---
title: "Skill Induction and Planning with Latent Language — Sharma 2021"
description: "The authors propose (SL)³, which treats subtask descriptions as latent variables in a generative model. 54 references, cited by 36 papers here."
authors: ["Pratyusha Sharma", "Antonio Torralba", "Jacob Andreas"]
year: 2021
venue: "arXiv:2110.01517"
source: "https://arxiv.org/abs/2110.01517"
doi: "10.48550/arxiv.2110.01517"
arxiv: "2110.01517"
licence: "arXiv — non-exclusive license to distribute"
licence_url: "http://arxiv.org/licenses/nonexclusive-distrib/1.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/sharma-2021-skill-induction-planning-latent-language/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/sharma-2021-skill-induction-planning-latent-language/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2021](https://md.research.elamur.ai/free/en/papers/2021/page/5/) / Skill Induction and Planning with Latent Language

# Skill Induction and Planning with Latent Language

**Authors:** Pratyusha Sharma, [Antonio Torralba](https://md.research.elamur.ai/free/en/authors/antonio-b-torralba/) and [Jacob Andreas](https://md.research.elamur.ai/free/en/authors/jacob-andreas/) · **Year:** 2021 · **Venue:** [arXiv:2110.01517](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2021](https://md.research.elamur.ai/free/en/venues/acl/2021/)) · **Licence:** [arXiv — non-exclusive license to distribute](http://arxiv.org/licenses/nonexclusive-distrib/1.0/)

**Links:** [Publisher's page](https://arxiv.org/abs/2110.01517) · [PDF](https://arxiv.org/pdf/2110.01517) · [DOI 10.48550/arxiv.2110.01517](https://doi.org/10.48550/arxiv.2110.01517) · [arXiv 2110.01517](https://arxiv.org/abs/2110.01517) · [OpenAlex](https://openalex.org/W3202187802)

**Citations:** 37 citations (OpenAlex, 2026-10-14) · [36](https://md.research.elamur.ai/free/en/papers/sharma-2021-skill-induction-planning-latent-language/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary

The paper tackles a practical obstacle for hierarchical agents: policies that plan in natural language normally need large datasets in which every demonstration is labelled with step-by-step instructions and aligned to actions. The authors propose (SL)³, which treats subtask descriptions as latent variables in a generative model. A goal produces a sequence of language subtasks, and each subtask produces low-level actions. Training alternates between three steps on mostly unlabelled demonstrations: segmenting action sequences into subtasks, inferring a text label for each segment, and updating the controller (goal→instructions) and executor (instruction→actions). A small seed set of annotated demonstrations grounds the language. A language model pretrained on ungrounded text (Raffel et al., 2020) serves as a backbone. On ALFRED, with instructions for only 10% of demonstrations and none at test time, the agent roughly matches a strong state-of-the-art system and beats several methods that see ground-truth plans. It also produces readable high-level plans. The work shows that rich text pretraining plus sparse supervision can substitute for costly fine-grained annotation.

## Key points

- Evaluated in the ALFRED household simulator. Only 10% of training demonstrations carry language annotations, no action–instruction alignments are given in training, and no instructions are provided at evaluation.
- Performance is comparable to Blukis et al. (2021). It also surpasses approaches that draw on more information during training and evaluation, namely those of Kim et al. (2021), Suglia et al. (2021) and Zhang and Chai (2021).
- Ablation averages (Table 2):
  - (SL)³ with 10% annotations: 56.
  - With 100% annotations: 58.
  - With ground-truth alignments: 65.
  - Without pretraining: 49 (pretraining adds about 7 points).
  - Without latent inference: 52 (latent inference adds about 4 points).
- Offline subtask success is already good with just 5–10% annotated data. Inferring latent instructions helps most in this low-data regime.
