Pith. sign in

REVIEW 4 major objections 5 minor 43 references

Deliberate Practice: Learning Robot Skills under a Budget

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read This paper proves that a robot can spend a fixed practice budget across its skills in a globally optimal way, by solving one bilinear program that jointly picks the budget allocation and the task plan.

desk verdict A genuinely new exact reformulation of budgeted skill practice, with a sound derivation and a caveat that 'budget-optimal' is only as good as the fitted competence model. read the letter →

arxiv 2608.13415 v1 pith:P7I4OLT4 submitted 2026-08-13 cs.RO cs.AI

Shivam Vats , Sudarshan Harithas , Mete Tuluhan Akbulut , Arvind Raghunathan , George Konidaris This is my paper · ORCID
classification cs.ROcs.AI
keywords budgetedskilllearningactivetaskandmotionplanningbileveloptimizationbilinearprogramrobotskillspracticebudgetlong-horizonmanipulation
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

Robots refining skills during deployment often have only a limited number of practice episodes. This paper introduces Deliberate Practice (DP), an algorithm that decides in advance how many episodes to spend on each skill in a library so that the expected reward from the best reachable task plan is maximized. The central result is Theorem 1: the bilevel problem of choosing a budget allocation while solving the resulting task MDP is exactly equivalent to one flat bilinear program, which off-the-shelf global solvers can optimize. If the competence-improvement model is right, the resulting practice plan is globally optimal for the budget. Experiments in simulated and real long-horizon manipulation show DP switching from an easy low-reward plan under a small budget to a harder high-reward plan under a larger one, whereas greedy baselines stay stuck.

What carries the argument

The load-bearing object is the bilinear program of Eq. (3), which jointly optimizes the practice allocation $b_u$ and dual state-action occupancy variables $\mu_s^a$, with skill transition probabilities $\bar P^a_{s's}(b) = f_{\text{improv}}(P^a_{s's}, b)$ entering bilinearly into the occupancy constraints. The competence prediction model $f_{\text{improv}}(u,b) = \min(1, p_u + \Delta_u b)$ is the mechanism that maps practice episodes into improved transitions, with $\Delta_u$ estimated online from observed improvement. The bilinear terms are relaxed by piecewise McCormick envelopes, and the resulting nonconvex program is solved to global optimality by spatial branch-and-bound, using finite bounds on $\mu$ and the transition probabilities to certify the result.

What would settle it

Run DP on a small MDP such as the five-state example in Section 4.1, enumerate every feasible allocation of a 20-episode budget by brute force, and solve each resulting MDP exactly; if any enumerated allocation achieves higher expected reward than DP's chosen allocation, Theorem 1 is false. On a real robot, give two robots the same budget but deliberately different competence priors: the one with optimistic priors should, per the paper's own limitation, sometimes pick a plan it cannot finish within the budget, showing that budget-optimality holds for predicted rather than realized improvement.

Watch

Extended reading notes

Core claim

The paper's central claim is that budgeted skill learning—maximize expected task value over a library of skills with a total practice budget—can be solved to global optimality by a single bilinear program (Eq. 3), rather than approximated by greedy one-step improvement. The move is to replace the inner MDP solve with its dual linear program, then fold the budget constraint into the dual feasibility equations, because the leader and follower optimize the same objective. Strong duality turns the bilevel max-min problem into a joint maximization over the allocation $b$ and the state-action occupancy variables $\mu$. The paper proves the reformulation is exact, gives finite bounds on the occupancy and transition variables so that piecewise McCormick envelopes produce valid convex relaxations, and invokes spatial branch-and-bound for global certification. The corollary, stated as Theorem 1, is that the returned allocation is budget-optimal: it maximizes expected cumulative reward among all allocations feasible within the budget, for the predicted competence dynamics.

Load-bearing premise

The whole plan is only as good as the robot's guess of how much each skill will improve per practice episode; if that guess is wrong, the 'optimal' allocation is optimal only inside the model.

