Hierarchical Task Decomposition for Execution Monitoring and Error Recovery: Understanding the Rationale Behind Task Demonstrations

Christoph Willibald; Dongheui Lee

REVIEW 3 major objections 6 minor 12 references

Hierarchical Task Decomposition for Execution Monitoring and Error Recovery: Understanding the Rationale Behind Task Demonstrations

T0 review · 3 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read A robot can learn a multi-step contact task from three demonstrations and recover from never-seen failures.

desk verdict A promising LfD framework with real-robot results, but the segmentation algorithm's Q-value computation is never defined, which makes the paper's core contribution non-reproducible as written. read the letter →

arxiv 2505.04565 v1 pith:KV4JU34Q submitted 2025-05-07 cs.RO

Christoph Willibald , Dongheui Lee This is my paper · ORCID

classification cs.RO MSC 68T4068T05

keywords learningfromdemonstrationunsupervisedtasksegmentationinversereinforcementanomalydetectionincrementalcontact-richmanipulationgraphGaussianmixtureregression

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper claims that a robot can learn a hierarchical model of a multi-step contact-rich manipulation task from as few as three human demonstrations on the real system, without simulation, without skill labels, and without a list of possible failures. The learned model has two levels: each skill is a low-level motion and force model with an expected feature region and a subgoal, and the skills are organized in a task graph that monitors progress and selects recovery behaviors. The authors argue that segmenting the demonstrations with a combination of inverse reinforcement learning and probabilistic feature clustering, rather than clustering trajectories or forces alone, is what makes the skill boundaries meaningful. They further claim that conditioning anomaly detection on the measured end-effector pose rather than on time lets the robot tell apart unfamiliar situations from normal variability, and react accordingly. If the claims hold, contact tasks that are expensive to simulate can be taught quickly, monitored online, and extended with recovery behaviors only when genuinely new failures occur.

What carries the argument

The load-bearing object is the optimality score $\varepsilon_d(s_i,a_i,g_k^d)=Q^{T_d}(s_i,a_i,g_k^d)/V^*(s_i,g_k^d)$, the ratio between the action value of the demonstrated policy and the optimal value for reaching a candidate subgoal; it lies in $[0,1]$, is fed through a softmax to give the probability that the observed action is directed to a given subgoal, and is multiplied by a multivariate Gaussian feature-constraint likelihood in the observation model (4). This combined likelihood is the basis for a Bayesian nonparametric Gaussian mixture with a Chinese restaurant process prior and Normal-Inverse-Wishart priors, inferred by collapsed Gibbs sampling (Algorithm 1). On the execution side, the machinery is Gaussian mixture regression conditioned on the measured end-effector pose: the predicted mean velocity and force and the conditional covariance define a Mahalanobis distance (15) whose threshold $D_{M,\max}$ is the largest distance over training data, while the mixture density (14) controls whether the prediction is confident. Together these pieces let the framework segment demonstrations unsupervised, monitor execution online, and decide when to learn a new recovery behavior.

What would settle it

Recompute the box-pushing segmentation from the same demonstrations using a different deterministic transition model, such as a different velocity discretization or a small perturbation, and check whether the inferred skill boundaries and subgoals stay the same; if they change, the claimed intention recognition is sensitive to the unstated dynamics rather than being determined by the demonstrations.

Watch

Extended reading notes

Core claim

The central claim is that a task performed in contact with the environment can be decomposed into skills by jointly scoring how each demonstrated action is directed toward a reachable subgoal and how the accompanying multimodal features cluster in a shared probabilistic model. The resulting algorithm, Bayesian Nonparametric Gaussian Inverse Reinforcement Learning (BNG-IRL), models each skill as a mixture component with a subgoal region in end-effector state space and a constraint region in feature space, and infers the number of skills, the skill assignments, and the subgoal of every demonstration by collapsed Gibbs sampling. On the execution side, the framework predicts the expected end-effector velocity and force from the measured pose via Gaussian mixture regression, flags a deviation when the Mahalanobis distance from the prediction exceeds the largest distance seen in training, and only trusts the flag when the measured pose lies in a region of low epistemic uncertainty. If a novel anomaly is confirmed, the user demonstrates a recovery behavior that is segmented and appended to the task graph at the skill where the anomaly occurred. The paper reports that this pipeline learns the box grasping and locking task from three demonstrations, detects four different anomaly types on a real robot without false positives during successful runs, and outperforms the compared segmentation and anomaly-detection baselines.

