Learning Symbolic Operators for Task and Motion Planning

@article{Silver2021LearningSO,
  title={Learning Symbolic Operators for Task and Motion Planning},
  author={Tom Silver and Rohan Chitnis and Joshua B. Tenenbaum and Leslie Pack Kaelbling and Tomas Lozano-Perez},
  journal={2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  year={2021},
  pages={3182-3189}
}
Robotic planning problems in hybrid state and action spaces can be solved by integrated task and motion planners (TAMP) that handle the complex interaction between motion-level decisions and task-level plan feasibility. TAMP approaches rely on domain-specific symbolic operators to guide the task-level search, making planning efficient. In this work, we formalize and study the problem of operator learning for TAMP. Central to this study is the view that operators define a lossy abstraction of… 

Figures and Tables from this paper

Learning to Search in Task and Motion Planning with Streams
TLDR
This work proposes a geometrically informed symbolic planner that expands the set of objects and facts in a best-first manner, prior-itized by a Graph Neural Network that is learned from prior search computations.
Learning Neuro-Symbolic Skills for Bilevel Planning
TLDR
The approach — bilevel planning with neuro-symbolic skills — can solve a wide range of tasks with varying initial states, goals, and objects, outperforming six baselines and ablations.
Accelerating Integrated Task and Motion Planning with Neural Feasibility Checking
TLDR
NFC is introduced, a simple yet effective visual heuristic for classifying the feasibility of proposed actions in TAMP that will identify infeasible actions of the task planner without the need for costly motion planning, hence reducing planning time in multistep manipulation tasks.
Learning to Plan with Optimistic Action Models
TLDR
This work proposes an active learning strategy which aids the robot in learning action models quickly, with the ultimate goal of using them within a planner, and provides initial optimistic action models which are a relaxation of the underlying unknown action model.
Inventing Relational State and Action Abstractions for Effective and Efficient Bilevel Planning
Effective and efficient planning in continuous state and action spaces is fundamentally hard, even when the transition model is deterministic and known. One way to alleviate this challenge is to
Learning Neuro-Symbolic Relational Transition Models for Bilevel Planning
TLDR
Experiments show that NSRTs can be learned very data-efficiently, and then used for fast planning in new tasks that require up to 60 actions and involve many more objects than were seen during training.
Robust Planning for Multi-stage Forceful Manipulation
TLDR
This work augments an existing task and motion planner with constraints that explicitly consider torque and frictional limits, captured through the proposed forceful kinematic chain constraint, and demonstrates how the system selects from among a combinatorial set of strategies.
Toward Creative Problem Solving Agents: Action Discovery through Behavior Babbling
TLDR
This work proposes a framework for CPS by robots for discovering novel actions via behavior babbling, capable of learning a representation of novel actions at both a symbolic planning level, and a sub-symbolic action controller level.
Creative Problem Solving in Artificially Intelligent Agents: A Survey and Framework
TLDR
This work presents a definition and a framework of CPS, which is adopted to categorize existing AI methods in this field, and consists of four main components of a CPS problem: problem formulation, knowledge representation, method of knowledge manipulation, and 4) method of evaluation.
Metal Wire Manipulation Planning for 3D Curving - How a Low Payload Robot Can Use a Bending Machine to Bend Stiff Metal Wire
This paper presents a combined task and motion planner for a robot arm to carry out 3D metal wire curving tasks by collaborating with a bending machine. We assume a collaborative robot that is safe
...
...

References

SHOWING 1-10 OF 32 REFERENCES
Active Model Learning and Diverse Action Sampling for Task and Motion Planning
TLDR
This work uses Gaussian process methods for learning the conditions of operator effectiveness from small numbers of expensive training examples collected by experimentation on a robot, and develops adaptive sampling methods for generating diverse elements of continuous sets during planning for solving a new task, so that planning is as efficient as possible.
Guided search for task and motion plans using learned heuristics
TLDR
This work presents an algorithm that searches the space of possible task and motion plans and uses statistical machine learning to guide the search process and shows significant improvements in performance over prior work.
Combined task and motion planning through an extensible planner-independent interface layer
TLDR
This work proposes a new approach that uses off-the-shelf task planners and motion planners and makes no assumptions about their implementation and uses a novel representational abstraction that requires only that failures in computing a motion plan for a high-level action be identifiable and expressible in the form of logical predicates at the task level.
Learning Quickly to Plan Quickly Using Modular Meta-Learning
TLDR
This work trains on multiple tasks and learns a variety of specializers that, on a new task, can be quickly adapted using relatively little data – thus, the system learns quickly to plan quickly using these specializers.
Logic-Geometric Programming: An Optimization-Based Approach to Combined Task and Motion Planning
TLDR
This work proposes to formulate the problem of sequential robot manipulation holistically as a 1st- order logic extension of a mathematical program: a non-linear constrained program over the full world trajectory where the symbolic state-action sequence defines the (in-)equality constraints.
Learning to guide task and motion planning using score-space representation
TLDR
A learning algorithm that speeds up the search in task and motion planning problems by predicting constraints on the search space based on a generic representation of a planning problem instance, called score-space, where the problem instance is represented in terms of the performance of a set of solutions attempted so far.
Incremental Task and Motion Planning: A Constraint-Based Approach
TLDR
The Iteratively Deepened Task and Motion Planning method is probabilistically-complete and offers improved performance and generality compared to a similar, state-of-theart, probabilistic-complete planner.
Deep Visual Reasoning: Learning to Predict Action Sequences for Task and Motion Planning from an Initial Scene Image
TLDR
A deep convolutional recurrent neural network that predicts action sequences for task and motion planning (TAMP) from an initial scene image such that ideally only one motion planning problem has to be solved to find a solution to the overall TAMP problem.
Learning value functions with relational state representations for guiding task-and-motion planning
We propose a novel relational state representation and an action-value function learning algorithm that learns from planning experience for geometric task-and-motion planning (GTAMP) problems, in
SPOTTER: Extending Symbolic Planning Operators through Targeted Reinforcement Learning
TLDR
An integrated framework named SPOTTER is proposed that uses RL to augment and support a planning agent by discovering new operators needed by the agent to accomplish goals that are initially unreachable for the agent.
...
...