Pith. sign in

REVIEW 3 major objections 4 minor 41 references

By coupling an LLM planner with a temporal knowledge-graph memory and exploration priced by information gain, ExRAP claims a 16.45% average success-rate gain and 3.40 fewer pending steps than LLM-Planner on VirtualHome, ALFRED, and CARLA.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review

2026-08-04 21:00 UTC pith:FSMOC5B4

load-bearing objection A strong empirical system for a genuinely new problem, but the exploration-value approximation is asserted rather than justified—worth peer review, not blind trust. the 3 major comments →

arxiv 2509.08222 v1 pith:FSMOC5B4 submitted 2025-09-10 cs.AI

Exploratory Retrieval-Augmented Planning For Continual Embodied Instruction Following

Minjong Yoo , Jinwoo Jang , Wei-jin Park , Honguk Woo This is my paper
classification cs.AI
keywords continual instruction followingembodied task planningretrieval-augmented planningtemporal knowledge graphnon-stationary environmentsexploration-exploitation balanceLLM-based agents
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

This paper targets a practical gap in embodied AI: instructions that arrive continuously — 'if the temperature is high, open the window' — in environments that keep changing, where a one-shot plan is never enough. ExRAP gives a large language model a two-sided job: answer conditional queries against a temporal knowledge-graph memory of the world, and choose skills that both complete tasks and deliberately reduce uncertainty about stale knowledge. The paper's central argument is that exploration should be priced by expected information gain, not left to the LLM's intuition, and that query confidence should decay as observations age. In VirtualHome, ALFRED, and CARLA, ExRAP reports higher success rates and fewer pending steps than ZSP, SayCan, ProgPrompt, and LLM-Planner across every tested level of non-stationarity. If the results hold, the framework points toward agents that handle ongoing user requests in real homes and roads without re-collecting the whole world state at each decision.

Core claim

ExRAP's central claim is that the bottleneck in continual embodied instruction following is memory freshness, not plan generation: when conditions change over time, an agent must know when its picture of the world is stale. The paper splits each instruction into a query against the environment plus an execution that fires when the query is satisfied, stores the world as a temporal embodied knowledge graph, and has an LLM evaluate queries against that graph while an exploration planner prices skills by how much they shrink query uncertainty. A temporal-consistency step forces evaluator confidence to fall as observations age. Across VirtualHome, ALFRED, and CARLA, ExRAP reports a 16.45% averag

What carries the argument

The load-bearing object is the temporal embodied knowledge graph (TEKG), a growing set of (source entity, relation, target entity, timestamp) quadruples that serve as the agent's environmental context memory. Retrieval from the TEKG feeds both the query evaluator and the exploitation planner, so one memory grounds condition-checking and skill choice. On top of it sits the exploration value vR(Gt,z) = sum over queries of H(P(q|Gt)) times [1 − d(Phi_R(Gz_t, {q})) / d(Phi_R(Gt, {q}))], which converts 'how much follow-up uncertainty will this skill remove' into a number the planner maximizes alongside the LLM's exploitation score. This proxy formula is the mechanism that balances keeping memory

Load-bearing premise

The exploration planner's value rests on the assumption that the graph distance between the agent and the retrieved knowledge accurately measures how much uncertainty a skill will remove, and that fully synchronized memory means zero query uncertainty; if that proxy is miscalibrated, exploration effort goes to the wrong places and the reported gains shrink.