- The planner can propose plausible plans for out-of-distribution goals, such as novel objects, verbs and subtask orders. It still fails on object affordances and on very long plans.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Massachusetts Institute of Technology](https://md.research.elamur.ai/free/en/universities/massachusetts-institute-of-technology/) — University, Cambridge, United States · 274 works · 82,013 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 19,503 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 18,805 works · [⚙️ Engineering](https://md.research.elamur.ai/free/en/shelves/engineering/) · 206 works · [⚙️ Control and Systems Engineering](https://md.research.elamur.ai/free/en/shelves/engineering/control-and-systems-engineering/) · 72 works

## Cite

Cite as: Pratyusha Sharma, Antonio Torralba and Jacob Andreas. 2021. Skill Induction and Planning with Latent Language. arXiv:2110.01517.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@misc{sharma2021skill,
  title = {{Skill Induction and Planning with Latent Language}},
  author = {Pratyusha Sharma, Antonio Torralba and Jacob Andreas},
  year = {2021},
  doi = {10.48550/arxiv.2110.01517},
  eprint = {2110.01517},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2110.01517}
}
```

- [← previous Hate Towards the Political Opponent: A Twitter Corpus Study of the 2020 US Elections on the Basis of Offensive Speech and Stance Detection](https://md.research.elamur.ai/free/en/papers/grimminger-2021-hate-political-opponent-twitter-corpus/)
- [next → Targeting the Benchmark: On Methodology in Current Natural Language Processing Research](https://md.research.elamur.ai/free/en/papers/schlangen-2021-targeting-benchmark-methodology-current-natural/)
- [↑ back to papers of 2021](https://md.research.elamur.ai/free/en/papers/2021/page/5/)

## References

54 references, 10 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Peter Anderson, Qi Wu, Damien Teney, Jake Bruce, Mark Johnson, Niko Sünderhauf, I. Reid, Stephen Gould, and A. V. Hengel. 2018. [Vision-and-language navigation: Interpreting visually-grounded navigation instructions in real environments](https://md.research.elamur.ai/free/en/papers/anderson-2017-vision-language-navigation-interpreting-visually/). 2018 IEEE/CVF Conference on Computer Vision and Pattern Recogni…
2. Jacob Andreas, D. Klein, and Sergey Levine. 2017. Modular multitask reinforcement learning with policy sketches. International Conference of Machine Learning. [find](https://scholar.google.com/scholar?q=Jacob+Andreas%2C+D%2E+Klein%2C+and+Sergey+Levine%2E+2017%2E+Modular+multitask+reinforcement+learning+with+policy+sketches%2E+International+Conference+of+Machine+Learning%2E)
3. Jacob Andreas, Dan Klein, and Sergey Levine. 2018. Learning with latent language. New Orleans, Louisiana. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Jacob+Andreas%2C+Dan+Klein%2C+and+Sergey+Levine%2E+2018%2E+Learning+with+latent+language%2E+New+Orleans%2C+Louisiana%2E+Association+for+Computational+Linguistics%2E)
4. P. Bacon, Jean Harb, and Doina Precup. 2017. The option-critic architecture. In AAAI. [find](https://scholar.google.com/scholar?q=P%2E+Bacon%2C+Jean+Harb%2C+and+Doina+Precup%2E+2017%2E+The+option%2Dcritic+architecture%2E+In+AAAI%2E)
5. L. Baum, T. Petrie, George W. Soules, and Norman Weiss. 1970. A maximization technique occurring in the statistical analysis of probabilistic functions of markov chains. Annals of Mathematical Statistics, 41:164–171. [find](https://scholar.google.com/scholar?q=L%2E+Baum%2C+T%2E+Petrie%2C+George+W%2E+Soules%2C+and+Norman+Weiss%2E+1970%2E+A+maximization+technique+occurring+in+the+statistical+analysis+of+probabilistic+functions+of+markov+chains%2E+Annals+of+Mathematical+Statist)
6. Valts Blukis, Chris Paxton, D. Fox, Animesh Garg, and Yoav Artzi. 2021. A persistent spatial semantic representation for high-level natural language instruction execution. ArXiv, abs/2107.05612. — [arXiv:2107.05612](https://arxiv.org/abs/2107.05612) [find](https://scholar.google.com/scholar?q=Valts+Blukis%2C+Chris+Paxton%2C+D%2E+Fox%2C+Animesh+Garg%2C+and+Yoav+Artzi%2E+2021%2E+A+persistent+spatial+semantic+representation+for+high%2Dlevel+natural+language+instruction+execution%2E+ArXiv%2C+abs%2F2107%2E05612%2E+%E2%80%94+%5Bar)