Load-bearing premise

The segmentation relies on an assumed but never specified model of how robot states change under actions and how the optimal value function is computed; if that transition model and action discretization are wrong, the optimality scores that drive skill inference lose their meaning.

Editorial extensions

If this is right

Contact-rich tasks that are hard to simulate can be taught on the real robot from three kinesthetic demonstrations, without an accurate physics model.
Skill boundaries found by combining subgoal intentions with feature constraints are more reliable than clustering on trajectories or forces alone, as measured against four segmentation baselines.
A detector that conditions on the measured end-effector pose can flag force deviations before they become visually observable, and it will not raise alarms in regions of the state space it does not know yet.
When a new anomaly is confirmed, the demonstrated recovery is added to the task graph for that skill, so the same recovery works regardless of when during the skill the anomaly appears.
Distinguishing aleatoric from epistemic uncertainty makes the anomaly detector more sensitive where training variance is low and switches to user supervision where the model has no data.

Reading between the lines

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

A direct stress test the paper leaves implicit is to alter the hidden transition model while keeping the demonstrations fixed; stable skill boundaries would indicate the segmentation is driven by the demonstrations' intentions, not by the model's assumptions.
The known weak case, the visually obvious but force-neutral 'box missed' anomaly, suggests a parallel fusion of this pose-conditioned force detector with a vision-based skill-state monitor, with the task graph arbitrating between their reports.
Because each skill's feature region is a multivariate Gaussian, the method can only capture linear correlations between features; a kernelized or mixture extension would be a natural test of whether nonlinear force-position trade-offs can still be learned from the same few demonstrations.
The paper considers reusing recovery skills across different skills; if skill preconditions were learned instead of hard-coded by graph position, one demonstrated recovery could in principle serve several skills, which would cut the number of required recovery demonstrations.

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.

Desk Editor's Note

read the letter

Worth a look if you work on learning from demonstration for contact-rich manipulation: this paper has a genuinely useful idea and real-robot experiments, but the central algorithm has a load-bearing gap that needs to be fixed before the results can be trusted. The new combination is BNG-IRL: subgoal-driven inverse RL fused with probabilistic feature clustering in a Bayesian nonparametric mixture, plus an extended-horizon optimality score that looks at the whole trajectory toward a subgoal rather than just the immediate action. That is a real step beyond prior BN-IRL and CBN-IRL. The two-step anomaly detection—epistemic confidence gate followed by aleatoric Mahalanobis distance—is a sensible, practical contribution, and the comparison against supervised vision-language baselines on force-based anomalies is a useful data point. Credit where due: they show a contact-rich box-locking task learned from three demonstrations, and their detector catches all four anomalies in their test set with no false positives on the successful runs. The soft spots are the problem. The stress-test note is right: Eqs. (1)–(2) require a transition model P(s'|s,a) and an action set A(s), and neither is ever specified. The state space is "all reachable EEF poses" and actions are "EEF velocities," but the Bellman equations sum over s' and maximize over a without any dynamics or discretization. In continuous 6D pose space, that max is ill-posed. This is not a minor implementation detail—Algorithm 1 precomputes Q-values and the segmentation likelihood in Eq. (4) depends directly on them, so the numerical segmentation results in Tables 1–2 cannot be independently reproduced or even checked. The anomaly thresholds are also fitted to the training data envelope (max Mahalanobis distance, min density), which is okay as a heuristic but should be described as such; the claim about "confidently detecting all anomalies" is really "detecting deviations beyond the tightest training variation," and the "box missed" case shows the method's limit. No code, no data, no error bars; several claims about reduced compute and data efficiency are not backed by numbers. In proportion: the experimental story is plausible, and the framework ideas are sound enough to deserve a proper referee. But as submitted, the segmentation core is not computable as written. This is a serious revise-and-resubmit. If they clarify the dynamics (even a simple kinematic model with explicit discretization), release code, and re-frame the anomaly-threshold fitting honestly, it could be a solid contribution. I'd send it to review, not desk-reject, but I would insist on those fixes.

