Pith. sign in

REVIEW 4 major objections 5 minor 34 references

Transformer-Based Model Predictive Path Integral Control

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A transformer-predicted start sequence lowers MPPI cost and sample requirements.

desk verdict TransformerMPPI is a plausible warm-start trick, but the paper's headline claims about computational speed and dynamic-obstacle robustness are not backed by what is actually measured. read the letter →

arxiv 2412.17118 v1 pith:CC7UDQC4 submitted 2024-12-22 cs.RO cs.SYeess.SY

Shrenik Zinage , Vrushabh Zinage , Efstathios Bakolas This is my paper · ORCID
classification cs.ROcs.SYeess.SY
keywords transformermodelpredictivepathintegralcontrolMPPIsampleefficiencyinformedinitializationautonomousracingcollisionavoidancedynamicobstacles
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

TransformerMPPI targets a weakness of Model Predictive Path Integral (MPPI) control: each cycle, samples are drawn around a mean control sequence that is typically zero or inherited from the previous step, so a large fraction of samples can explore unproductive regions. The paper proposes training a transformer to map a short history of states plus an environment context vector to a full control horizon, and using that predicted sequence as the mean of the MPPI sampling distribution. In simulated 2D obstacle navigation and autonomous racing, the authors report that this informed initialization consistently lowers average cost, reduces the number of samples needed, and speeds up computation compared with standard MPPI, including when dynamic obstacles are added even though the transformer was trained only on static obstacle layouts. The payoff, if the claim holds, is that sampling-based controllers can run with smaller sampling budgets and still plan well, which is directly relevant to real-time robot control.

What carries the argument

The load-bearing object is the transformer-initialized mean control sequence. In standard MPPI the mean is zero-initialized or carried over from the previous iteration; in TransformerMPPI an encoder-decoder transformer with self-attention takes an encoder input of the last $k$ states concatenated with the context vector $c$ and autoregressively predicts the $H$-step mean sequence $\hat u_t, \dots, \hat u_{t+H-1}$. MPPI then samples $u^k_{t+i} = \hat u_{t+i} + \epsilon^k_{t+i}$ with $\epsilon^k_{t+i} \sim \mathcal N(0, \Sigma_u)$, so the exponential weighting by trajectory cost operates on trajectories that are already near promising behavior. The paper argues that self-attention is what lets the model exploit long-horizon patterns and adapt to time-varying environments, giving it an advantage over prior informed-sampling approaches such as conditional variational autoencoders.

What would settle it

An experiment that would settle the central claim: train the transformer only on static obstacle layouts, then evaluate it on dynamic obstacle layouts while withholding obstacle velocities from $c$; if TransformerMPPI's cost advantage over MPPI disappears, or if retraining on dynamic obstacle data produces a large additional drop in cost, then the paper's dynamic-generalization claim is falsified. The same experiment should report whether the dynamic-obstacle advantage survives when the context vector is ablated to contain no obstacle information at all.

Watch

Extended reading notes

Core claim

The central claim is that where MPPI centers its sampling distribution matters as much as how many samples are used, and a transformer trained on past optimal control data can supply a much better center than the standard choices. The paper defines a mapping $\Pi_\theta$ from the last $k$ states and a context vector $c$ to the next $H$ control inputs, trains it with teacher forcing on data generated by MPPI itself, and then runs MPPI with samples drawn as predicted mean plus Gaussian noise. The reported result is that this TransformerMPPI attains lower average cost than MPPI across every tested sample size, from 50 to 500 in the 2D navigation task and from 5000 to 10000 in autonomous racing, needs fewer steps to reach the goal in the navigation task, and keeps its cost advantage as the number of dynamic obstacles grows. The two methods converge only when the sampling budget becomes very large, which the paper presents as evidence that the benefit of the informed initialization is concentrated exactly in the low-sample regime where real-time control operates.

Load-bearing premise

The load-bearing premise is that the transformer, trained on static obstacle data, produces a mean control sequence that is also good in environments with moving obstacles, even though the paper never specifies exactly what environmental information the context vector $c$ carries and never verifies that $c$ captures the time-varying obstacle state.