What would settle it

Log, for each skill in a controlled VirtualHome run, the exploration value predicted by Eq. (11) against the entropy drop actually measured in the query evaluator after the skill's observation; if high-predicted-value skills consistently yield near-zero entropy drops, the graph-distance proxy does not track information gain. A second check: in a static environment, compare ExRAP against the exploitation-only variant ExRAP-EXP — if the success gap vanishes, the exploration term's value rests entirely on the decay assumption.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • Continual instructions can be handled as persistent queries over a knowledge-graph memory instead of re-planned from scratch at each step; this is why ExRAP's edge over LLM-Planner widens as the number of simultaneous instructions grows, from an 18.49-point SR gap at small scale to 22.04 points at large scale.
  • Temporal consistency is doing real work: dropping it lowers success by 15.56% on average in the ablation, so modeling knowledge decay is part of what keeps query evaluation reliable.
  • Pricing exploration by information gain pays off most when the world changes fastest: the reported SR gap over LLM-Planner widens from 15.35 points at low non-stationarity to 18.58 points at high non-stationarity.
  • The memory-plus-exploration design, not a simulator-specific trick, carries the result: the same framework reports consistent gains in household (VirtualHome, ALFRED) and driving (CARLA) domains.
  • The approach tolerates smaller language models: with Gemma-2B, ExRAP holds about 52.75% success where LLM-Planner drops to about 23.31%, suggesting memory compensation reduces dependence on model scale.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Because the exploration value uses graph distance as a stand-in for uncertainty reduction, a natural extension is metric or continuous-state environments (object positions, temperatures, moving vehicles), where the distance proxy would need re-derivation and could be validated directly against measured entropy drops.
  • The temporal-consistency rule — confidence must fall as memory ages — may misfire in genuinely periodic environments, where old observations of a regularly toggled state still carry information; per-query decay rates would be a testable refinement.
  • The reported weights (exploration value 1.0, exploitation value 0.01) make ExRAP heavily exploration-biased by construction; tuning this ratio per environment or per instruction set is an untested lever that could trade steps for success differently.
  • Robustness to a small backbone hints that the memory and exploration machinery, not the language model, does most of the work; a deployment-relevant test would run ExRAP on a robot with fixed compute and measure throughput against the reported retrieval speedup.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. ExRAP is an LLM-based planning framework for continual embodied instruction following in non-stationary environments. The agent maintains a temporal embodied knowledge graph (TEKG) as environmental memory; an instruction interpreter decomposes each continual instruction into a query and an execution; and a memory-augmented query evaluator uses retrieved TEKG quadruples to estimate query satisfaction with an entropy-based temporal consistency refinement. For planning, ExRAP combines an exploitation value (LLM in-context planning with demonstrations) with an exploration value derived from a graph-distance proxy for mutual-information reduction. The framework is evaluated in VirtualHome, ALFRED, and CARLA against ZSP, SayCan, ProgPrompt, and LLM-Planner across three non-stationarity levels, instruction scales/types, LLM sizes, and ablations. The central empirical claim is that ExRAP consistently improves success rate and reduces pending steps, e.g., 16.45% average SR gain and 3.40 PS reduction over LLM-Planner.

Significance. The empirical scope is a clear strength: three simulators, multiple non-stationarity levels, instruction types, instruction scales, LLM backbones, and ablation conditions, plus runtime-overhead measurement and qualitative analyses mapping ExRAP behavior to known heuristics. If the reported gains hold up, ExRAP would be a strong practical approach to continual instruction following, particularly because it maintains performance with small LLMs. However, the information-based exploration estimator that is central to the method is asserted rather than derived and is not validated against the quantity it approximates. Because Table A.4 makes this estimator the dominant term in skill selection, the empirical results do not currently establish the proposed exploration mechanism as the cause of the gains. The baselines are also author-adapted and no code is released, so independent replication is not yet possible. The manuscript would need additional validation or more modest claims before the central contribution is fully supported.