7. S. Branavan, Harr Chen, Luke Zettlemoyer, and R. Barzilay. 2009. Reinforcement learning for mapping instructions to actions. In ACL. [find](https://scholar.google.com/scholar?q=S%2E+Branavan%2C+Harr+Chen%2C+Luke+Zettlemoyer%2C+and+R%2E+Barzilay%2E+2009%2E+Reinforcement+learning+for+mapping+instructions+to+actions%2E+In+ACL%2E)
8. R. Brockett. 1993. Hybrid models for motion control systems. [find](https://scholar.google.com/scholar?q=R%2E+Brockett%2E+1993%2E+Hybrid+models+for+motion+control+systems%2E)
9. David L. Chen and R. Mooney. 2011. Learning to interpret natural language navigation instructions from observations. In AAAI 2011. [find](https://scholar.google.com/scholar?q=David+L%2E+Chen+and+R%2E+Mooney%2E+2011%2E+Learning+to+interpret+natural+language+navigation+instructions+from+observations%2E+In+AAAI+2011%2E)
10. Rodolfo Corona, Daniel Fried, Coline Devin, D. Klein, and Trevor Darrell. 2021. Modular networks for compositional instruction following. In NAACL. [find](https://scholar.google.com/scholar?q=Rodolfo+Corona%2C+Daniel+Fried%2C+Coline+Devin%2C+D%2E+Klein%2C+and+Trevor+Darrell%2E+2021%2E+Modular+networks+for+compositional+instruction+following%2E+In+NAACL%2E)
11. Christian Daniel, G. Neumann, and Jan Peters. 2012. Hierarchical relative entropy policy search. J. Mach. Learn. Res., 17:93:1–93:50. [find](https://scholar.google.com/scholar?q=Christian+Daniel%2C+G%2E+Neumann%2C+and+Jan+Peters%2E+2012%2E+Hierarchical+relative+entropy+policy+search%2E+J%2E+Mach%2E+Learn%2E+Res%2E%2C+17%3A93%3A1%E2%80%9393%3A50%2E)
12. Abhishek Das, Samyak Datta, Georgia Gkioxari, Stefan Lee, Devi Parikh, and Dhruv Batra. 2018. Embodied question answering. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pages 2135–213509. [find](https://scholar.google.com/scholar?q=Abhishek+Das%2C+Samyak+Datta%2C+Georgia+Gkioxari%2C+Stefan+Lee%2C+Devi+Parikh%2C+and+Dhruv+Batra%2E+2018%2E+Embodied+question+answering%2E+2018+IEEE%2FCVF+Conference+on+Computer+Vision+and+Pattern+Recognition+Workshops)
13. P. Dayan and Geoffrey E. Hinton. 1992. Feudal reinforcement learning. In NIPS. [find](https://scholar.google.com/scholar?q=P%2E+Dayan+and+Geoffrey+E%2E+Hinton%2E+1992%2E+Feudal+reinforcement+learning%2E+In+NIPS%2E)
14. Thomas G Dietterich. 1999. Hierarchical reinforcement learning with the MAXQ value function decomposition. [find](https://scholar.google.com/scholar?q=Thomas+G+Dietterich%2E+1999%2E+Hierarchical+reinforcement+learning+with+the+MAXQ+value+function+decomposition%2E)
15. R. Fikes, P. Hart, and N. Nilsson. 1972. [Learning and executing generalized robot plans](https://md.research.elamur.ai/free/en/papers/fikes-1972-learning-executing-generalized-robot-plans/). Artif. Intell., 3:251–288.
16. Roy Fox, S. Krishnan, I. Stoica, and Ken Goldberg. 2017. Multi-level discovery of deep options. ArXiv, abs/1703.08294. [find](https://scholar.google.com/scholar?q=Roy+Fox%2C+S%2E+Krishnan%2C+I%2E+Stoica%2C+and+Ken+Goldberg%2E+2017%2E+Multi%2Dlevel+discovery+of+deep+options%2E+ArXiv%2C+abs%2F1703%2E08294%2E)
17. Andrea Frome, G. Corrado, Jonathon Shlens, Samy Bengio, J. Dean, Marc’Aurelio Ranzato, and Tomas Mikolov. 2013. Devise: A deep visual-semantic embedding model. In NIPS. [find](https://scholar.google.com/scholar?q=Andrea+Frome%2C+G%2E+Corrado%2C+Jonathon+Shlens%2C+Samy+Bengio%2C+J%2E+Dean%2C+Marc%E2%80%99Aurelio+Ranzato%2C+and+Tomas+Mikolov%2E+2013%2E+Devise%3A+A+deep+visual%2Dsemantic+embedding+model%2E+In+NIPS%2E)
18. Daniel Gordon, Aniruddha Kembhavi, Mohammad Rastegari, Joseph Redmon, D. Fox, and Ali Farhadi. 2018. Iqa: Visual question answering in interactive environments. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4089– 4098. [find](https://scholar.google.com/scholar?q=Daniel+Gordon%2C+Aniruddha+Kembhavi%2C+Mohammad+Rastegari%2C+Joseph+Redmon%2C+D%2E+Fox%2C+and+Ali+Farhadi%2E+2018%2E+Iqa%3A+Visual+question+answering+in+interactive+environments%2E+2018+IEEE%2FCVF+Conference+on+Computer+Vi)
19. Kaiming He, X. Zhang, Shaoqing Ren, and Jian Sun. 2016. [Deep residual learning for image recognition](https://md.research.elamur.ai/free/en/papers/he-2015-deep-residual-learning-image-recognition/). 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 770–778.