Editorial extensions

If this is right

  • At low sampling budgets (50 samples in navigation), TransformerMPPI produces smoother, straighter trajectories and lower costs than MPPI, which is exactly the regime where standard MPPI struggles.
  • Across all tested sample sizes in both environments, TransformerMPPI reports lower average cost, with the gap shrinking only as the sample count approaches 10,000 in racing.
  • In the navigation task, TransformerMPPI also reduces the average number of steps to the goal at low sample counts, indicating better planned trajectories, not just lower cost.
  • The dynamic-obstacle experiments show lower average cost for TransformerMPPI as obstacle count grows, even though the transformer was not trained on dynamic obstacle configurations.
  • Fewer samples and cheap transformer forward passes make per-cycle computation faster, supporting real-time use; the initialization plugs into the standard MPPI update, so it can be composed with other MPPI variants.

Reading between the lines

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

  • An implication the authors leave implicit is that the exact contents of the context vector $c$ matter for dynamic generalization; if $c$ carries only static obstacle geometry, the reported dynamic-obstacle gains must come from the transformer exploiting smooth trajectory structure rather than from perceiving moving obstacles, and encoding obstacle velocities explicitly would be a natural testable
  • A further step the authors leave implicit is to let the transformer shape the sampling covariance as well as the mean; centering and shrinking the proposal distribution together would likely amplify the sample-efficiency gain beyond what mean-only initialization achieves.
  • Because the evaluations are entirely simulated, a hardware deployment would test whether the simulation gains survive model mismatch, actuation latency, and perception noise, which are not modeled here.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes TransformerMPPI, a variant of model predictive path integral (MPPI) control in which a transformer encoder-decoder maps a window of past states and an environmental context vector to an initial mean control sequence; MPPI then samples perturbations around this sequence rather than around a zero or previous-iteration initialization. The transformer is trained with teacher forcing on sequences generated by the baseline MPPI (Section 3.3), and the method is evaluated in two simulated tasks: 2D navigation among static and dynamic obstacles, and autonomous racing with lane information. The reported results are lower median and average costs and, in the navigation task, fewer steps across several sample sizes, along with cost comparisons as the number of dynamic obstacles is varied; the abstract additionally claims improved computational speed.

Significance. Learned initialization of sampling-based MPC is a useful research direction, and the paper offers a simple, modular integration with standard MPPI. The authors provide explicit simulation configurations, a clear algorithm listing, and a stated plan to release source code, which are strengths. However, the evidence as written does not support the full headline claim: no wall-clock time is measured despite the computational-speed claim, and the dynamic-obstacle results lack a mechanism because the transformer inputs, as specified, do not include moving-obstacle information and the transformer was not trained on dynamic obstacle configurations. If these gaps are closed, especially with runtime measurements and either a dynamic-obstacle input channel or a qualified claim, the approach would be a solid incremental contribution to informed MPPI; currently its significance is limited by the mismatch between the claims and the evidence.

major comments (4)
  1. [Abstract and Section 5 vs. Section 4] The abstract and Section 5 state that TransformerMPPI improves 'computational speed' and reduces 'computational burden,' but Sections 4.1 and 4.2 report only cost and step counts; no wall-clock time, per-iteration latency, or sample-generation throughput is reported anywhere, despite the hardware description on page 9. Since the computational-speed claim is one of the three headline outcomes, it must be supported by direct runtime measurements, such as median time per control step or end-to-end episode time at matched sample counts.
  2. [Section 4.1, Fig. 7b and Fig. 11d] The mechanism for dynamic-obstacle adaptation is not present in the method as specified. Section 3.1 defines the transformer input as xenc = [x_{t-k+1}, ..., x_t, c] with c described only as 'obstacle coordinates' or 'lane information,' and the navigation state is x_k = [x, y, theta]^T. The paper does not state that c is updated with dynamic obstacle positions or velocities at run time. Moreover, Section 4.1 explicitly says the transformer 'was not trained on dynamic obstacle configurations.' Consequently, the transformer output is independent of the time-varying obstacle configuration, so the lower costs reported for dynamic obstacles cannot be attributed to the proposed mechanism. The authors should either specify the dynamic-obstacle input channel and include such configurations in training, or remove or substantially qualify the dynamic-obstacle claim.
  3. [Sections 4.1-4.2, Figs. 6-11] All cost and step results are averages over only 10 successful episodes, with no confidence intervals, standard deviations, or significance tests, and the paper does not report success rates. With n=10, the claim of 'consistently' lower cost is weak: Fig. 6c shows TransformerMPPI better in 8 of 10 episodes, and Fig. 7a shows mixed step counts across episodes. Please report more episodes, error bars or confidence intervals, success rates, and, ideally, paired statistical comparisons.
  4. [Section 3.3 and Fig. 10b] The transformer is trained on trajectories generated by the baseline MPPI, which the text calls 'optimal.' This makes the learned initialization a fitted function of the teacher policy, so any advantage must come from warm-starting MPPI's sampling rather than from access to a better solution. The paper would be strengthened by an ablation against the standard MPPI warm-start from the previous iteration (the baseline in [31]) and by reporting whether the cost advantage persists as the sample count grows; the convergence of both methods at K=10,000 in Fig. 10b suggests such a comparison is needed to characterize the regime of benefit.
minor comments (5)
  1. [Section 3.1] The dimension of xenc is written as R^{(k+1)x(n+p)}, but concatenating k past states and one context vector should give dimension k*n+p unless the context is replicated across positions; please correct the notation.
  2. [Algorithm 1] Line 9 of Algorithm 1 reads f(xt, u^k_{t+H}) but should use u^k_{t+i} for each future step, and lines 7-13 use the inconsistent index range {0,...,H} with a horizon of H steps; the indexing should be made uniform.
  3. [Figure 1 caption] The caption contains the typo 'lower no: of samples,' which should read 'lower number of samples.'
  4. [Section 3.3 and Table 1] No separate validation or test split is described; Section 3.3 says Nenv environments are generated for training but does not state how the test environments used in Figures 5-11 are sampled. Please specify the train/test protocol.
  5. [Section 4.2] In the racing cost function R(x_k) = 2|v_k| - |d| - 5000 R*_beta - 1000000 R*_k, the sign conventions and the meaning of beta and d should be clarified, since the text does not state whether higher or lower R corresponds to better performance.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the transformer is trained on MPPI outputs, but the claimed cost/sample-efficiency improvements are empirical and not forced by construction.