major comments (3)
  1. [Section 3.4, Eq. (10)] The equality \(I(Q;G_{t+1})-I(Q;G_t)=\sum_q H(P(q|G_t))-H(P(q|G_{t+1}))\) does not follow from the definitions. Mutual information expands as \(H(Q)-H(Q|G)\); the difference equals \(H(Q|G_t)-H(Q|G_{t+1})\) only if the marginal \(H(Q)\) is invariant, and \(H(Q|G)\) is not \(\sum_q H(P(q|G))\) unless the queries are independent Bernoulli variables. These assumptions are not stated. Eq. (11) then substitutes a graph-distance ratio for the true entropy reduction; the stated endpoint condition (entropy zero when memory is fully synchronized) does not imply the linear interpolation, and the distance function \(d\) is never formally defined. As written, Eqs. (10)-(11) are heuristics, not derivations.
  2. [Section 4 / Table A.4] With \(w_R=1.0\) and \(w_T=0.01\), the argmax in Eq. (12) is effectively the exploration value \(v_R\) alone. Since \(v_R\) is defined through the same query evaluator it is meant to improve, any miscalibration in Eq. (11) directly decides which skill is executed. No ablation varies \(w_R/w_T\), and no experiment replaces Eq. (11) with a computed mutual-information reduction or an alternative exploration bonus. The reported 16.45% SR gain over LLM-Planner (Section 4.1) is attributed to this exploration mechanism, so the central performance claim currently rests on an unvalidated proxy. Please add weight sweeps and a validation of Eq. (11), or soften the information-based-exploration claim.
  3. [Section 3.3, Eqs. (6)-(7)] The temporal consistency rule \(H(R(q|G_{t-1}))>H(P(q|G_{t-1}))\) is an unrestricted monotonicity assumption. Enforcing it by discarding all sampled responses that violate it can systematically bias the prior when the environment is stable or when the previous estimate was overconfident; no formal justification is given. The ablation in Table 4 shows a 15.56% average SR difference, so this scheme is load-bearing for query evaluation. At minimum, report sensitivity to the refinement procedure or compare against a properly derived Bayesian decay model.
minor comments (4)
  1. [Table 1 / Appendix E] Table 1 header spells 'CALRA'; it should be 'CARLA'. Appendix E heading 'Anaylsis' should be 'Analysis'. Section C.2 says 'hyperparameter settings for the baselines' but Table A.4 is for ExRAP.
  2. [Eq. (11)] The formula is typeset with mismatched parentheses; it should be \(1 - d(\Phi_R(G^z_t,\{q\}))/d(\Phi_R(G_t,\{q\}))\). Also, the average graph distance \(d\) is never defined (shortest path, embedding distance, etc.).
  3. [Algorithm 1] Algorithm 1 is inconsistent with Eq. (7): line 14 appends \(P(q|G_{t-1})\) after computing \(R(q|G_{t-1})\), while Eq. (7) averages the accepted \(\Phi_{LLM}\) responses. Line 19 uses \(G_{1:t-1}\) where Eq. (5) implies \(G_{1:t}\).
  4. [Appendix D.2] ExRAP-MA appears in Table A.6 without definition or discussion in the main text; please clarify whether it is a memory-augmented variant and how it differs from ExRAP.

Circularity Check

0 steps flagged

No significant circularity: the empirical SR/PS claim is external and not definitionally tied to the exploration heuristic; Eq. (11)'s unvalidated proxy is a correctness risk, not a circular step.

full rationale

The paper's central claim is empirical: ExRAP outperforms baselines on SR and PS in VirtualHome, ALFRED, and CARLA. Those metrics are measured from environment rollouts, not defined in terms of the exploration planner or the query evaluator's entropy. Eq. (10) and Eq. (11) define an exploration value as a difference/approximation of mutual information based on the query evaluator's entropy; the final policy in Eq. (12) does combine this with an LLM-based exploitation value. However, no prediction is 'fitted' and then re-reported as a result: the weights (w_R=1.0, w_T=0.01) are fixed hyperparameters, not estimated from the SR/PS data, and the exploration value is a heuristic surrogate rather than a constructed consequence of the success metric. The same-query-evaluator entropy is used both for exploration and for memory-augmented query evaluation, but this is an active-learning-style loop, not a definitional reduction of the reported outcomes to the exploration objective. The 'mild assumption' in Section 3.4 and the entropy/distance approximation are unvalidated and mathematically sloppy (Eq. (10) equates an MI difference to an entropy difference without the needed conditional-entropy identity), but these are correctness and generalization concerns, not circularity. The only self-citation is to DREAM [17] for mutual-information-based exploration, and it is not load-bearing: the paper states its own approximation and does not invoke an external uniqueness theorem or forbid alternatives. Therefore no specific circular step can be quoted, and the honest non-finding is a low score.