20. M. Hoffman, David M. Blei, Chong Wang, and J. Paisley. 2013. Stochastic variational inference. ArXiv, abs/1206.7051. [find](https://scholar.google.com/scholar?q=M%2E+Hoffman%2C+David+M%2E+Blei%2C+Chong+Wang%2C+and+J%2E+Paisley%2E+2013%2E+Stochastic+variational+inference%2E+ArXiv%2C+abs%2F1206%2E7051%2E)
21. Hengyuan Hu, Denis Yarats, Qucheng Gong, Yuandong Tian, and M. Lewis. 2019. Hierarchical decision making by generating and following natural language instructions. In NeurIPS. [find](https://scholar.google.com/scholar?q=Hengyuan+Hu%2C+Denis+Yarats%2C+Qucheng+Gong%2C+Yuandong+Tian%2C+and+M%2E+Lewis%2E+2019%2E+Hierarchical+decision+making+by+generating+and+following+natural+language+instructions%2E+In+NeurIPS%2E)
22. Athul Paul Jacob, M. Lewis, and Jacob Andreas. 2021. Multitasking inhibits semantic drift. ArXiv, abs/2104.07219. [find](https://scholar.google.com/scholar?q=Athul+Paul+Jacob%2C+M%2E+Lewis%2C+and+Jacob+Andreas%2E+2021%2E+Multitasking+inhibits+semantic+drift%2E+ArXiv%2C+abs%2F2104%2E07219%2E)
23. Yiding Jiang, S. Gu, K. Murphy, and Chelsea Finn. 2019. Language as an abstraction for hierarchical deep reinforcement learning. In NeurIPS. [find](https://scholar.google.com/scholar?q=Yiding+Jiang%2C+S%2E+Gu%2C+K%2E+Murphy%2C+and+Chelsea+Finn%2E+2019%2E+Language+as+an+abstraction+for+hierarchical+deep+reinforcement+learning%2E+In+NeurIPS%2E)
24. L P Kaelbling and T Lozano-Pérez. 2011. Hierarchical task and motion planning in the now. 2011 IEEE International. [find](https://scholar.google.com/scholar?q=L+P+Kaelbling+and+T+Lozano%2DP%C3%A9rez%2E+2011%2E+Hierarchical+task+and+motion+planning+in+the+now%2E+2011+IEEE+International%2E)
25. Byeonghwi Kim, Suvaansh Bhambri, Kunal Pratap Singh, Roozbeh Mottaghi, and Jonghyun Choi. 2021. Agent with the big picture: Perceiving surroundings for interactive instruction following. In Embodied AI Workshop CVPR. [find](https://scholar.google.com/scholar?q=Byeonghwi+Kim%2C+Suvaansh+Bhambri%2C+Kunal+Pratap+Singh%2C+Roozbeh+Mottaghi%2C+and+Jonghyun+Choi%2E+2021%2E+Agent+with+the+big+picture%3A+Perceiving+surroundings+for+interactive+instruction+following%2E+In+Embodied+A)
26. Diederik P. Kingma and Max Welling. 2014. [Autoencoding variational bayes](https://md.research.elamur.ai/free/en/papers/kingma-2013-auto-encoding-variational-bayes/). CoRR, abs/1312.6114.