Referee Report

3 major / 6 minor

Summary. The paper proposes an incremental learning-from-demonstration framework for contact-rich manipulation. User demonstrations are segmented by a new unsupervised algorithm, BNG-IRL, which combines feature-space Gaussian clustering with subgoal-based inverse reinforcement learning; each inferred skill receives a subgoal region, a feature constraint region, and an assigned set of demonstrations. The low-level execution model uses GMM/GMR to generate pose-conditioned velocity and force commands, and the anomaly detector computes Mahalanobis distances against the trained skill model, with a two-step confidence scheme intended to separate epistemic and aleatoric uncertainty. Skills are organized in a task graph, and recovery behaviors are incrementally appended when novel anomalies are detected. The experiments evaluate segmentation on a simulated box-pushing task and on two real robots (DLR LWR IV and DLR SARA), including a box grasping and locking task with four injected anomaly types, and compare anomaly detection against vision-language and supervised video baselines.

Significance. If the central segmentation derivation can be made concrete, this would be a valuable applied contribution: the framework addresses a real gap by learning hierarchical, force-sensitive manipulation skills from very few demonstrations on physical robots, and it explicitly tackles incremental recovery. The real-robot experiments, the comparisons against several baselines, and the honest limitations section in Sec. 7.4 are strengths. The segmentation idea of combining subgoal intentions with feature constraints is sensible, and the paper is right that the segmentation itself is not circular because BNG-IRL infers skills from the data rather than assuming them. However, the key algorithmic step, the computation of Q-values and the optimal value function in Eqs. (1) and (2), is never specified, which currently prevents independent reproduction and leaves the numerical segmentation results in Tables 1 and 2 dependent on an undisclosed transition model and action discretization.

major comments (3)

[§5.1.1, Eqs. (1)–(4), Algorithm 1 line 4] The optimality score ε_d(s_i,a_i,g) that drives the segmentation likelihood (4) is not computable as stated. Equations (1) and (2) define Q_Td and V* through Bellman recurrences over a transition model P(s′|s,a) and an action set A(s), but neither is specified anywhere. Section 4.1 says only that S is 'all reachable robot EEF poses' and actions are 'EEF velocities'; for the contact tasks S is 6D with quaternion orientation (Sec. 7.3.2). In this continuous setting the max over A(s) in Eq. (2) is undefined without a discretization or a dynamical model, and in contact phases such as pushing or locking a simple kinematic s′ = s + a·dt model is inconsistent because the next EEF pose depends on the object and the contact state. If a learned or empirical transition model was used, it must be described; if a kinematic model was used, its validity in the presence of contact must be justified. Algorithm 1 line 4 ('Precompute Q(·,·,R_g)') does not specify how Q and V* are obtained. This is load-bearing because the numerical segmentation results in Tables 1 and 2 depend on the hidden choice, and the same issue affects the BN-IRL baseline reimplementation. Please provide the exact transition model, action discretization, discount factor, and reward used in every experiment.
[§6.1, Eqs. (15)–(17)] The anomaly detection thresholds are defined as training-data extrema: D_M,max = max over the training Mahalanobis distances and P_min(s) = min over the training set of P(s). This makes the detector a comparator against the most extreme training observation rather than a calibrated predictor, and the 'confidence' threshold P_min(s) is not a principled uncertainty level. Consequently, the paper's claim that the method leverages epistemic and aleatoric uncertainty to make confident predictions (Secs. 6.1 and 8) overstates what these quantities can certify. The empirical results may still be valid, but the authors should either provide a calibration or ablative analysis showing how the thresholds behave with varying amounts of training data, or revise the uncertainty/confidence language so that it matches the operational definition used in Eqs. (16) and (17).
[§7.3.6, Tables 4 and 5] The conclusion that the method is 'the only one to confidently detect all anomalies, while not triggering any false positive detection during successful executions' is only partially supported by the reported tables. For anomaly IV ('box missed'), Table 4 shows GMR(our) with recall 34.6 and a detection delay of 4.3 s, and the text states that ConditionNET performs better in this case. Table 5 reports 100% task-level accuracy for the same case, so the two tables are not necessarily contradictory, but the phrase 'confidently detect all anomalies' obscures the fact that frame-wise detection of case IV is weak and late. Please state explicitly that detection for case IV usually occurs at the execution level and after a long delay, and report the successful-execution false-positive result alongside the per-case detection accuracy so that the comparison is not read as frame-wise superiority on every anomaly.

minor comments (6)