full rationale

The paper's derivation chain is: MPPI defines an optimal control distribution via free energy and importance sampling; a transformer is trained on control sequences collected from MPPI; at execution the transformer output is used as the mean of the MPPI sampling distribution. The fact that the training labels come from the same MPPI algorithm is a provenance limitation, not a circular reduction: the transformer output is a function of current state and context and is not asserted to be identical to the baseline's previous-iteration mean, and the reported cost/sample reductions are empirical outcomes that could in principle fail on out-of-distribution environments. The dynamic-obstacle experiments in Section 4.1 and 4.2 are a genuine evidentiary gap -- the paper admits the transformer 'was not trained on dynamic obstacle configurations' and never specifies that context c carries time-varying obstacle information -- but this is a missing mechanism, not an equation-level equivalence or a fitted quantity renamed as a prediction. The self-citations ([3], [34]) are incidental and not load-bearing. Therefore no circular steps are present.

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

The central claim rests on fitted transformer weights and several hand-chosen MPPI hyperparameters, plus the assumption that MPPI's own outputs are good enough to supervise the transformer. No new physical entities are introduced.

free parameters (5)
  • Transformer weights theta = Trained on MPPI-generated data, not reported individually
    The central prediction is produced by these learned weights; no architecture state or checkpoint is shipped.
  • MPPI temperature lambda = Not reported
    Controls the sensitivity of importance weights in Eq. (5); not listed in Table 1.
  • Control noise covariance Sigma_u = Not reported
    Determines the sampling distribution for MPPI and is central to the algorithm.
  • Cost weights = 10000 obstacle penalty; racing weights 5000 and 1000000
    Hand-chosen reward design that shapes all reported costs.
  • Transformer hyperparameters = hidden 256, layers 3, heads 8, dropout 0.1, learning rate 5e-4
    Chosen without ablation; results may depend on these values.
assumptions (5)
  • domain assumption Known transition model f for the system dynamics
    MPPI rollouts require f, used in Eq. (1) and Algorithm 1.
  • domain assumption Context vector c contains sufficient information to predict optimal controls
    Section 3.1 defines the mapping with c but never verifies that c captures dynamic obstacle states.
  • ad hoc to paper MPPI-generated trajectories used as training targets are sufficiently close to optimal
    Section 3.3 calls the labels optimal, but they are stochastic MPPI solutions, not verified global optima.
  • domain assumption Supervised learning generalization from past states and context to future controls
    Standard machine learning assumption, not proven beyond the reported episodes.
  • standard math Standard MPPI importance-sampling derivation is valid
    Section 2.2 reproduces the known Feynman-Kac and KL-minimization derivation from Williams et al.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Transformer-Based Model Predictive Path Integral Control." pith.science (2026). https://pith.science/paper/CC7UDQC4