Axiom & Free-Parameter Ledger

5 free parameters · 5 axioms · 0 invented entities

The paper introduces the TEKG data structure and query/exploration planners, but these are algorithmic components, not new physical entities; no independent falsifiable handle outside the framework is claimed.

free parameters (5)
  • Filtering threshold theta = 0.5
    Selects executions whose query probability exceeds threshold; value chosen in Table A.4 without sensitivity analysis.
  • Exploitation weight w_T = 0.01
    Balances task value against exploration in Eq (12); hand-set in Table A.4.
  • Exploration weight w_R = 1.0
    Balances exploration value in Eq (12); hand-set in Table A.4.
  • Number of retrieved quadruples k = 12
    Retriever sample size in Section 3.3; set in Table A.4.
  • Number of query evaluator iterations = 10
    Monte Carlo sampling for prior R(q|G_{t-1}) in Eq (7); set in Table A.4.
axioms (5)
  • ad hoc to paper Entropy of query evaluator reaches zero when TEKG memory is fully synchronized with the environment
    Invoked in Section 3.4 before Eq (11) to justify approximating exploration value by distance ratio; not proven or validated.
  • ad hoc to paper Entropy of the prior query response at time t should be larger than at time t-1 (Eq 6)
    Temporal consistency constraint used to filter query priors; asserted as information decay without empirical validation beyond ablation.
  • ad hoc to paper TEKG graph distance between retrieved quadruples and the agent is a valid proxy for information gain
    Used in Eq (11) to replace the intractable mutual information difference; no theoretical justification.
  • standard math Standard probability and mutual information properties
    Eq (10) uses standard identities; assumed as background.
  • domain assumption Conditional instructions can be expressed as independent query-execution pairs through Phi_I
    Section 3.2, Eq (4): the framework assumes each instruction decomposes cleanly into a query and an execution; no mechanism for decomposition failure is provided.

reviewed 2026-08-04 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Exploratory Retrieval-Augmented Planning For Continual Embodied Instruction Following." pith.science (2026). https://pith.science/paper/FSMOC5B4

@misc{pith2026250908222,
  author       = {Pith},
  title        = {Pith review of: Exploratory Retrieval-Augmented Planning For Continual Embodied Instruction Following},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FSMOC5B4}},
  note         = {Machine review of arXiv:2509.08222}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

This study presents an Exploratory Retrieval-Augmented Planning (ExRAP) framework, designed to tackle continual instruction following tasks of embodied agents in dynamic, non-stationary environments. The framework enhances Large Language Models' (LLMs) embodied reasoning capabilities by efficiently exploring the physical environment and establishing the environmental context memory, thereby effectively grounding the task planning process in time-varying environment contexts. In ExRAP, given multiple continual instruction following tasks, each instruction is decomposed into queries on the environmental context memory and task executions conditioned on the query results. To efficiently handle these multiple tasks that are performed continuously and simultaneously, we implement an exploration-integrated task planning scheme by incorporating the {information-based exploration} into the LLM-based planning process. Combined with memory-augmented query evaluation, this integrated scheme not only allows for a better balance between the validity of the environmental context memory and the load of environment exploration, but also improves overall task performance. Furthermore, we devise a {temporal consistency refinement} scheme for query evaluation to address the inherent decay of knowledge in the memory. Through experiments with VirtualHome, ALFRED, and CARLA, our approach demonstrates robustness against a variety of embodied instruction following scenarios involving different instruction scales and types, and non-stationarity degrees, and it consistently outperforms other state-of-the-art LLM-based task planning approaches in terms of both goal success rate and execution efficiency.

Figures

Figures reproduced from arXiv: 2509.08222 by Honguk Woo, Jinwoo Jang, Minjong Yoo, Wei-Jin Park.

Figure 1
Figure 1. Figure 1: Concept of ExRAP. In the embodied environment, this framework manages continual instructions, a set of instructions for embodied instruction following tasks that are conducted continu￾ously and simultaneously. At each step, it operates through (a) memory-augmented query evaluation, and (b) exploration-integrated task planning coupled with environmental context memory updates (as shown in the left side of t… view at source ↗
Figure 2
Figure 2. Figure 2: Overall procedures of ExRAP. (a) Query evaluation: The instruction interpreter ΦI produces queries and executions, as well as a condition function from continual instructions. The memory￾augmented query evaluator ΦM then evaluates these queries probabilistically using the LLM with a retrieved TEKG from the environmental context memory. (b) Exploration-integrated task planning: The LLM-based exploitation pl… view at source ↗
Figure 3
Figure 3. Figure 3: Knowledge exploration heatmap. Darker color represents high frequency in exploration. [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