[§5.1.2 (heading)] The heading contains a typo: 'Probabilistc Feature Clustering' should be 'Probabilistic Feature Clustering'.
[Algorithm 1, lines 3–4] The pseudocode is ambiguous about the loop over subgoal candidates: line 3 sets 'g←s_i' inside a loop over i, but it is not clear whether Q is precomputed once per candidate state or once per demonstration, and how the precomputed Q-values are indexed during the Gibbs updates. Please rewrite the pseudocode to make the subgoal candidate indexing and the Q-value precomputation explicit.
[Eq. (1), §5.1.1] After the deterministic-transition simplification, the notation R_g(s,a) appears in the last line of Eq. (1), while the reward was defined as R_g(s,a,s′) = 1(s′=g). Please define the reduced reward R_g(s,a) or write the expectation over s′ explicitly to avoid ambiguity.
[Table 4] For the GMRwo rows in the 'slide locked' and 'box missed' columns, precision, recall, and F1 are all 0 but a numeric detection delay is still reported (8.3 and 6.6). A delay for a detector that never fires is not meaningful; mark these entries as 'no detection' or define the delay for missed detections.
[§7.3.6] The anomaly trigger condition is described as 'more than 300 ms' in the text (Secs. 7.2.2 and 7.3.6), but the parameter ε in Eq. (16) and Algorithm 2 is not given in cycles or seconds. Please specify the control rate and the value of ε used in the experiments.
[General] The paper would benefit from a supplementary release of the code or a detailed appendix containing the discretization and dynamics used for the Q-value computation; without this, even after adding a textual description, the segmentation results in Tables 1 and 2 cannot be independently checked.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation; BNG-IRL and the GMR anomaly detector are self-contained inference procedures, with the main weakness being an underspecified transition model (a correctness issue, not circularity).

full rationale

The paper's central segmentation contribution is a Bayesian nonparametric mixture model whose likelihood (Eq. 4) combines an IRL-based intention term with a feature-clustering term; both terms are defined from the demonstrations and priors, and the latent skills/subgoals are inferred by collapsed Gibbs sampling rather than taken from the output. The anomaly detector (Eqs. 15-17) uses thresholds equal to training-data extrema (DM,max and Pmin(s)); this is a standard, transparent novelty-detection rule and does not rename a fit as a prediction, since the decision boundary is a statistic of the training distribution and is applied to new measurements. Self-citations (Willibald et al. 2020; Eiband et al. 2023b; Willibald and Lee 2022) are contextual and do not carry the derivation. The most serious weakness is that Eqs. (1)-(2) require a transition model P(s'|s,a) and an action space A(s) that are never specified, making the optimality score epsilon_d and hence the segmentation likelihood not fully computable as written; however, that is an under-specification/reproducibility concern, not a circular reduction of the claimed result to its inputs.

Assumptions & free parameters 8 free parameters · 6 assumptions · 0 invented entities

The central learning pipeline rests on several unstated modeling choices. The most important is the existence of a deterministic transition model for Q-value computation, which is never specified; the model also assumes each skill pursues a single subgoal, that skill features are Gaussian, and that pose-conditioning is sufficient for prediction. Hyperparameters for the Bayesian priors and softmax temperature are not reported. The paper introduces no new physical entities.

free parameters (8)

alpha (demonstration confidence)
Softmax temperature in Eq (3) controls reliance on the optimality score; no value is reported in the experiments.
eta (CRP concentration)
Chinese Restaurant Process prior over the number of skills in Eq (6); no value is reported.
NIW hyperparameters beta_G, beta_C
Normal-Inverse-Wishart priors for subgoal and constraint regions (Eq 7); no values are reported.
Gibbs sampling steps T = 1000
Number of sampling iterations used in all experiments (Sec 7.1.4); convergence is not assessed.
Number of GMM components per skill E = 2
Hand-chosen in Sec 7.3.5 for motion generation and anomaly detection models.
Anomaly trigger window epsilon = 300 ms
Number of consecutive anomalous cycles required before declaring an anomaly (Sec 7.2.2, 7.3.6).
Anomaly threshold D_M,max and confidence threshold P_min(s) = max/min over training data Ok
Defined in Eqs (16)-(17) as the maximum Mahalanobis distance and minimum likelihood over the training set; the detector is calibrated to the training envelope.
SVM sliding window length = 100
Used for the anomaly-type classifier in Sec 7.3.7; no sensitivity analysis is provided.