@misc{pith2026241217118,
  author       = {Pith},
  title        = {Pith review of: Transformer-Based Model Predictive Path Integral Control},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CC7UDQC4}},
  note         = {Machine review of arXiv:2412.17118}
}
read the original abstract

This paper presents a novel approach to improve the Model Predictive Path Integral (MPPI) control by using a transformer to initialize the mean control sequence. Traditional MPPI methods often struggle with sample efficiency and computational costs due to suboptimal initial rollouts. We propose TransformerMPPI, which uses a transformer trained on historical control data to generate informed initial mean control sequences. TransformerMPPI combines the strengths of the attention mechanism in transformers and sampling-based control, leading to improved computational performance and sample efficiency. The ability of the transformer to capture long-horizon patterns in optimal control sequences allows TransformerMPPI to start from a more informed control sequence, reducing the number of samples required, and accelerating convergence to optimal control sequence. We evaluate our method on various control tasks, including avoidance of collisions in a 2D environment and autonomous racing in the presence of static and dynamic obstacles. Numerical simulations demonstrate that TransformerMPPI consistently outperforms traditional MPPI algorithms in terms of overall average cost, sample efficiency, and computational speed in the presence of static and dynamic obstacles.

Figures

Figures reproduced from arXiv: 2412.17118 by the authors.

Figure 1
Figure 1. Instead of using the mean control sequence from the previous iteration as in the MPPI (Fig. 1a), we leverage an informed [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Schematic of TransformerMPPI: Our approach enhances the computational efficiency of the MPPI framework by lever [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Transformer architecture for mean control sequence prediction [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Transformer predictions for a random sample from test data for navigation 2D [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Trajectories generated via MPPI and TransformerMPPI for navigation 2D when the number of samples are low (i.e, 50). [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Performance comparison of TransformerMPPI versus MPPI for navigation 2D [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Performance comparison of TransformerMPPI versus MPPI for navigation 2D [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Transformer predictions for a random sample from test data for autonomous racing [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Trajectories generated via MPPI and TransformerMPPI for autonomous racing when the number of samples is 5000. [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]
Figure 10
Figure 10. Figure 10: Comparison of TransformerMPPI versus MPPI for autonomous racing [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 11
Figure 11. Figure 11: Comparison of TransformerMPPI versus MPPI for autonomous racing [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 24 canonical work pages

  1. [20]

    Informed information the- oretic model predictive control

    Raphael Kusumoto, Luigi Palmieri, Markus Spies, Akos Csiszar, and Kai O Arras. Informed information the- oretic model predictive control. In 2019 International Conference on Robotics and Automation (ICRA), pages 2047–2053. IEEE, 2019

    work page 2019

  2. [31]

    Information-theoretic model predictive control: Theory and applications to autonomous driving

    Grady Williams, Paul Drews, Brian Goldfain, James M Rehg, and Evangelos A Theodorou. Information-theoretic model predictive control: Theory and applications to autonomous driving. IEEE Transactions on Robotics, 34 (6):1603–1622, 2018

    work page 2018

  3. [1]

    Model predictive control

    Basil Kouvaritakis and Mark Cannon. Model predictive control. Switzerland: Springer International Publishing, 38:13–56, 2016

    work page 2016

  4. [2]

    Review on model predictive control: An en- gineering perspective

    Max Schwenzer, Muzaffer Ay, Thomas Bergs, and Dirk Abel. Review on model predictive control: An en- gineering perspective. The International Journal of Advanced Manufacturing Technology, 117(5):1327–1349, 2021

    work page 2021

  5. [3]

    Transformermpc: Accelerating model predictive con- trol via transformers

    Vrushabh Zinage, Ahmed Khalil, and Efstathios Bakolas. Transformermpc: Accelerating model predictive con- trol via transformers. arXiv preprint arXiv:2409.09266, 2024

    arXiv 2024

  6. [4]

    Real-time model predictive control for quadrotors

    Moses Bangura and Robert Mahony. Real-time model predictive control for quadrotors. IFAC Proceedings V olumes, 47(3):11773–11780, 2014

    work page 2014

  7. [5]

    Nonlinear model predictive control of a human- sized quadrotor

    Andrea Zanelli, Greg Horn, Gianluca Frison, and Moritz Diehl. Nonlinear model predictive control of a human- sized quadrotor. In 2018 European Control Conference (ECC), pages 1542–1547. IEEE, 2018

    work page 2018

  8. [6]

