG, “Concepts of Design Assurance for Neural Networks (CoDANN),” Technical report. 2020

    work page 2020

  5. [5]

    Artificial intelligence – uncertainty quantification in machine learning,

    DIN, “Artificial intelligence – uncertainty quantification in machine learning,” Deutsches Institut für Normung, Berlin, Germany, DIN SPEC 92005, 2024, Text in English, p. 46

    work page 2024

  6. [6]

    Probabilistic Parameter Estimators and Calibration Metrics for Pose Estimation from Image Features,

    R. Valentin, S. M. Katz, J. Lee, D. Walker, M. Sorgenfrei, and M. J. Kochenderfer, “Probabilistic Parameter Estimators and Calibration Metrics for Pose Estimation from Image Features,” in 2024 AIAA DATC/IEEE 43rd Digital Avionics Systems Conference (DASC) , 2024, pp. 01–09

    work page 2024

  7. [7]

    LARD - landing approach runway detection - dataset for vision based landing,

    M. Ducoffe, M. Carrere, L. Féliers, A. Gauffriau, V . Mussot, C. Pagetti, and T. Sammour, “LARD - landing approach runway detection - dataset for vision based landing,” 2023

    work page 2023

  8. [8]

    End-to-End Race Driving with Deep Reinforcement Learning,

    M. Jaritz, R. de Charette, M. Toromanoff, E. Perot, and F. Nashashibi, “End-to-End Race Driving with Deep Reinforcement Learning,” in 2018 IEEE International Conference on Robotics and Automation (ICRA), 2018, pp. 2070–2075

    work page 2018

Show all 52 references
  1. [9]

    Driving Autonomy with Event-Based Cameras: Algorithm and Hardware Perspectives,

    N. M. Rahman, U. Kamal, M. Nagaraj, S. Roy, and S. Mukhopad- hyay, “Driving Autonomy with Event-Based Cameras: Algorithm and Hardware Perspectives,” in 2024 Design, Automation & Test in Europe Conference & Exhibition (DATE) , 2024, pp. 1–6

    2024

  2. [10]

    Autonomous Vehicles on the Edge: A Survey on Autonomous Vehicle Racing,

    J. Betz, H. Zheng, A. Liniger, U. Rosolia, P. Karle, M. Behl, V . Krovi, and R. Mangharam, “Autonomous Vehicles on the Edge: A Survey on Autonomous Vehicle Racing,” IEEE Open Journal of Intelligent Transportation Systems, vol. 3, pp. 458–488, 2022

    2022

  3. [11]

    Autonomous Drone Racing: A Survey,

    D. Hanover, A. Loquercio, L. Bauersfeld, A. Romero, R. Penicka, Y . Song, G. Cioffi, E. Kaufmann, and D. Scaramuzza, “Autonomous Drone Racing: A Survey,” IEEE Transactions on Robotics , vol. 40, pp. 3044–3067, 2024

    2024

  4. [12]

    Champion-level drone racing using deep reinforce- ment learning,

    E. Kaufmann, L. Bauersfeld, A. Loquercio, M. Müller, V . Koltun, and D. Scaramuzza, “Champion-level drone racing using deep reinforce- ment learning,” Nature, vol. 620, no. 7976, pp. 982–987, 2023

    2023

  5. [13]

    A vision- based approach for unmanned aerial vehicle landing,

    C. Patruno, M. Nitti, A. Petitti, E. Stella, and T. D’Orazio, “A vision- based approach for unmanned aerial vehicle landing,” Journal of Intelligent & Robotic Systems , vol. 95, no. 2, pp. 645–664, 2019

    2019

  6. [14]

    Detecting flying objects using a single moving camera,

    A. Rozantsev, V . Lepetit, and P. Fua, “Detecting flying objects using a single moving camera,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 39, no. 5, pp. 879–892, 2016

    2016

  7. [15]