assumptions (6)

domain assumption A deterministic transition model P(s'|s,a) exists and is known for computing Q-values and V*.
Used in Eqs (1)-(2) and Algorithm 1; the actual dynamics are never specified, making V* computation undefined in continuous state space.
domain assumption Each skill is directed at a single subgoal state, encoded by sparse reward R_g(s,a,s') = 1(s'=g).
Central to intention recognition in Sec 5.1.1; assumes every action within a skill targets one terminal subgoal.
domain assumption Skill features follow a multivariate Gaussian distribution.
Constraint regions are modeled as Gaussians in Sec 5.1.2; the authors acknowledge in Sec 7.4 that nonlinear feature relations are not captured.
domain assumption Skill behavior depends only on the current end-effector pose, not on time.
Pose-conditioned GMR in Sec 6.1 assumes force/velocity is a function of pose, ignoring dynamic or history-dependent effects.
domain assumption User demonstrations are approximately optimal with respect to the skill subgoal.
The optimality score in Eq (3) and the softmax parameter alpha rely on this near-optimality assumption.
standard math Chinese Restaurant Process and Normal-Inverse-Wishart priors are appropriate for skill discovery.
Standard Bayesian nonparametric machinery used in Eqs (5)-(8); no custom derivation is provided.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Hierarchical Task Decomposition for Execution Monitoring and Error Recovery: Understanding the Rationale Behind Task Demonstrations." pith.science (2026). https://pith.science/paper/KV4JU34Q

@misc{pith2026250504565,
  author       = {Pith},
  title        = {Pith review of: Hierarchical Task Decomposition for Execution Monitoring and Error Recovery: Understanding the Rationale Behind Task Demonstrations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KV4JU34Q}},
  note         = {Machine review of arXiv:2505.04565}
}

read the original abstract

Multi-step manipulation tasks where robots interact with their environment and must apply process forces based on the perceived situation remain challenging to learn and prone to execution errors. Accurately simulating these tasks is also difficult. Hence, it is crucial for robust task performance to learn how to coordinate end-effector pose and applied force, monitor execution, and react to deviations. To address these challenges, we propose a learning approach that directly infers both low- and high-level task representations from user demonstrations on the real system. We developed an unsupervised task segmentation algorithm that combines intention recognition and feature clustering to infer the skills of a task. We leverage the inferred characteristic features of each skill in a novel unsupervised anomaly detection approach to identify deviations from the intended task execution. Together, these components form a comprehensive framework capable of incrementally learning task decisions and new behaviors as new situations arise. Compared to state-of-the-art learning techniques, our approach significantly reduces the required amount of training data and computational complexity while efficiently learning complex in-contact behaviors and recovery strategies. Our proposed task segmentation and anomaly detection approaches outperform state-of-the-art methods on force-based tasks evaluated on two different robotic systems.

Figures

Figures reproduced from arXiv: 2505.04565 by the authors.

**Figure 1.** The different phases and challenges of the box grasping and locking task. The upper row shows EEF configurations in the subgoal region of the respective skills during the execution which trigger a transition to the successor skills. The bottom row highlights the difficulties during each skill. In the first skill, the lower part of the gripper must not collide with the box, while the movable slides are positioned abo… view at source ↗

**Figure 2.** Our proposed incremental high- and low-level task learning framework. The process of learning a new task starts with an initial teaching sequence ⃝1 (Sec. 3.1), where the user provides several demonstrations of the intended task that are segmented using BNG-IRL to learn an initial task model. The low-level skills of this initial model can then be further refined during the skill refinement teaching sequence ⃝2 (Sec.… view at source ↗

**Figure 3.** Model and parameter inference of our proposed unsupervised task segmentation approach BNG-IRL. when taking action a, hence we can simplify the equations according to the last line. V ∗ (s, g) = max a∈A(s) Q ∗ (a, s, g) = max a∈A(s) X s ′∈S P(s ′ |s, a)(Rg(s, a, s′ ) + γV ∗ (s ′ , g)) = max a∈A(s) (Rg(s, a) + γV ∗ (s ′ , g)) (2) Similar to Michini et al. (2015) and Park et al. (2020), we use a sparse reward function … view at source ↗

Figures from the paper (15 more)