    Spacecraft attitude control using explicit model predictive control

    Øyvind Hegrenæs, Jan Tommy Gravdahl, and Petter Tøndel. Spacecraft attitude control using explicit model predictive control. Automatica, 41(12):2107–2114, 2005

    work page 2005

Show all 34 references
  1. [7]

    A comparative study of different active heave compensation ap- proaches

    Shrenik Zinage and Abhilash Somayajula. A comparative study of different active heave compensation ap- proaches. Ocean Systems Engineering, 10(4):373, 2020

    2020

  2. [8]

    COMPARATIVEANALYSISOF DIFFERENT CONTROL STRATEGIESFOR ACTIVE HEA VECOMPENSATION

    SHRENIK ZINAGE. COMPARATIVEANALYSISOF DIFFERENT CONTROL STRATEGIESFOR ACTIVE HEA VECOMPENSATION. PhD thesis, INDIAN INSTITUTE OF TECHNOLOGY MADRAS, 2021

    2021

  3. [9]

    Autonomous navigation of agvs in unknown cluttered environments: log-mppi control strategy

    Ihab S Mohamed, Kai Yin, and Lantao Liu. Autonomous navigation of agvs in unknown cluttered environments: log-mppi control strategy. IEEE Robotics and Automation Letters, 7(4):10240–10247, 2022

    2022

  4. [10]

    Towards a safe real-time motion planning framework for autonomous driving systems: An mppi approach

    Mehdi Testouri, Gamal Elghazaly, and Raphael Frank. Towards a safe real-time motion planning framework for autonomous driving systems: An mppi approach. arXiv preprint arXiv:2308.01654, 2023

    2023 arXiv

  5. [11]

    Model predictive path integral control framework for partially observable navigation: A quadrotor case study

    Ihab S Mohamed, Guillaume Allibert, and Philippe Martinet. Model predictive path integral control framework for partially observable navigation: A quadrotor case study. In 2020 16th International Conference on Control, Automation, Robotics and Vision (ICARCV), pages 196–203. I...

    2020

  6. [12]

    Model predictive path integral control for agile unmanned aerial vehicles

    Michal Minarik, Robert Penicka, V ojtech V onasek, and Martin Saska. Model predictive path integral control for agile unmanned aerial vehicles. arXiv preprint arXiv:2407.09812, 2024

    2024 arXiv

  7. [13]

    Real-time perception-limited motion plan- ning using sampling-based mpc

    Hanchen Lu, Qun Zong, Shupeng Lai, Bailing Tian, and Lihua Xie. Real-time perception-limited motion plan- ning using sampling-based mpc. IEEE Transactions on Industrial Electronics, 69(12):13182–13191, 2022

    2022

  8. [14]

    Safety barrier certificates for path integral control: Safety-critical control of quadrotors

    Tao Jin, Jian Di, Xinghu Wang, and Haibo Ji. Safety barrier certificates for path integral control: Safety-critical control of quadrotors. IEEE Robotics and Automation Letters, 2023. 14 Transformer MPPI A PREPRINT

    2023

  9. [15]

    Multi-modal mppi and active inference for reactive task and motion planning

    Yuezhe Zhang, Corrado Pezzato, Elia Trevisan, Chadi Salmi, Carlos Hernández Corbato, and Javier Alonso- Mora. Multi-modal mppi and active inference for reactive task and motion planning. IEEE Robotics and Automation Letters, 2024

    2024

  10. [16]

    Model-based generalization under parameter uncertainty using path integral control

    Ian Abraham, Ankur Handa, Nathan Ratliff, Kendall Lowrey, Todd D Murphey, and Dieter Fox. Model-based generalization under parameter uncertainty using path integral control. IEEE Robotics and Automation Letters, 5(2):2864–2871, 2020

    2020

  11. [17]

    Un- certainty averse pushing with model predictive path integral control

    Ermano Arruda, Michael J Mathew, Marek Kopicki, Michael Mistry, Morteza Azad, and Jeremy L Wyatt. Un- certainty averse pushing with model predictive path integral control. In 2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids), pages 497–502. IEEE, 2017

    2017

  12. [18]