Editorial extensions

If this is right

  • A robot with a known downtime window can compute, before practicing, exactly which skills to train and for how long, instead of adapting greedily after each episode.
  • Under small budgets the optimal behavior is conservative—practice only the easy, low-reward plan—while under larger budgets the same procedure shifts to hard high-reward plans that require multiple skills, as demonstrated in simulation and on a real robot.
  • Greedy active-learning baselines plateau at local optima because one-step task improvement is zero for skills that only pay off when practiced together; DP's joint lookahead avoids that failure.
  • The bilinear program scales to task MDPs with about 5,000 abstract states and 22 skills, solving within minutes; if interrupted before proving optimality, the solver still returns an allocation with an optimality-gap certificate.

Reading between the lines

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

  • The duality reformulation is not tied to the particular linear competence model: any competence predictor that enters the transition probabilities as a function of $b$ and yields finite bounds could be placed in the same program, so the machinery may extend to saturating or learned improvement curves with the same global-optimality framework.
  • Because Theorem 1 holds for the predicted competence model, a natural stress test is to compare DP's allocation under optimistically biased priors against a robust or uncertainty-aware allocation; the paper itself flags optimistic priors as the key failure mode.
  • The same 'replace the follower with its dual' move applies to any bilevel resource-allocation problem whose inner problem is a linear program with an aligned objective, so budget-optimal skill learning may transfer to settings such as allocating teaching time among sub-skills or tuning a fleet of policies.
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

4 major / 5 minor

Summary. The paper addresses the problem of allocating a limited practice budget across a library of robot skills in a sequential task-and-motion-planning setting. It formulates budgeted skill learning as a bilevel optimization problem (Eq. 1) whose inner level solves the task MDP induced by the budget-dependent competence model, and whose outer level chooses a budget allocation. The main contribution is Deliberate Practice (DP), which models competence improvement with a piecewise-linear function (Eq. 2), derives an exact single-level bilinear reformulation of the bilevel problem via LP duality (Eq. 3 and Appendix A.2), and solves it with off-the-shelf global solvers. The authors claim that DP is provably budget-optimal (Theorem 1) and validate it in two simulated manipulation domains and one real-robot breakfast-making task.

Significance. The idea of coupling a practice budget to long-horizon task planning is timely and practically motivated, and the LP-duality reformulation is an interesting technical contribution that goes beyond greedy active-learning baselines. If the reformulation and its optimality proof are correct, the paper offers a principled way to make budget-aware decisions in robot skill learning, with experiments spanning nontrivial state spaces (up to 5000 abstract states and 22 skills) and a real robot. The paper also gives a clear, reproducible problem statement and an explicit comparison to myopic baselines. However, the optimality claim is conditional on the competence model and the proof contains fixable but nontrivial gaps; the empirical evidence does not yet establish that the computed allocations are optimal or even feasible for the real robot.