**Figure 4.** Figure 4: Simplified 2D example of subgoal-driven intention recognition based on IRL for a single demonstration. 2D example in [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]

**Figure 5.** Figure 5: , where we simulate a user demonstration of a pushing task. The robot starts from an arbitrary start configuration [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]

**Figure 6.** Figure 6: The top row shows the segmentation result of the three compared methods BNG-IRL, BNGMM (Rasmussen 1999), and BN-IRL (Michini et al. 2015), mapped onto the end-effector (EEF) trajectory of the task demonstration. The skill assignment z is represented by the color of eac…

**Figure 8.** Figure 8: The Cartesian end-effector trajectory during the user demonstration of the manipulation task. Samples with the same color are assigned to the same skill. The state subgoals of all six skills are depicted with an X. Several subgoals are inferred in the vicinity of the o…

**Figure 7.** Figure 7: The task graph structures the low-level skills on a higher level of abstraction. The green sequence of skills represents the intended task flow from the initial demonstration. The recovery behavior in red restores a situation from which the robot can continue with the …

**Figure 9.** Figure 9: Anomaly detection mechanism illustrated with the expected force during the transportation skill. In blue are the samples collected during the user demonstration ⃝1 - ⃝2 which were used to infer the skill’s expected feature region, represented by a 2D Gaussian (light bl…

**Figure 10.** Figure 10: The mean EEF poses µG,k of the skills’ subgoals in the sequence of occurrence during the task. The subgoals are defined as regions of 6D EEF poses. If the robot reaches the subgoal region of a skill, the next intended task flow skill with a new subgoal is scheduled […

**Figure 11.** Figure 11: Passive box gripper design with movable linear slides to grasp and lock a euro box. into the neutral configuration. In both cases, configuration 3 can not be reached. If configuration 3 has been reached, the gripper can move closer to the box, such that the locking pi…

**Figure 12.** Figure 12: , the further the task progresses, the more restricted the subgoal (right) and constraint regions (left) become, because the process requires more precision once contact between the robot and the box is established. As expected, the subgoal regions are located at the …

**Figure 14.** Figure 14: Comparison of the ground truth segmentation of demonstration 1 with the task segmentation results of BNGIRL, BOCPD, and AWE. BOCPD suffers from over segmentation in the final phase of the task, while AWE detects too many segments in the beginning. AWE’s chosen error …

**Figure 13.** Figure 13: Characteristic features during the initial task demonstrations, which are utilized to segment the task into skills based on consistent correlations among the demonstrations. The features define the constraint regions Ck of each skill. The colors represent the differen…

**Figure 15.** Figure 15: Force Refinement of skill 4. The relation between z-position and commanded force in z-direction for skill 4, is illustrated by the 2D excerpt of the GMM for motion- and force generation (left). The measured and commanded forces in zdirection during the entire task ar…

**Figure 16.** Figure 16: Nominal execution of skill 3. The relation between fx and x, as well as fz and z for skill 3, is illustrated by the 2D excerpts of the GMM for motion- and force generation (left). As seen on the left, the measured forces during the execution fx and fz are within the e…

**Figure 17.** Figure 17: When interfering with the task, our method consistently detects anomalies before potentially dangerous situations can occur. As shown in rows 2 and 3, when the measured forces fx and fz are outside the expected range with respect to the measured EEF pose, our unsuperv…

**Figure 18.** Figure 18: The Task Graph for the box grasping and locking task, including recovery behaviors for the different anomalies I) - IV) of skill 3. After the initial demos, the TG consists of the intended task flow skills in green. As new anomalies are identified, the TG can be incre…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

12 extracted references · 4 canonical work pages

[1]

arXiv preprint arXiv:0710.3742

Adams RP and MacKay DJ (2007) Bayesian online changepoint detection. arXiv preprint arXiv:0710.3742 . Agia C, Sinha R, Yang J, Cao Za, Antonova R, Pavone M and Bohg J (2024) Unpacking failure modes of generative policies: Runtime monitoring of consistency and progress. arXiv preprint arXiv:2410.04640 . Agrawal P, Antoniak S, Hanna EB, Bout B, Chaplot D, C...

arXiv 2007
[5]