    Self- adapting recurrent models for object pushing from learning in simulation

    Lin Cong, Michael Grner, Philipp Ruppel, Hongzhuo Liang, Norman Hendrich, and Jianwei Zhang. Self- adapting recurrent models for object pushing from learning in simulation. In 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 5304–5310. IEEE, 2020

    2020

  13. [19]

    Constrained stochastic optimal control with learned importance sampling: A path integral approach

    Jan Carius, René Ranftl, Farbod Farshidian, and Marco Hutter. Constrained stochastic optimal control with learned importance sampling: A path integral approach. The International Journal of Robotics Research, 41(2): 189–209, 2022

    2022

  14. [21]

    Auto-encoding variational bayes

    Diederik P Kingma. Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114, 2013

    2013 arXiv

  15. [22]

    Acceleration of gradient-based path integral method for efficient optimal and inverse optimal control

    Masashi Okada and Tadahiro Taniguchi. Acceleration of gradient-based path integral method for efficient optimal and inverse optimal control. In 2018 IEEE International Conference on Robotics and Automation (ICRA), pages 3013–3020. IEEE, 2018

    2018

  16. [23]

    Stein variational model pre- dictive control

    Alexander Lambert, Adam Fishman, Dieter Fox, Byron Boots, and Fabio Ramos. Stein variational model pre- dictive control. arXiv preprint arXiv:2011.07641, 2020

    2011 arXiv

  17. [24]

    Stein variational gradient descent: A general purpose bayesian inference algorithm

    Qiang Liu and Dilin Wang. Stein variational gradient descent: A general purpose bayesian inference algorithm. Advances in neural information processing systems, 29, 2016

    2016

  18. [25]

    Flying complex maneuvers with model predictive path integral control

    Jintasit Pravitra, Evangelos Theodorou, and Eric N Johnson. Flying complex maneuvers with model predictive path integral control. In AIAA Scitech 2021 Forum, page 1957, 2021

    2021

  19. [26]

    Variational inference mpc using tsallis divergence

    Ziyi Wang, Oswin So, Jason Gibson, Bogdan Vlahov, Manan S Gandhi, Guan-Horng Liu, and Evangelos A Theodorou. Variational inference mpc using tsallis divergence. arXiv preprint arXiv:2104.00241, 2021

    2021 arXiv

  20. [27]

    Robust sampling based model predictive control with sparse objective information

    Grady Williams, Brian Goldfain, Paul Drews, Kamil Saigol, James M Rehg, and Evangelos A Theodorou. Robust sampling based model predictive control with sparse objective information. In Robotics: Science and Systems, volume 14, page 2018, 2018

    2018

  21. [28]

    L1- adaptive mppi architecture for robust and agile control of multirotors

    Jintasit Pravitra, Kasey A Ackerman, Chengyu Cao, Naira Hovakimyan, and Evangelos A Theodorou. L1- adaptive mppi architecture for robust and agile control of multirotors. In 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 7661–7666. IEEE, 2020

    2020

  22. [29]

    Trajectory distribution control for model predictive path integral control using covariance steering

    Ji Yin, Zhiyuan Zhang, Evangelos Theodorou, and Panagiotis Tsiotras. Trajectory distribution control for model predictive path integral control using covariance steering. In 2022 International Conference on Robotics and Automation (ICRA), pages 1478–1484. IEEE, 2022

    2022

  23. [30]

    Attention is all you need

    A Vaswani. Attention is all you need. Advances in Neural Information Processing Systems, 2017

    2017

  24. [32]

    Aggressive driving with model predictive path integral control

    Grady Williams, Paul Drews, Brian Goldfain, James M Rehg, and Evangelos A Theodorou. Aggressive driving with model predictive path integral control. In2016 IEEE International Conference on Robotics and Automation (ICRA), pages 1433–1440. IEEE, 2016

    2016

  25. [33]

    Learning phrase representations using rnn encoder-decoder for statistical machine transla- tion

    Kyunghyun Cho, Bart Van Merriënboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. Learning phrase representations using rnn encoder-decoder for statistical machine transla- tion. arXiv preprint arXiv:1406.1078, 2014

    2014 arXiv

  26. [34]

    Leveraging gated recurrent units for iterative online precise attitude control for geodetic missions

    Vrushabh Zinage, Shrenik Zinage, Srinivas Bettadpur, and Efstathios Bakolas. Leveraging gated recurrent units for iterative online precise attitude control for geodetic missions. arXiv preprint arXiv:2405.15159, 2024. 15

    2024 arXiv

Pith tools

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