major comments (4)
  1. [Appendix A.2, Eq. (7) and Section 5.2, Eq. (3)] The displayed bilinear program omits the nonnegativity constraint on the dual variables mu_s^a. In the standard dual LP of an MDP these variables are state-action occupancies and must satisfy mu_s^a >= 0. Without this constraint, the feasible set allows negative values for individual actions as long as the sum over actions satisfies the flow equality. Because the objective is linear in mu, a solver could shift occupancy from a low-reward action to a higher-reward action with identical transition probabilities, improving the objective while violating the intended MDP interpretation. The claim that Eq. (3) is an exact reformulation of Eq. (1) therefore is not supported as written. The fix is straightforward -- add mu_s^a >= 0 -- but the theorem and proof must be updated to state and use this constraint.
  2. [Appendix A.2, Lemma 1] The proof of Slater's condition is incorrect: the proposed candidate hat v_s = alpha/gamma^2 does not strictly satisfy the constraints for discount factors with gamma > 0.618. For example, with gamma = 0.9 the left-hand side alpha/0.81 is smaller than the maximum right-hand side alpha + alpha/0.9 = alpha(1 + 1/gamma). A correct choice would be hat v_s = C with C > alpha/(1-gamma), which yields strict satisfaction. Since the LP is feasible and bounded for discounted MDPs, strong duality can be established by standard LP arguments, but the lemma as stated is mathematically wrong and must be corrected.
  3. [Section 4 and Section 5.1, Eq. (2)] The budget B is defined as a number of trials or episodes, but the optimization in Eq. (3) treats the allocation b as a continuous variable with no integrality constraint and no rounding analysis. The competence model f_improv(u,b) is evaluated at real-valued b, while practice episodes are indivisible. The proof of global optimality in Theorem 1 is therefore over the continuous relaxation, not over integer allocations. The paper does not show that the continuous optimum can be rounded to an integer allocation without losing optimality or feasibility. This is a load-bearing gap between the mathematical claim and the actual robot-setting where budgets are integer episode counts.
  4. [Section 8 and Section 6.1 (Fig. 6)] Theorem 1 certifies optimality only with respect to the competence prediction model of Eq. (2), not with respect to the robot's realized competence. The paper's own Section 8 states that optimistic priors can make the chosen plan infeasible within the budget, which means the 'provably budget-optimal' claim is conditional on an unvalidated model. The real-robot results in Fig. 6 are reported without error bars, without repeated trials, and without any calibration check comparing predicted competence f_improv to the competence actually achieved after practice. The abstract's wording that DP 'maximizes expected cumulative reward while being learnable within the budget' overstates what is established. The claims should be qualified to optimality under the assumed competence model, and the empirical section should include calibration evidence or a sensitivity analysis over priors.
minor comments (5)
  1. [Appendix A.5] The text uses 'Auroco Markers' and 'Auroco Detection'; the correct spelling is ArUco.
  2. [Section 5.1] The online estimation of Delta_u is not fully specified: the update uses p_t values, but it is not stated how p_t is estimated from the outcome of a single practice episode or how many episodes are used. A binary success/failure per episode makes a running average of differences very noisy; please specify the estimation window and any smoothing used in the experiments.
  3. [Section 6.1, Fig. 6] The real-robot figure lacks information about the number of independent runs and does not show variability. Reporting mean plus standard deviation or individual trial outcomes would strengthen the claim that DP reliably selects the correct plan under each budget.
  4. [Appendix A.2, Eq. (6) and Eq. (7)] The transition-probability indexing is inconsistent between the general notation P^a_{s's} in Eq. (7) and the earlier use of \bar P^a_{s's}(b) in Eq. (3); the reader must infer which index is the source state. Please unify the notation.
  5. [Section 5.2] The sentence claiming 'to our knowledge, this is the first exact formulation of robot learning on a budget as an optimization problem' is an overclaim and difficult to verify; tempering it to 'to our knowledge, the first exact single-level formulation of this budgeted skill-learning problem' would be more defensible.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: Theorem 1 is a model-conditional optimality proof via standard LP duality; the competence model is an explicit assumption, not a fitted prediction renamed as a result.

full rationale

Theorem 1 (Section 5.2) claims that Deliberate Practice computes a globally optimal allocation of the practice budget. The proof is a self-contained mathematical derivation: the bilevel problem in Eq. (1) is instantiated as the LP bilevel program in Eq. (4), the inner LP is replaced by its dual using strong duality (Slater's condition verified in Lemma 1, Appendix A.2), and the resulting max-max problem is merged into the single-level bilinear program Eq. (7)/Eq. (3). This is an exact reformulation of the stated optimization problem, not an identity that assumes the conclusion. The optimal allocation is not set equal to a fitted constant by construction, and the global-optimality argument rests on standard MDP LP duality (Puterman [30]) and branch-and-bound global optimization [32], both independent of the authors. The competence prediction model f_improv in Eq. (2) is an explicit modeling assumption, with a domain-specific prior and an online estimate of Delta_u; the theorem's budget-optimality is conditional on that model. That is a model-validity limitation, explicitly acknowledged in Section 8 (optimistic priors may make the chosen plan infeasible), but it is not circularity: a theorem can be correct about a model while the model is imperfect. The empirical validation tests the practical usefulness of the model, but the derivation chain itself does not reduce to the fitted competence rates. No load-bearing self-citation or imported uniqueness theorem appears; the paper's own LP-duality derivation is shown in the appendix. Therefore no significant circularity is present.