    Learning to detect aircraft for long-range vision-based sense-and-avoid systems,

    J. James, J. J. Ford, and T. L. Molloy, “Learning to detect aircraft for long-range vision-based sense-and-avoid systems,” IEEE Robotics and Automation Letters, vol. 3, no. 4, pp. 4383–4390, 2018

    2018

  8. [16]

    Visual-based obstacle detection and tracking, and conflict detection for small UAS sense and avoid,

    R. Opromolla and G. Fasano, “Visual-based obstacle detection and tracking, and conflict detection for small UAS sense and avoid,” Aerospace Science and Technology , vol. 119, p. 107 167, 2021

    2021

  9. [17]

    Small aircraft detection based on feature enhancement and context information,

    J. Ying, H. Li, H. Yang, and Y . Jiang, “Small aircraft detection based on feature enhancement and context information,” Journal of Aerospace Information Systems, pp. 1–12, 2022

    2022

  10. [18]

    Domain Generalization: A Survey,

    K. Zhou, Z. Liu, Y . Qiao, T. Xiang, and C. C. Loy, “Domain Generalization: A Survey,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 4, pp. 4396–4415, 2023

    2023

  11. [19]

    A Holistic Assessment of the Reliability of Machine Learning Systems,

    A. Corso, D. Karamadian, R. Valentin, M. Cooper, and M. J. Kochen- derfer, “A Holistic Assessment of the Reliability of Machine Learning Systems,” arXiv, arXiv: 2307.10586 (cs), 2023

    2023 arXiv

  12. [20]

    Does progress on ImageNet transfer to real-world datasets?

    A. Fang, S. Kornblith, and L. Schmidt, “Does progress on ImageNet transfer to real-world datasets?” arXiv, arXiv: 2301.04644 (cs), 2023

    2023 arXiv

  13. [21]

    Reachability analysis of neural feedback loops,

    M. Everett, G. Habibi, C. Sun, and J. P. How, “Reachability analysis of neural feedback loops,” IEEE Access, vol. 9, pp. 163 938–163 953, 2021

    2021

  14. [22]

    OVERT: An algorithm for safety verification of neural network control policies for nonlinear systems,

    C. Sidrane, A. Maleki, A. Irfan, and M. J. Kochenderfer, “OVERT: An algorithm for safety verification of neural network control policies for nonlinear systems,” Journal of Machine Learning Research , vol. 23, no. 117, pp. 1–45, 2022

    2022

  15. [23]

    Reachable polyhedral marching (RPM): A safety verification algorithm for robotic systems with deep neural network components,

    J. A. Vincent and M. Schwager, “Reachable polyhedral marching (RPM): A safety verification algorithm for robotic systems with deep neural network components,” in IEEE International Conference on Robotics and Automation (ICRA) , IEEE, 2021, pp. 9029–9035

    2021

  16. [24]

    Reachability analysis for neural network aircraft collision avoidance systems,

    K. D. Julian and M. J. Kochenderfer, “Reachability analysis for neural network aircraft collision avoidance systems,” AIAA Journal of Guidance, Control, and Dynamics , vol. 44, no. 6, pp. 1132–1142, 2021

    2021

  17. [25]

    Verification of neural network compression of ACAS Xu lookup tables with star set reachability,

    D. M. Lopez, T. Johnson, H.-D. Tran, S. Bak, X. Chen, and K. L. Hobbs, “Verification of neural network compression of ACAS Xu lookup tables with star set reachability,” in AIAA Scitech Forum, 2021, p. 0995

    2021

  18. [26]

    Ver- ification of image-based neural network controllers using generative models,

    S. M. Katz, A. L. Corso, C. A. Strong, and M. J. Kochenderfer, “Ver- ification of image-based neural network controllers using generative models,” Journal of Aerospace Information Systems , vol. 19, no. 9, pp. 574–584, 2022

    2022

  19. [27]