41 extracted references · 22 canonical work pages · 1 internal anchor

  1. [1]

    Language reward modulation for pretraining reinforcement learning.arXiv preprint arXiv:2308.12270, 2023

    Ademi Adeniji, Amber Xie, Carmelo Sferrazza, Younggyo Seo, Stephen James, and Pieter Abbeel. Language reward modulation for pretraining reinforcement learning.arXiv preprint arXiv:2308.12270, 2023

    Pith/arXiv arXiv 2023

  2. [2]

    Llama 3 model card

    AI@Meta. Llama 3 model card. 2024

    2024

  3. [3]

    Self-rag: Learning to retrieve, generate, and critique through self-reflection.arXiv preprint arXiv:2310.11511, 2023

    Akari Asai, Zeqiu Wu, Yizhong Wang, Avirup Sil, and Hannaneh Hajishirzi. Self-rag: Learning to retrieve, generate, and critique through self-reflection.arXiv preprint arXiv:2310.11511, 2023

    Pith/arXiv arXiv 2023

  4. [4]

    Do as i can, not as i say: Grounding language in robotic affordances

    Anthony Brohan, Yevgen Chebotar, Chelsea Finn, Karol Hausman, Alexander Herzog, Daniel Ho, Julian Ibarz, Alex Irpan, Eric Jang, Ryan Julian, et al. Do as i can, not as i say: Grounding language in robotic affordances. InProceedings of the 6th Conference on Robot Learning (CoRL), pages 287–318. PMLR, 2023

    work page 2023

  5. [5]

    Exploration by random network distillation.arXiv preprint arXiv:1810.12894, 2018

    Yuri Burda, Harrison Edwards, Amos Storkey, and Oleg Klimov. Exploration by random network distillation.arXiv preprint arXiv:1810.12894, 2018

    Pith/arXiv arXiv 2018

  6. [6]

    RegaVAE: A Retrieval-Augmented Gaussian Mixture Variational Auto-Encoder for Language Modeling

    Jingcheng Deng, Liang Pang, Huawei Shen, and Xueqi Cheng. Regavae: A retrieval- augmented gaussian mixture variational auto-encoder for language modeling.arXiv preprint arXiv:2310.10567, 2023

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  7. [7]

    CARLA: An open urban driving simulator

    Alexey Dosovitskiy, German Ros, Felipe Codevilla, Antonio Lopez, and Vladlen Koltun. CARLA: An open urban driving simulator. InProceedings of the 1st Conference on Robot Learning (CoRL), pages 1–16, 2017

    work page 2017

  8. [8]

    Lin Guan, Karthik Valmeekam, Sarath Sreedharan, and Subbarao Kambhampati. Leveraging pre-trained large language models to construct and utilize world models for model-based task planning.Proceedings of the 36th Advances in Neural Information Processing Systems (NeurIPS), 36:79081–79094, 2023

    work page 2023

  9. [9]

    Jina 10 embeddings 2: 8192-token general-purpose text embeddings for long documents.arXiv preprint arXiv:2310.19923, 2023

    Michael Günther, Jackmin Ong, Isabelle Mohr, Alaeddine Abdessalem, Tanguy Abel, Moham- mad Kalim Akram, Susana Guzman, Georgios Mastrapas, Saba Sturua, Bo Wang, et al. Jina 10 embeddings 2: 8192-token general-purpose text embeddings for long documents.arXiv preprint arXiv:2310.19923, 2023

    Pith/arXiv arXiv 2023

  10. [10]

    Reasoning with language model is planning with world model.arXiv preprint arXiv:2305.14992, 2023

    Shibo Hao, Yi Gu, Haodi Ma, Joshua Jiahua Hong, Zhen Wang, Daisy Zhe Wang, and Zhiting Hu. Reasoning with language model is planning with world model.arXiv preprint arXiv:2305.14992, 2023