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

The central claim is budget-optimality of the computed allocation. This rests on the competence model (Eq 2), the LP duality reformulation, and the continuous treatment of the budget. No new physical entities are introduced; the algorithm is the contribution.

free parameters (3)
  • competence improvement rate Delta_u = estimated online via running average (Eq 2)
    Each skill's competence improvement per practice episode is a parameter of the prediction model; these rates determine the transition function P(b) in Eq (3) and hence the optimal allocation.
  • smoothing factor epsilon = chosen in [0,1], value not specified in the paper
    Used in the running average Delta^t_u = epsilon Delta^{t-1}_u + (1-epsilon)(p^t_u - p^{t-1}_u); a hyperparameter that affects the rate estimate.
  • initial competence priors p_u = domain-specific prior, values not given
    The model is initialized with a prior over each skill's competence (Section 5.1); these priors are inputs to the optimization and influence which plans appear learnable.
assumptions (5)
  • standard math Standard LP duality for MDPs
    The reformulation in Appendix A.2 relies on the primal and dual LPs for the MDP value function having equal optimal values (Puterman [30]).
  • standard math Slater's condition holds for the inner LP
    Lemma 1 proves the existence of a strictly feasible v, enabling strong duality.
  • domain assumption Skill competence evolves as min(1, p + Delta b)
    Competence prediction model in Eq (2); assumes linear improvement with a cap, which is a modeling choice.
  • domain assumption The abstract MDP and task planner are correct
    The robot is given high-level skill specifications and a planner that solves the abstract MDP optimally (Section 3); the optimization evaluates value functions of this MDP.
  • domain assumption Budget b is treated as continuous
    The optimization (Eq 3) has no integrality constraints, while B is the number of episodes in Eq (1); no rounding scheme is given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Deliberate Practice: Learning Robot Skills under a Budget." pith.science (2026). https://pith.science/paper/P7I4OLT4

@misc{pith2026260813415,
  author       = {Pith},
  title        = {Pith review of: Deliberate Practice: Learning Robot Skills under a Budget},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P7I4OLT4}},
  note         = {Machine review of arXiv:2608.13415}
}
read the original abstract

We consider the problem of autonomously learning robot skills under a limited practice budget for sequential tasks. We propose an active skill learning algorithm, \emph{Deliberate Practice (DP)}, that computes a provably \emph{budget-optimal} allocation---practicing skills that maximize expected cumulative reward while being learnable within the budget. DP estimates both the time needed to master skills and the cumulative reward of the task plans that the skills unlock. Computing a budget-optimal allocation is challenging as it requires reasoning about combinatorially many skill plans over a large practice budget. Our key contribution is a bilinear program that can compute this exactly using off-the-shelf solvers. Through simulated and real-world experiments on long-horizon manipulation tasks, we show that our approach allows robots to optimally use limited practice time to acquire useful policies and improve long-horizon planning.

Figures

Figures reproduced from arXiv: 2608.13415 by the authors.

Figure 1
Figure 1. Deliberate Practice allows robots to intelligently use a limited practice budget to actively learn skills to maximize long-horizon task reward. (top) Under a small budget, the robot practices an easy task plan that toasts bread and provides reward 1. (bottom) Under a larger budget, it practices a more difficult but higher-reward task plan that microwaves oatmeal and provides reward 2. or use additional practice time… view at source ↗
Figure 2
Figure 2. An example illustrat￾ing how greedy active learning can make suboptimal decisions. Prior active skill learning methods [27, 28] for sequential tasks greedily practice the skill that will most improve the expected task performance Jtask after one round of training. These meth￾ods can be viewed as performing one-step reasoning to solve equation 1. However, this is inherently myopic: it can over￾look higher-performing … view at source ↗
Figure 3
Figure 3. Overview: Our approach predicts skill competence as a function of practice budget, computes a budget allocation for practice, and practices skills by interacting with the environment. model provides a good balance between performance and efficiency. fimprov(u, b) = min(1, pu + ∆ub), (2) where ∆u is the rate of improvement of a skill u estimated from practice. ∆u is estimated online as a running average of competence… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: (a) Breakfast task requires learning forceful manipulation skills to interact with novel articulated objects. (b) Cleanup task requires the robot to place objects in the top, middle, or bottom drawer with rewards 1, 2 and 4 respectively. 3. Breakfast (real-robot). A Fr…
Figure 5
Figure 5. Figure 5: We evaluate all methods on Cleanup with budgets (B) of 100 (low), 150 (medium), and 250 (high) episodes. While they perform similarly under a low budget, DP is significantly better under medium and high budgets. We report the mean and standard deviation averaged over 5…
Figure 6
Figure 6. Figure 6: Real-world Breakfast task. Real-world Validation. We conduct experiments on a real Franka Panda robot to validate our design choices. To complete the Breakfast task, the robot must learn either to toast bread or microwave oatmeal. As shown in [PITH_FULL_IMAGE:figures/…
Figure 7
Figure 7. Figure 7: (a) Breakfast domain. The robot can either toast bread to achieve reward 1 or microwave oatmeal to achieve reward 2. The former requires practicing StartToaster while the latter requires practicing OpenMicrowave and CloseMicrowave. (b) Cleanup-multi domain. The achieve…
Figure 8
Figure 8. Figure 8: Breakfast domain. After pressing the microwave button, the robot estimates the position of the microwave door using a RealSense wrist camera. This information is used to construct a collision model of the scene for motion planning. to the object properties and task con…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 30 canonical work pages