27. G. Konidaris, S. Kuindersma, R. Grupen, and A. Barto. 2012. Robot learning from demonstration by constructing skill trees. The International Journal of Robotics Research, 31:360 – 375. [find](https://scholar.google.com/scholar?q=G%2E+Konidaris%2C+S%2E+Kuindersma%2C+R%2E+Grupen%2C+and+A%2E+Barto%2E+2012%2E+Robot+learning+from+demonstration+by+constructing+skill+trees%2E+The+International+Journal+of+Robotics+Research%2C+31%3A360+%E2%80%93+375%2E)
28. Tejas D. Kulkarni, Karthik Narasimhan, A. Saeedi, and J. Tenenbaum. 2016. Hierarchical deep reinforcement learning: Integrating temporal abstraction and intrinsic motivation. In NIPS. [find](https://scholar.google.com/scholar?q=Tejas+D%2E+Kulkarni%2C+Karthik+Narasimhan%2C+A%2E+Saeedi%2C+and+J%2E+Tenenbaum%2E+2016%2E+Hierarchical+deep+reinforcement+learning%3A+Integrating+temporal+abstraction+and+intrinsic+motivation%2E+In+NIPS%2E)
29. I. Loshchilov and F. Hutter. 2019. Decoupled weight decay regularization. In ICLR. [find](https://scholar.google.com/scholar?q=I%2E+Loshchilov+and+F%2E+Hutter%2E+2019%2E+Decoupled+weight+decay+regularization%2E+In+ICLR%2E)
30. Chih-Yao Ma, Jiasen Lu, Zuxuan Wu, G. Al-Regib, Z. Kira, R. Socher, and Caiming Xiong. 2019. Self-monitoring navigation agent via auxiliary progress estimation. ArXiv, abs/1901.03035. [find](https://scholar.google.com/scholar?q=Chih%2DYao+Ma%2C+Jiasen+Lu%2C+Zuxuan+Wu%2C+G%2E+Al%2DRegib%2C+Z%2E+Kira%2C+R%2E+Socher%2C+and+Caiming+Xiong%2E+2019%2E+Self%2Dmonitoring+navigation+agent+via+auxiliary+progress+estimation%2E+ArXiv%2C+abs%2F1901%2E03035%2E)
31. A. McGovern and A. Barto. 2001. Automatic discovery of subgoals in reinforcement learning using diverse density. In ICML. [find](https://scholar.google.com/scholar?q=A%2E+McGovern+and+A%2E+Barto%2E+2001%2E+Automatic+discovery+of+subgoals+in+reinforcement+learning+using+diverse+density%2E+In+ICML%2E)
32. So Yeon Min, Devendra Singh Chaplot, Pradeep Ravikumar, Yonatan Bisk, and Ruslan Salakhutdinov. 2021. FILM: following instructions in language with modular methods. CoRR, abs/2110.07342. [find](https://scholar.google.com/scholar?q=So+Yeon+Min%2C+Devendra+Singh+Chaplot%2C+Pradeep+Ravikumar%2C+Yonatan+Bisk%2C+and+Ruslan+Salakhutdinov%2E+2021%2E+FILM%3A+following+instructions+in+language+with+modular+methods%2E+CoRR%2C+abs%2F2110%2E07342%2E)
33. Dipendra Kumar Misra, J. Langford, and Yoav Artzi. 2017. [Mapping instructions and visual observations to actions with reinforcement learning](https://md.research.elamur.ai/free/en/papers/misra-2017-mapping-instructions-visual-observations-actions/). In EMNLP.
34. A. Newell. 1973. Human problem solving. [find](https://scholar.google.com/scholar?q=A%2E+Newell%2E+1973%2E+Human+problem+solving%2E)
35. Alexander Pashevich, Cordelia Schmid, and Chen Sun. 2021. Episodic transformer for vision-and-language navigation. CoRR, abs/2105.06453. — [arXiv:2105.06453](https://arxiv.org/abs/2105.06453) [find](https://scholar.google.com/scholar?q=Alexander+Pashevich%2C+Cordelia+Schmid%2C+and+Chen+Sun%2E+2021%2E+Episodic+transformer+for+vision%2Dand%2Dlanguage+navigation%2E+CoRR%2C+abs%2F2105%2E06453%2E+%E2%80%94+%5BarXiv%3A2105%2E06453%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2105%2E06453%29)
36. Lawrence R. Rabiner. 1989. A tutorial on hidden markov models and selected applications. Proceedings of the IEEE. [find](https://scholar.google.com/scholar?q=Lawrence+R%2E+Rabiner%2E+1989%2E+A+tutorial+on+hidden+markov+models+and+selected+applications%2E+Proceedings+of+the+IEEE%2E)
37. Colin Raffel, Noam M. Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, W. Li, and Peter J. Liu. 2020. [Exploring the limits of transfer learning with a unified text-to-text transformer](https://md.research.elamur.ai/free/en/papers/raffel-2019-exploring-limits-transfer-learning-unified/). ArXiv, abs/1910.10683.