    Pith/arXiv arXiv 2023

  11. [11]

    Saycanpay: Heuristic planning with large language models using learnable domain knowledge

    Rishi Hazra, Pedro Zuidberg Dos Martires, and Luc De Raedt. Saycanpay: Heuristic planning with large language models using learnable domain knowledge. InProceedings of the 42nd AAAI Conference on Artificial Intelligence, volume 38, pages 20123–20133, 2024

    work page 2024

  12. [12]

    Language models as zero-shot planners: Extracting actionable knowledge for embodied agents

    Wenlong Huang, Pieter Abbeel, Deepak Pathak, and Igor Mordatch. Language models as zero-shot planners: Extracting actionable knowledge for embodied agents. InProceedings of the 39th International Conference on Machine Learning (ICML), pages 9118–9147. PMLR, 2022

    work page 2022

  13. [13]

    Inner monologue: Embodied reasoning through planning with language models.arXiv preprint arXiv:2207.05608, 2022

    Wenlong Huang, Fei Xia, Ted Xiao, Harris Chan, Jacky Liang, Pete Florence, Andy Zeng, Jonathan Tompson, Igor Mordatch, Yevgen Chebotar, et al. Inner monologue: Embodied reasoning through planning with language models.arXiv preprint arXiv:2207.05608, 2022

    Pith/arXiv arXiv 2022

  14. [14]

    Active retrieval augmented generation.arXiv preprint arXiv:2305.06983, 2023

    Zhengbao Jiang, Frank F Xu, Luyu Gao, Zhiqing Sun, Qian Liu, Jane Dwivedi-Yu, Yiming Yang, Jamie Callan, and Graham Neubig. Active retrieval augmented generation.arXiv preprint arXiv:2305.06983, 2023

    Pith/arXiv arXiv 2023

  15. [15]

    Interactive task planning with language models.arXiv preprint arXiv:2310.10645, 2023

    Boyi Li, Philipp Wu, Pieter Abbeel, and Jitendra Malik. Interactive task planning with language models.arXiv preprint arXiv:2310.10645, 2023

    Pith/arXiv arXiv 2023

  16. [16]

    Code as policies: Language model programs for embodied control

    Jacky Liang, Wenlong Huang, Fei Xia, Peng Xu, Karol Hausman, Brian Ichter, Pete Florence, and Andy Zeng. Code as policies: Language model programs for embodied control. In Proceedings of the 40th International Conference on Robotics and Automation (ICRA), pages 9493–9500. IEEE, 2023

    work page 2023

  17. [17]

    Decoupling exploration and exploitation for meta-reinforcement learning without sacrifices

    Evan Z Liu, Aditi Raghunathan, Percy Liang, and Chelsea Finn. Decoupling exploration and exploitation for meta-reinforcement learning without sacrifices. InProceedings of the 38th International Conference on Machine Learning (ICML), pages 6925–6935. PMLR, 2021

    work page 2021

  18. [18]

    Continual queries for internet scale event-driven information delivery.IEEE Transactions on Knowledge and Data Engineering, 11(4):610–628, 1999

    Ling Liu, Calton Pu, and Wei Tang. Continual queries for internet scale event-driven information delivery.IEEE Transactions on Knowledge and Data Engineering, 11(4):610–628, 1999

    work page 1999

  19. [19]

    Text embed- dings reveal (almost) as much as text.arXiv preprint arXiv:2310.06816, 2023

    John X Morris, V olodymyr Kuleshov, Vitaly Shmatikov, and Alexander M Rush. Text embed- dings reveal (almost) as much as text.arXiv preprint arXiv:2310.06816, 2023

    Pith/arXiv arXiv 2023

  20. [20]

    Skill-based meta-reinforcement learning.arXiv preprint arXiv:2204.11828, 2022

    Taewook Nam, Shao-Hua Sun, Karl Pertsch, Sung Ju Hwang, and Joseph J Lim. Skill-based meta-reinforcement learning.arXiv preprint arXiv:2204.11828, 2022

    Pith/arXiv arXiv 2022

  21. [21]

    Do embodied agents dream of pixelated sheep: Embodied decision making using language guided world modelling