IEEE Robotics and Automation Letters

Sliwowski D and Lee D (2024) Conditionnet: Learning precondi- tions and effects for execution monitoring. IEEE Robotics and Automation Letters . Steinmetz F, Nitsch V and Stulp F (2019) Intuitive task- level programming by demonstration through semantic skill recognition. IEEE Robotics and Automation Letters 4(4): 3742–3749. Su Z, Kroemer O, Loeb GE, Sukh...

arXiv 2024
[10]

In: 2015 IEEE international conference on robotics and automation (ICRA)

Prepared using sagej.cls 24 Journal Title XX(X) Guerin KR, Lea C, Paxton C and Hager GD (2015) A framework for end-user instruction of a robot assistant for manufacturing. In: 2015 IEEE international conference on robotics and automation (ICRA). IEEE, pp. 6167–6174. Hagos TM, Suomalainen M and Kyrki V (2018) Segmenting and sequencing of compliant motions....

arXiv 2015
[11]

Segmentation of Robot Movements using Position and Contact Forces

Karlsson M, Robertsson A and Johansson R (2019) Segmentation of robot movements using position and contact forces. arXiv preprint arXiv:1909.08289 . Khansari-Zadeh SM and Billard A (2011) Learning stable nonlinear dynamical systems with gaussian mixture models. IEEE Transactions on Robotics 27(5): 943–957. Konidaris G, Kuindersma S, Grupen R and Barto A (...

work page Pith review arXiv 2019
[45]

In: 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

Mayr M, Chatzilygeroudis K, Ahmad F, Nardi L and Krueger V (2021) Learning of parameters in behavior trees for movement skills. In: 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, pp. 7572–7579. Meier F, Theodorou E, Stulp F and Schaal S (2011) Movement segmentation using a primitive library. In: 2011 IEEE/RSJ Intern...

work page 2021
[137]

DOI:10.1016/j

Der Kiureghian A and Ditlevsen O (2009) Aleatory or epistemic? does it matter? Structural Safety 31: 105–112. DOI:10.1016/j. strusafe.2008.06.020. Driess D, Xia F, Sajjadi MS, Lynch C, Chowdhery A, Ichter B, Wahid A, Tompson J, Vuong Q, Yu T et al. (2023) Palm- e: An embodied multimodal language model. arXiv preprint arXiv:2303.03378 . Du Y , Konyushkova ...

arXiv 2009
[232]

PMLR, pp. 120–136. Eiband T, Liebl J, Willibald C and Lee D (2023a) Online task segmentation by merging symbolic and data-driven skill recognition during kinesthetic teaching. Robotics and Autonomous Systems 162: 104367. Eiband T, Saveriano M and Lee D (2019) Intuitive programming of conditional tasks by demonstration of multiple solutions. IEEE Robotics ...

arXiv 2023
[560]

(2022) A generalist agent

Reed S, Zolna K, Parisotto E, Colmenarejo SG, Novikov A, Barth- Maron G, Gimenez M, Sulsky Y , Kay J, Springenberg JT et al. (2022) A generalist agent. arXiv preprint arXiv:2205.06175 . Romeres D, Jha DK, Yerazunis W, Nikovski D and Dau HA (2019) Anomaly detection for insertion tasks in robotic assembly using gaussian process models. In: 2019 18th Europea...

arXiv 2022

Show all 12 references

[629]

In: Conference on Robot Learning

Park D, Noseworthy M, Paul R, Roy S and Roy N (2020) Inferring task goals and constraints using bayesian nonparametric inverse reinforcement learning. In: Conference on Robot Learning. pp. 1005–1014. Pastor P, Kalakrishnan M, Chitta S, Theodorou E and Schaal S (2011) Skill lea...

2020
[680]

arXiv preprint arXiv:2304.08587

Zhang X, Ding Y , Amiri S, Yang H, Kaminski A, Esselink C and Zhang S (2023) Grounding classical task planners via vision- language models. arXiv preprint arXiv:2304.08587 . Ziebart BD, Maas AL, Bagnell JA, Dey AK et al. (2008) Maximum entropy inverse reinforcement learning. I...

2023 arXiv
[1307]

Intelligent service robotics 9: 1–29

Calinon S (2016) A tutorial on task-parameterized movement learning and retrieval. Intelligent service robotics 9: 1–29. Calinon S (2020) Gaussians on riemannian manifolds: Applications for robot learning and adaptive control. IEEE Robotics & Automation Magazine 27(2): 33–45. ...