    Neural network compression of ACAS Xu early prototype is unsafe: Closed-loop verification through quantized state backreachability,

    S. Bak and H.-D. Tran, “Neural network compression of ACAS Xu early prototype is unsafe: Closed-loop verification through quantized state backreachability,” in NASA Formal Methods Symposium (NFM) , Springer, 2022, pp. 280–298

    2022

  20. [28]

    A Survey of Algorithms for Black-Box Safety Validation of Cyber- Physical Systems,

    A. Corso, R. J. Moss, M. Koren, R. Lee, and M. J. Kochenderfer, “A Survey of Algorithms for Black-Box Safety Validation of Cyber- Physical Systems,” Journal of Artificial Intelligence Research , vol. 72,

  21. [29]

    Correlated encounter model for cooperative aircraft in the national airspace system,

    M. J. Kochenderfer, L. P. Espindle, J. K. Kuchar, and J. D. Griffith, “Correlated encounter model for cooperative aircraft in the national airspace system,” Massachusetts Institute of Technology, Lincoln Lab- oratory, Project Report ATC-344, 2008

    2008

  22. [30]

    Bayesian safety validation for failure probability estimation of black-box sys- tems,

    R. J. Moss, M. J. Kochenderfer, M. Gariel, and A. Dubois, “Bayesian safety validation for failure probability estimation of black-box sys- tems,” Journal of Aerospace Information Systems , vol. 21, no. 7, pp. 533–546, 2024

    2024

  23. [31]

    Diffusion-based failure sampling for cyber-physical systems,

    H. Delecki, M. R. Schlichting, M. Arief, A. Corso, M. Vazquez- Chanlatte, and M. J. Kochenderfer, “Diffusion-based failure sampling for cyber-physical systems,” in IEEE International Conference on Engineering Reliable Autonomous Systems (ERAS) , 2025

    2025

  24. [32]

    Failure probability estimation for black-box autonomous systems using state-dependent importance sampling proposals,

    H. Delecki, S. Katz, and M. J. Kochenderfer, “Failure probability estimation for black-box autonomous systems using state-dependent importance sampling proposals,” in International Conference on Con- trol, Decision and Information Technologies (CoDIT) , 2025

    2025

  25. [33]

    Guidance for level 1 & 2 machine learning applications,

    European Union Aviation Safety Agency, “Guidance for level 1 & 2 machine learning applications,” EASA, Concept Paper, 2024

    2024

  26. [34]

    A review of uncertainty quantification in deep learning: Techniques, applications and challenges,

    M. Abdar, F. Pourpanah, S. Hussain, D. Rezazadegan, L. Liu, M. Ghavamzadeh, P. Fieguth, X. Cao, A. Khosravi, U. R. Acharya, et al., “A review of uncertainty quantification in deep learning: Techniques, applications and challenges,” Information Fusion, vol. 76, pp. 243–297, 2021

    2021

  27. [35]

    Conformal prediction: A gentle introduction,

    A. N. Angelopoulos, S. Bates, et al., “Conformal prediction: A gentle introduction,” Foundations and Trends in Machine Learning , vol. 16, no. 4, pp. 494–591, 2023

    2023

  28. [36]

    Individual calibration with randomized forecasting,

    S. Zhao, T. Ma, and S. Ermon, “Individual calibration with randomized forecasting,” in Proceedings of the 37th International Conference on Machine Learning, vol. 119, JMLR.org, 2020, pp. 11 387–11 397

    2020

  29. [37]

    Solution Separation Versus Residual-Based RAIM,

    M. Joerger, F.-C. Chan, and B. Pervan, “Solution Separation Versus Residual-Based RAIM,” Navigation, vol. 61, no. 4, pp. 273–291, 2014

    2014

  30. [38]

    GNSS receiver autonomous integrity monitoring (RAIM) performance analysis,

    S. Hewitson and J. Wang, “GNSS receiver autonomous integrity monitoring (RAIM) performance analysis,” GPS Solutions , vol. 10, pp. 155–170, 2006