    Kolby Nottingham, Prithviraj Ammanabrolu, Alane Suhr, Yejin Choi, Hannaneh Hajishirzi, Sameer Singh, and Roy Fox. Do embodied agents dream of pixelated sheep: Embodied decision making using language guided world modelling. InProceedings of the 40th International Conference on Machine Learning (ICML), pages 26311–26325. PMLR, 2023

    work page 2023

  22. [22]

    Accelerating reinforcement learning with learned skill priors

    Karl Pertsch, Youngwoon Lee, and Joseph Lim. Accelerating reinforcement learning with learned skill priors. InProceedings of the 4th Conference on Robot Learning (CoRL), pages 188–204. PMLR, 2021

    work page 2021

  23. [23]

    Guided reinforcement learning with learned skills.arXiv preprint arXiv:2107.10253, 2021

    Karl Pertsch, Youngwoon Lee, Yue Wu, and Joseph J Lim. Guided reinforcement learning with learned skills.arXiv preprint arXiv:2107.10253, 2021

    Pith/arXiv arXiv 2021

  24. [24]

    Virtualhome: Simulating household activities via programs

    Xavier Puig, Kevin Ra, Marko Boben, Jiaman Li, Tingwu Wang, Sanja Fidler, and Antonio Torralba. Virtualhome: Simulating household activities via programs. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 8494–8502, 2018

    work page 2018

  25. [25]

    Replug: Retrieval-augmented black-box language models.arXiv preprint arXiv:2301.12652, 2023

    Weijia Shi, Sewon Min, Michihiro Yasunaga, Minjoon Seo, Rich James, Mike Lewis, Luke Zettlemoyer, and Wen-tau Yih. Replug: Retrieval-augmented black-box language models.arXiv preprint arXiv:2301.12652, 2023. 11

    Pith/arXiv arXiv 2023

  26. [26]

    Alfred: A benchmark for interpreting grounded instructions for everyday tasks

    Mohit Shridhar, Jesse Thomason, Daniel Gordon, Yonatan Bisk, Winson Han, Roozbeh Mot- taghi, Luke Zettlemoyer, and Dieter Fox. Alfred: A benchmark for interpreting grounded instructions for everyday tasks. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 10740–10749, 2020

    work page 2020

  27. [27]

    An intrinsic reward mechanism for efficient exploration

    Özgür ¸ Sim¸ sek and Andrew G Barto. An intrinsic reward mechanism for efficient exploration. InProceedings of the 23rd International Conference on Machine Learning (ICML), pages 833–840, 2006

    work page 2006

  28. [28]

    Progprompt: Generating situated robot task plans using large language models

    Ishika Singh, Valts Blukis, Arsalan Mousavian, Ankit Goyal, Danfei Xu, Jonathan Tremblay, Dieter Fox, Jesse Thomason, and Animesh Garg. Progprompt: Generating situated robot task plans using large language models. InProceedings of the 40th International Conference on Robotics and Automation (ICRA), pages 11523–11530. IEEE, 2023

    work page 2023

  29. [29]

    Llm-planner: Few-shot grounded planning for embodied agents with large language models

    Chan Hee Song, Jiaman Wu, Clayton Washington, Brian M Sadler, Wei-Lun Chao, and Yu Su. Llm-planner: Few-shot grounded planning for embodied agents with large language models. In Proceedings of the 19th International Conference on Computer Vision (ICCV), pages 2998–3009, 2023

    work page 2023

  30. [30]

    Adaplanner: Adaptive planning from feedback with language models.Proceedings of the 37th Advances in Neural Information Processing Systems (NeurIPS), 36, 2024

    Haotian Sun, Yuchen Zhuang, Lingkai Kong, Bo Dai, and Chao Zhang. Adaplanner: Adaptive planning from feedback with language models.Proceedings of the 37th Advances in Neural Information Processing Systems (NeurIPS), 36, 2024

    work page 2024

  31. [31]

    Gemma: Open models based on gemini research and technology.arXiv preprint arXiv:2403.08295, 2024