  1. [1]

    Pastor, M

    P. Pastor, M. Kalakrishnan, S. Chitta, E. Theodorou, and S. Schaal. Skill learning and task outcome prediction for manipulation. In2011 IEEE international conference on robotics and automation, pages 3828–3834. IEEE, 2011

    work page 2011

  2. [2]

    Y . Zhu, Z. Jiang, P. Stone, and Y . Zhu. Learning generalizable manipulation policies with object-centric 3d representations.arXiv preprint arXiv:2310.14386, 2023

    arXiv 2023

  3. [3]

    C. Chi, Z. Xu, S. Feng, E. Cousineau, Y . Du, B. Burchfiel, R. Tedrake, and S. Song. Diffusion policy: Visuomotor policy learning via action diffusion.The International Journal of Robotics Research, 44(10-11):1684–1704, 2025

    2025

  4. [4]

    Cambon, R

    S. Cambon, R. Alami, and F. Gravot. A hybrid approach to intricate motion, manipulation and task planning.The International Journal of Robotics Research, 28(1):104–126, 2009

    work page 2009

  5. [5]

    C. R. Garrett, R. Chitnis, R. Holladay, B. Kim, T. Silver, L. P. Kaelbling, and T. Lozano-P´erez. Integrated task and motion planning.Annual review of control, robotics, and autonomous systems, 4(1):265–293, 2021

    2021

  6. [6]

    Task and Skill Planning: Hierarchical Robot Planning with Black-Box Skills

    B. Hedegaard, Z. Yang, Y . Wei, A. Jaafar, S. Tellex, G. Konidaris, and N. Shah. Beyond task and motion planning: Hierarchical robot planning with general-purpose policies.arXiv preprint arXiv:2504.17901, 2025

    work page Pith review arXiv 2025

  7. [7]

    Liang, M

    J. Liang, M. Sharma, A. LaGrassa, S. Vats, S. Saxena, and O. Kroemer. Search-based task planning with learned skill effect models for lifelong robotic manipulation. In2022 Interna- tional Conference on Robotics and Automation (ICRA), pages 6351–6357. IEEE, 2022

    work page 2022

  8. [8]

    Mishani, Y

    I. Mishani, Y . Shaoul, and M. Likhachev. Mosaic: A skill-centric algorithmic framework for long-horizon manipulation planning.arXiv preprint arXiv:2504.16738, 2025

    arXiv 2025

Show all 43 references
  1. [9]

    Kober, J

    J. Kober, J. A. Bagnell, and J. Peters. Reinforcement learning in robotics: A survey.The International Journal of Robotics Research, 32(11):1238–1274, 2013

    2013

  2. [10]

    M. P. Deisenroth, G. Neumann, and J. Peters. A survey on policy search for robotics.Founda- tions and Trends® in Robotics, 2(1-2):1–142, 2013

    2013

  3. [11]

    S. Gu, E. Holly, T. Lillicrap, and S. Levine. Deep reinforcement learning for robotic manipula- tion with asynchronous off-policy updates. In2017 IEEE international conference on robotics and automation (ICRA), pages 3389–3396. IEEE, 2017