    2006

  31. [39]

    A baseline GPS RAIM scheme and a note on the equivalence of three RAIM methods,

    R. G. Brown, “A baseline GPS RAIM scheme and a note on the equivalence of three RAIM methods,” Navigation, vol. 39, no. 3, pp. 301–316, 1992

    1992

  32. [40]

    Navigation system integrity monitoring using redundant measurements,

    M. A. Sturza, “Navigation system integrity monitoring using redundant measurements,” Navigation, vol. 35, no. 4, pp. 483–501, 1988

    1988

  33. [41]

    Strutz, Data Fitting and Uncertainty

    T. Strutz, Data Fitting and Uncertainty . Springer, 2011

    2011

  34. [42]

    M. J. Kochenderfer and T. A. Wheeler, Algorithms for Optimization . MIT Press, 2019

    2019

  35. [43]

    Probabilistic forecasts, calibration and sharpness,

    T. Gneiting, F. Balabdaoui, and A. E. Raftery, “Probabilistic forecasts, calibration and sharpness,” Journal of the Royal Statistical Society: Series B (Statistical Methodology) , vol. 69, no. 2, pp. 243–268, 2007

    2007

  36. [44]

    Probabilistic Forecasting,

    T. Gneiting and M. Katzfuss, “Probabilistic Forecasting,” Annual Review of Statistics and Its Application , vol. 1, no. 1, pp. 125–151, 2014

    2014

  37. [45]

    DeepPose: Human Pose Estimation via Deep Neural Networks,

    A. Toshev and C. Szegedy, “DeepPose: Human Pose Estimation via Deep Neural Networks,” in 2014 IEEE Conference on Computer Vision and Pattern Recognition , 2014, pp. 1653–1660. arXiv: 1312 . 4659 [cs]

    2014

  38. [46]

    U-Net: Convolutional Net- works for Biomedical Image Segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-Net: Convolutional Net- works for Biomedical Image Segmentation,” in Medical Image Com- puting and Computer-Assisted Intervention – MICCAI 2015 , Springer International Publishing, 2015, pp. 234–241

    2015

  39. [47]

    Stacked Hourglass Networks for Human Pose Estimation,

    A. Newell, K. Yang, and J. Deng, “Stacked Hourglass Networks for Human Pose Estimation,” arXiv, arXiv: 1603.06937 (cs), 2016

    2016 arXiv

  40. [48]

    Deep High-Resolution Representation Learning for Human Pose Estimation,

    K. Sun, B. Xiao, D. Liu, and J. Wang, “Deep High-Resolution Representation Learning for Human Pose Estimation,” arXiv, arXiv: 1902.09212 (cs), 2019

    1902 arXiv

  41. [49]

    End-to-End Training of Deep Visuomotor Policies,

    S. Levine, C. Finn, T. Darrell, and P. Abbeel, “End-to-End Training of Deep Visuomotor Policies,” arXiv, arXiv: 1504.00702 (cs), 2016

    2016 arXiv

  42. [50]

    Estimating the mean and variance of the target probability distribution,

    D. A. Nix and A. S. Weigend, “Estimating the mean and variance of the target probability distribution,” in Proceedings of 1994 ieee international conference on neural networks (ICNN’94) , IEEE, vol. 1, 1994, pp. 55–60

    1994

  43. [51]

    Deep Residual Learning for Image Recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep Residual Learning for Image Recognition,” arXiv, arXiv: 1512.03385 (cs), 2015

    2015 arXiv

  44. [52]

    EfficientNet: Rethinking Model Scaling for Con- volutional Neural Networks,

    M. Tan and Q. Le, “EfficientNet: Rethinking Model Scaling for Con- volutional Neural Networks,” in Proceedings of the 36th International Conference on Machine Learning , PMLR, 2019, pp. 6105–6114

    2019

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.