Adaptive Agents in Minecraft: A Hybrid Paradigm for Combining Domain Knowledge with Reinforcement Learning

Parashar, Priyam; Sheneman, Bradley; Goel, Ashok K.

doi:10.1007/978-3-319-71679-4_6

“…We conclude that knowledge transfer using metareasoning makes a robotic system more flexible than one with only classical planning. On the other hand, previous results [42] suggest that learning with metareasoning requires more structured knowledge but less data.…”

Section: Discussionmentioning

confidence: 87%

“…Jones and Goel [29] present "Empirical Verification Procedures" which ground all highlevel concepts and axioms known to the agent in lower-level precepts in a video game. Prior work [32,42,43], combines metareasoning with reinforcement learning using purely visual form expectations. However, they still use symbolic descriptions or computerized descriptions of visuals which simplifies the perception part of the problem.…”

Section: Related Workmentioning

confidence: 99%

Using meta-reasoning for incremental repairs in multi-object robot manipulation tasks

Parashar

¹

,

Goel

²

,

Christensen

³

2022

Front. Phys.

Self Cite

0

View full text Add to dashboard Cite

Robots tasked with object assembly by manipulation of parts require not only a high-level plan for order of placement of parts but also detailed low-level information on how to place and pick the part based on its state. This is a complex multi-level problem prone to failures at various levels. This paper employs meta reasoning architecture along with robotics principles and proposes dual encoding of state expectations during the progression of task to ground nominal scenarios. We present our results on table-top scenario using perceptual expectations based in the concept of occupancy grids and key point representations. Our results in a constrained manipulation setting suggest using low-level information or high-level expectations alone the system performs worse than if the architecture uses them both. We then outline a complete architecture and system which tackles this problem for repairing more generic assembly plans with objects moving in spaces with 6 degrees of freedom.

show abstract

“…The Minecraft domain has been used extensively for various tasks in AI (Aluru et al, 2015;Parashar et al, 2017), including planning (Roberts et al, 2017) and natural language understanding (Gray et al, 2019). Regarding NLG specifically, Narayan-Chen et al ( 2019) trained a neural model to generate building instructions in Minecraft; in the absence of symbolic domain knowledge, their model struggles to generate correct instructions.…”

Section: Related Workmentioning

confidence: 99%

Generating Instructions at Different Levels of Abstraction

Köhn¹,

Wichlacz²,

Torralba³

et al. 2020

Proceedings of the 28th International Conference on Computational Linguistics

View full text Add to dashboard Cite

When generating technical instructions, it is often convenient to describe complex objects in the world at different levels of abstraction. A novice user might need an object explained piece by piece, while for an expert, talking about the complex object (e. g. a wall or railing) directly may be more succinct and efficient. We show how to generate building instructions at different levels of abstraction in Minecraft. We introduce the use of hierarchical planning to this end, a method from AI planning which can capture the structure of complex objects neatly. A crowdsourcing evaluation shows that the choice of abstraction level matters to users, and that an abstraction strategy which balances low-level and high-level object descriptions compares favorably to ones which don't.

show abstract

“…The Minecraft domain has been used extensively for various tasks in AI (Aluru et al, 2015;Parashar et al, 2017), including planning (Roberts et al, 2017) and natural language understanding (Gray et al, 2019). Regarding NLG specifically, Narayan-Chen et al ( 2019) trained a neural model to generate building instructions in Minecraft; in the absence of symbolic domain knowledge, their model struggles to generate correct instructions.…”

Section: Related Workmentioning

confidence: 99%

Generating Instructions at Different Levels of Abstraction

Köhn¹,

Wichlacz²,

Torralba³

et al. 2020

Preprint

0

View full text Add to dashboard Cite

When generating technical instructions, it is often convenient to describe complex objects in the world at different levels of abstraction. A novice user might need an object explained piece by piece, while for an expert, talking about the complex object (e. g. a wall or railing) directly may be more succinct and efficient. We show how to generate building instructions at different levels of abstraction in Minecraft. We introduce the use of hierarchical planning to this end, a method from AI planning which can capture the structure of complex objects neatly. A crowdsourcing evaluation shows that the choice of abstraction level matters to users, and that an abstraction strategy which balances low-level and high-level object descriptions compares favorably to ones which don't.

show abstract

Adaptive Agents in Minecraft: A Hybrid Paradigm for Combining Domain Knowledge with Reinforcement Learning

Cited by 5 publications

References 18 publications

Using meta-reasoning for incremental repairs in multi-object robot manipulation tasks

Using meta-reasoning for incremental repairs in multi-object robot manipulation tasks

Generating Instructions at Different Levels of Abstraction

Generating Instructions at Different Levels of Abstraction

Contact Info

Product

Resources

About