    2017

  4. [12]

    V . Mnih, K. Kavukcuoglu, D. Silver, A. A. Rusu, J. Veness, M. G. Bellemare, A. Graves, M. Riedmiller, A. K. Fidjeland, G. Ostrovski, et al. Human-level control through deep rein- forcement learning.nature, 518(7540):529–533, 2015

    2015

  5. [13]

    Kalashnikov, A

    D. Kalashnikov, A. Irpan, P. Pastor, J. Ibarz, A. Herzog, E. Jang, D. Quillen, E. Holly, M. Kalakrishnan, V . Vanhoucke, et al. Scalable deep reinforcement learning for vision-based robotic manipulation. InConference on robot learning, pages 651–673. PMLR, 2018

    2018

  6. [14]

    Anders Ericsson

    K. Anders Ericsson. Deliberate practice and acquisition of expert performance: a general overview.Academic emergency medicine, 15(11):988–994, 2008

    2008

  7. [15]

    Da Silva, G

    B. Da Silva, G. Konidaris, and A. Barto. Active learning of parameterized skills. InInterna- tional Conference on Machine Learning, pages 1737–1745. PMLR, 2014

    2014

  8. [16]

    Chernova and M

    S. Chernova and M. Veloso. Interactive policy learning through confidence-based autonomy. Journal of Artificial Intelligence Research, 34:1–25, 2009

    2009

  9. [17]

    Fabisch and J

    A. Fabisch and J. H. Metzen. Active contextual policy search.The Journal of Machine Learn- ing Research, 15(1):3371–3399, 2014. 10

    2014

  10. [18]

    S. Vats, O. Kroemer, and M. Likhachev. Synergistic scheduling of learning and allocation of tasks in human-robot teams. In2022 International Conference on Robotics and Automation (ICRA), pages 2789–2795. IEEE, 2022

    2022

  11. [19]

    S. Vats, M. Zhao, P. Callaghan, M. Jia, M. Likhachev, O. Kroemer, and G. Konidaris. Optimal interactive learning on the job via facility location planning.Robotics: Science and Systems (RSS), 2025

    2025

  12. [20]

    Srivastava, E

    S. Srivastava, E. Fang, L. Riano, R. Chitnis, S. Russell, and P. Abbeel. Combined task and motion planning through an extensible planner-independent interface layer. In2014 IEEE international conference on robotics and automation (ICRA), pages 639–646. IEEE, 2014

    2014

  13. [21]

    Fox and D

    M. Fox and D. Long. Pddl2. 1: An extension to pddl for expressing temporal planning domains. Journal of artificial intelligence research, 20:61–124, 2003

    2003

  14. [22]

    K. Hauser. Task planning with continuous actions and nondeterministic motion planning queries. InProc. of AAAI Workshop on Bridging the Gap between Task and Motion Plan- ning, 2010

    2010

  15. [23]

    C. R. Garrett, T. Lozano-P´erez, and L. P. Kaelbling. Pddlstream: Integrating symbolic planners and blackbox samplers via optimistic adaptive planning. InProceedings of the international conference on automated planning and scheduling, volume 30, pages 440–448, 2020

    2020

  16. [24]

    Chitnis, D

    R. Chitnis, D. Hadfield-Menell, A. Gupta, S. Srivastava, E. Groshev, C. Lin, and P. Abbeel. Guided search for task and motion plans using learned heuristics. In2016 IEEE International Conference on Robotics and Automation (ICRA), pages 447–454. IEEE, 2016

    2016

  17. [25]

    Silver, R

    T. Silver, R. Chitnis, J. Tenenbaum, L. P. Kaelbling, and T. Lozano-P ´erez. Learning sym- bolic operators for task and motion planning. In2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 3182–3189. IEEE, 2021

    2021

  18. [26]

    Shah and S

    N. Shah and S. Srivastava. Hierarchical planning and learning for robots in stochastic set- tings using zero-shot option invention. InProceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 10358–10367, 2024

    2024

  19. [27]

    S. Vats, M. Likhachev, and O. Kroemer. Efficient recovery learning using model predictive meta-reasoning. In2023 IEEE International Conference on Robotics and Automation (ICRA), pages 7258–7264. IEEE, 2023