2016
[1618]

In: 2015 IEEE international conference on robotics and automation (ICRA)

Kroemer O, Daniel C, Neumann G, Van Hoof H and Peters J (2015) Towards learning hierarchical skills for multi-phase manipulation tasks. In: 2015 IEEE international conference on robotics and automation (ICRA). IEEE, pp. 1503–1510. Kr¨uger V , Tikhanoff V , Natale L and Sandini...

2015

Pith tools

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

[1] [1]

arXiv preprint arXiv:0710.3742

Adams RP and MacKay DJ (2007) Bayesian online changepoint detection. arXiv preprint arXiv:0710.3742 . Agia C, Sinha R, Yang J, Cao Za, Antonova R, Pavone M and Bohg J (2024) Unpacking failure modes of generative policies: Runtime monitoring of consistency and progress. arXiv preprint arXiv:2410.04640 . Agrawal P, Antoniak S, Hanna EB, Bout B, Chaplot D, C...

arXiv 2007

[2] [5]

IEEE Robotics and Automation Letters

Sliwowski D and Lee D (2024) Conditionnet: Learning precondi- tions and effects for execution monitoring. IEEE Robotics and Automation Letters . Steinmetz F, Nitsch V and Stulp F (2019) Intuitive task- level programming by demonstration through semantic skill recognition. IEEE Robotics and Automation Letters 4(4): 3742–3749. Su Z, Kroemer O, Loeb GE, Sukh...

arXiv 2024

[3] [10]

In: 2015 IEEE international conference on robotics and automation (ICRA)

Prepared using sagej.cls 24 Journal Title XX(X) Guerin KR, Lea C, Paxton C and Hager GD (2015) A framework for end-user instruction of a robot assistant for manufacturing. In: 2015 IEEE international conference on robotics and automation (ICRA). IEEE, pp. 6167–6174. Hagos TM, Suomalainen M and Kyrki V (2018) Segmenting and sequencing of compliant motions....

arXiv 2015

[4] [11]

Segmentation of Robot Movements using Position and Contact Forces

Karlsson M, Robertsson A and Johansson R (2019) Segmentation of robot movements using position and contact forces. arXiv preprint arXiv:1909.08289 . Khansari-Zadeh SM and Billard A (2011) Learning stable nonlinear dynamical systems with gaussian mixture models. IEEE Transactions on Robotics 27(5): 943–957. Konidaris G, Kuindersma S, Grupen R and Barto A (...

work page Pith review arXiv 2019

[5] [45]

In: 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

Mayr M, Chatzilygeroudis K, Ahmad F, Nardi L and Krueger V (2021) Learning of parameters in behavior trees for movement skills. In: 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, pp. 7572–7579. Meier F, Theodorou E, Stulp F and Schaal S (2011) Movement segmentation using a primitive library. In: 2011 IEEE/RSJ Intern...

work page 2021

[6] [137]

DOI:10.1016/j

Der Kiureghian A and Ditlevsen O (2009) Aleatory or epistemic? does it matter? Structural Safety 31: 105–112. DOI:10.1016/j. strusafe.2008.06.020. Driess D, Xia F, Sajjadi MS, Lynch C, Chowdhery A, Ichter B, Wahid A, Tompson J, Vuong Q, Yu T et al. (2023) Palm- e: An embodied multimodal language model. arXiv preprint arXiv:2303.03378 . Du Y , Konyushkova ...

arXiv 2009

[7] [232]

PMLR, pp. 120–136. Eiband T, Liebl J, Willibald C and Lee D (2023a) Online task segmentation by merging symbolic and data-driven skill recognition during kinesthetic teaching. Robotics and Autonomous Systems 162: 104367. Eiband T, Saveriano M and Lee D (2019) Intuitive programming of conditional tasks by demonstration of multiple solutions. IEEE Robotics ...

arXiv 2023

[8] [560]

(2022) A generalist agent

Reed S, Zolna K, Parisotto E, Colmenarejo SG, Novikov A, Barth- Maron G, Gimenez M, Sulsky Y , Kay J, Springenberg JT et al. (2022) A generalist agent. arXiv preprint arXiv:2205.06175 . Romeres D, Jha DK, Yerazunis W, Nikovski D and Dau HA (2019) Anomaly detection for insertion tasks in robotic assembly using gaussian process models. In: 2019 18th Europea...

arXiv 2022