    Gemma Team, Thomas Mesnard, Cassidy Hardin, Robert Dadashi, Surya Bhupatiraju, Shreya Pathak, Laurent Sifre, Morgane Rivière, Mihir Sanjay Kale, Juliette Love, et al. Gemma: Open models based on gemini research and technology.arXiv preprint arXiv:2403.08295, 2024

    Pith/arXiv arXiv 2024

  32. [32]

    Query result caching for multiple event-driven continuous queries.Information systems, 35(1):94–110, 2010

    Yousuke Watanabe and Hiroyuki Kitagawa. Query result caching for multiple event-driven continuous queries.Information systems, 35(1):94–110, 2010

    work page 2010

  33. [33]

    Embodied task planning with large language models.arXiv preprint arXiv:2307.01848, 2023

    Zhenyu Wu, Ziwei Wang, Xiuwei Xu, Jiwen Lu, and Haibin Yan. Embodied task planning with large language models.arXiv preprint arXiv:2307.01848, 2023

    Pith/arXiv arXiv 2023

  34. [34]

    Retrieval meets long context large language models.arXiv preprint arXiv:2310.03025, 2023

    Peng Xu, Wei Ping, Xianchao Wu, Lawrence McAfee, Chen Zhu, Zihan Liu, Sandeep Subra- manian, Evelina Bakhturina, Mohammad Shoeybi, and Bryan Catanzaro. Retrieval meets long context large language models.arXiv preprint arXiv:2310.03025, 2023

    Pith/arXiv arXiv 2023

  35. [35]

    Octopus: Embodied vision-language programmer from environmental feedback.arXiv preprint arXiv:2310.08588, 2023

    Jingkang Yang, Yuhao Dong, Shuai Liu, Bo Li, Ziyue Wang, Chencheng Jiang, Haoran Tan, Jiamu Kang, Yuanhan Zhang, Kaiyang Zhou, et al. Octopus: Embodied vision-language programmer from environmental feedback.arXiv preprint arXiv:2310.08588, 2023

    Pith/arXiv arXiv 2023

  36. [36]

    Language to rewards for robotic skill synthesis.arXiv preprint arXiv:2306.08647, 2023

    Wenhao Yu, Nimrod Gileadi, Chuyuan Fu, Sean Kirmani, Kuang-Huei Lee, Montse Gonzalez Arenas, Hao-Tien Lewis Chiang, Tom Erez, Leonard Hasenclever, Jan Humplik, et al. Language to rewards for robotic skill synthesis.arXiv preprint arXiv:2306.08647, 2023

    Pith/arXiv arXiv 2023

  37. [37]

    If the green building calls, load the goods from the green building to the red building

    Wenhao Yu, Hongming Zhang, Xiaoman Pan, Kaixin Ma, Hongwei Wang, and Dong Yu. Chain-of-note: Enhancing robustness in retrieval-augmented language models.arXiv preprint arXiv:2311.09210, 2023. 12 A Broader impact Our work does not involve activities associated with negative societal impacts, such as disseminating disinformation, creating fake profiles, or ...

    Pith/arXiv arXiv 2023

  38. [40]

    If your towel isn’t stored somewhere else, put it in the closet. Timesteps: 10 Environmental knowledge: (TV , inside, livingroom, 7), (desk, inside, bedroom, 3), (apple, inside, kitchen, 9), (apple, on, kitchencounter, 9), (kitchen, adjacent, bedroom, 1), (kitchen, adjacent, bathroom, 1), (towel, inside, bathroom) ... Query response for instruction 1: Yes...

    work page

  39. [41]

    If you have an apple somewhere, bring it to your desk

    work page

  40. [42]

    If no one is watching the TV , turn it on

    work page

  41. [43]

    If your towel isn’t stored somewhere else, put it in the closet. Timesteps: 92 Query response for instruction 1: Yes: 92%, No: 8% Query response for instruction 2: Yes: 87%, No: 13% Query response for instruction 3: Yes: 50%, No: 50% Environmental knowledge: (TV , inside, livingroom, 2), (desk, inside, bedroom, 4), (apple, inside, kitchen, 89), (tv, is, o...

    work page

This paper was first reviewed by deepseek-v4-flash on August 4, 2026.