38. E. Sacerdoti. 1973. Planning in a hierarchy of abstraction spaces. Artif. Intell., 5:115–135. [find](https://scholar.google.com/scholar?q=E%2E+Sacerdoti%2E+1973%2E+Planning+in+a+hierarchy+of+abstraction+spaces%2E+Artif%2E+Intell%2E%2C+5%3A115%E2%80%93135%2E)
39. K. Shiarlis, Markus Wulfmeier, S. Salter, S. Whiteson, and I. Posner. 2018. Taco: Learning task decomposition via temporal alignment for control. In ICML. [find](https://scholar.google.com/scholar?q=K%2E+Shiarlis%2C+Markus+Wulfmeier%2C+S%2E+Salter%2C+S%2E+Whiteson%2C+and+I%2E+Posner%2E+2018%2E+Taco%3A+Learning+task+decomposition+via+temporal+alignment+for+control%2E+In+ICML%2E)
40. Mohit Shridhar, Jesse Thomason, Daniel Gordon, Yonatan Bisk, Winson Han, Roozbeh Mottaghi, Luke Zettlemoyer, and Dieter Fox. 2020. [ALFRED: A Benchmark for Interpreting Grounded Instructions for Everyday Tasks](https://md.research.elamur.ai/free/en/papers/shridhar-2019-alfred-benchmark-interpreting-grounded-instructions/). In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR).
41. Mohit Shridhar, Xingdi Yuan, Marc-Alexandre Côté, Yonatan Bisk, Adam Trischler, and M. Hausknecht. 2021. [Alfworld: Aligning text and embodied environments for interactive learning](https://md.research.elamur.ai/free/en/papers/shridhar-2020-alfworld-aligning-text-embodied-environments/). ArXiv, abs/2010.03768.
42. Tianmin Shu, Caiming Xiong, and R. Socher. 2018. Hierarchical and interpretable skill acquisition in multi-task reinforcement learning. ArXiv, abs/1712.07294. [find](https://scholar.google.com/scholar?q=Tianmin+Shu%2C+Caiming+Xiong%2C+and+R%2E+Socher%2E+2018%2E+Hierarchical+and+interpretable+skill+acquisition+in+multi%2Dtask+reinforcement+learning%2E+ArXiv%2C+abs%2F1712%2E07294%2E)
43. Kunal Pratap Singh, Suvaansh Bhambri, Byeonghwi Kim, Roozbeh Mottaghi, and Jonghyun Choi. 2020. Moca: A modular object-centric approach for interactive instruction following. arXiv preprint arXiv:2012.03208. — [arXiv:2012.03208](https://arxiv.org/abs/2012.03208) [find](https://scholar.google.com/scholar?q=Kunal+Pratap+Singh%2C+Suvaansh+Bhambri%2C+Byeonghwi+Kim%2C+Roozbeh+Mottaghi%2C+and+Jonghyun+Choi%2E+2020%2E+Moca%3A+A+modular+object%2Dcentric+approach+for+interactive+instruction+following%2E+arXiv+preprint+arXiv%3A2012)
44. Alessandro Suglia, Qiaozi Gao, Jesse Thomason, Govind Thattai, and Gaurav Sukhatme. 2021. Embodied BERT: A transformer model for embodied, language-guided visual task completion. arXiv. [find](https://scholar.google.com/scholar?q=Alessandro+Suglia%2C+Qiaozi+Gao%2C+Jesse+Thomason%2C+Govind+Thattai%2C+and+Gaurav+Sukhatme%2E+2021%2E+Embodied+BERT%3A+A+transformer+model+for+embodied%2C+language%2Dguided+visual+task+completion%2E+arXiv%2E)
45. R S Sutton, D Precup, and S Singh. 1999. Between MDPs and semi-MDPs: A framework for temporal abstraction in reinforcement learning. Artif. Intell. [find](https://scholar.google.com/scholar?q=R+S+Sutton%2C+D+Precup%2C+and+S+Singh%2E+1999%2E+Between+MDPs+and+semi%2DMDPs%3A+A+framework+for+temporal+abstraction+in+reinforcement+learning%2E+Artif%2E+Intell%2E)
46. Stefanie Tellex, N. Gopalan, H. Kress-Gazit, and Cynthia Matuszek. 2020. Robots that use language. [find](https://scholar.google.com/scholar?q=Stefanie+Tellex%2C+N%2E+Gopalan%2C+H%2E+Kress%2DGazit%2C+and+Cynthia+Matuszek%2E+2020%2E+Robots+that+use+language%2E)
47. Stefanie Tellex, T. Kollar, Steven Dickerson, Matthew R. Walter, A. Banerjee, S. Teller, and N. Roy. 2011. Understanding natural language commands for robotic navigation and mobile manipulation. In AAAI. [find](https://scholar.google.com/scholar?q=Stefanie+Tellex%2C+T%2E+Kollar%2C+Steven+Dickerson%2C+Matthew+R%2E+Walter%2C+A%2E+Banerjee%2C+S%2E+Teller%2C+and+N%2E+Roy%2E+2011%2E+Understanding+natural+language+commands+for+robotic+navigation+and+mobile+manipulation%2E+In+AA)
48. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Lukasz Kaiser, and Illia Polosukhin. 2017. [Attention is all you need](https://md.research.elamur.ai/free/en/papers/vaswani-2017-attention-all-need/).