    2023

  20. [28]

    Kumar, T

    N. Kumar, T. Silver, W. McClinton, L. Zhao, S. Proulx, T. Lozano-P ´erez, L. P. Kaelbling, and J. Barry. Practice makes perfect: Planning to learn skill parameter policies. InRobotics: Science and Systems (RSS), 2024

    2024

  21. [29]

    Abbatematteo, E

    B. Abbatematteo, E. Rosen, S. Thompson, T. Akbulut, S. Rammohan, and G. Konidaris. Com- posable interaction primitives: A structured policy class for efficiently learning sustained- contact manipulation skills. In2024 IEEE International Conference on Robotics and Automa- tion ...

    2024

  22. [30]

    M. L. Puterman.Markov decision processes: discrete stochastic dynamic programming. John Wiley & Sons, 2014

    2014

  23. [31]

    Gurobi Optimizer Reference Manual, 2024

    Gurobi Optimization, LLC. Gurobi Optimizer Reference Manual, 2024. URLhttps:// www.gurobi.com

    2024

  24. [32]

    Tawarmalani and N

    M. Tawarmalani and N. V . Sahinidis.Convexification and global optimization in continuous and mixed-integer nonlinear programming: theory, algorithms, software, and applications, volume 65. Springer Science & Business Media, 2013. 11

    2013

  25. [33]

    Todorov, T

    E. Todorov, T. Erez, and Y . Tassa. Mujoco: A physics engine for model-based control. In2012 IEEE/RSJ international conference on intelligent robots and systems, pages 5026–5033. IEEE, 2012

    2012

  26. [34]

    B. Liu, Y . Zhu, C. Gao, Y . Feng, Q. Liu, Y . Zhu, and P. Stone. Libero: Benchmarking knowl- edge transfer for lifelong robot learning.arXiv preprint arXiv:2306.03310, 2023

    2023 arXiv

  27. [35]

    Stout and A

    A. Stout and A. G. Barto. Competence progress intrinsic motivation. In2010 IEEE 9th inter- national conference on development and learning, pages 257–262. IEEE, 2010

    2010

  28. [36]

    Colas, P

    C. Colas, P. Fournier, M. Chetouani, O. Sigaud, and P.-Y . Oudeyer. Curious: intrinsically motivated modular multi-goal reinforcement learning. InInternational conference on machine learning, pages 1331–1340. PMLR, 2019

    2019

  29. [37]

    N. Hogan. Impedance control: An approach to manipulation: Part i—theory.Journal of Dynamic Systems, Measurement, and Control, 107(1):1–7, 1985

    1985

  30. [38]

    Hansen and A

    N. Hansen and A. Ostermeier. Completely derandomized self-adaptation in evolution strate- gies.Evolutionary computation, 9(2):159–195, 2001

    2001

  31. [39]

    G. P. McCormick. Computability of global solutions to factorable nonconvex programs: Part i—convex underestimating problems.Mathematical programming, 10(1):147–175, 1976

    1976

  32. [40]

    Garrido-Jurado, R

    S. Garrido-Jurado, R. Mu ˜noz-Salinas, F. J. Madrid-Cuevas, and M. J. Mar ´ın-Jim´enez. Auto- matic generation and detection of highly reliable fiducial markers under occlusion.Pattern Recognition, 47(6):2280–2292, 2014

    2014

  33. [41]

    G. Bradski. The opencv library.Dr. Dobb’s Journal: Software Tools for the Professional Programmer, 25(11):120–123, 2000

    2000

  34. [42]

    I. A. Sucan, M. Moll, and L. E. Kavraki. The open motion planning library.IEEE Robotics & Automation Magazine, 19(4):72–82, 2012

    2012

  35. [43]

    Chitta, I

    S. Chitta, I. Sucan, and S. Cousins. Moveit![ros topics].IEEE robotics & automation maga- zine, 19(1):18–19, 2012. 12 A Appendix A.1 Linear Programming Formulation of MDPs We leverage the linear programming (LP) formulation of MDPs to instantiateSolveMDPin Equa- tion 1. Using ...

    2012

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.