49. Martin J. Wainwright and M.I. Jordan. 2008. Graphical models, exponential families, and variational inference. Found. Trends Mach. Learn., 1:1–305. [find](https://scholar.google.com/scholar?q=Martin+J%2E+Wainwright+and+M%2EI%2E+Jordan%2E+2008%2E+Graphical+models%2C+exponential+families%2C+and+variational+inference%2E+Found%2E+Trends+Mach%2E+Learn%2E%2C+1%3A1%E2%80%93305%2E)
50. Sam Wiseman, S. Shieber, and Alexander M. Rush. 2018. [Learning neural templates for text generation](https://md.research.elamur.ai/free/en/papers/wiseman-2018-learning-neural-templates-text-generation/). ArXiv, abs/1808.10122.
51. Catherine Wong, Kevin Ellis, J. Tenenbaum, and Jacob Andreas. 2021. Leveraging language to learn program abstractions and search heuristics. In ICML. [find](https://scholar.google.com/scholar?q=Catherine+Wong%2C+Kevin+Ellis%2C+J%2E+Tenenbaum%2C+and+Jacob+Andreas%2E+2021%2E+Leveraging+language+to+learn+program+abstractions+and+search+heuristics%2E+In+ICML%2E)
52. Rowan Zellers, Ari Holtzman, Matthew E. Peters, R. Mottaghi, Aniruddha Kembhavi, Ali Farhadi, and Yejin Choi. 2021. Piglet: Language grounding through neuro-symbolic interaction in a 3d world. In ACL/IJCNLP. [find](https://scholar.google.com/scholar?q=Rowan+Zellers%2C+Ari+Holtzman%2C+Matthew+E%2E+Peters%2C+R%2E+Mottaghi%2C+Aniruddha+Kembhavi%2C+Ali+Farhadi%2C+and+Yejin+Choi%2E+2021%2E+Piglet%3A+Language+grounding+through+neuro%2Dsymbolic+interaction+in+a+3d+world%2E+In+ACL%2F)
53. Yichi Zhang and Joyce Chai. 2021. Hierarchical task learning from language instructions with unified transformers and self-monitoring. CoRR, abs/2106.03427. [find](https://scholar.google.com/scholar?q=Yichi+Zhang+and+Joyce+Chai%2E+2021%2E+Hierarchical+task+learning+from+language+instructions+with+unified+transformers+and+self%2Dmonitoring%2E+CoRR%2C+abs%2F2106%2E03427%2E)
54. Brian D Ziebart, J Andrew Bagnell, and Anind K Dey. 2013. The principle of maximum causal entropy for estimating interacting processes. IEEE Transactions on Information Theory, 59(4):1966–1980. [find](https://scholar.google.com/scholar?q=Brian+D+Ziebart%2C+J+Andrew+Bagnell%2C+and+Anind+K+Dey%2E+2013%2E+The+principle+of+maximum+causal+entropy+for+estimating+interacting+processes%2E+IEEE+Transactions+on+Information+Theory%2C+59%284%29%3A1966%E2%80%931980%2E)

*Updated 3 November 2026*