Language and Learning for Robots

Robot technology will find wide-scale use only when a robotic device can be given commands and taught new tasks in a natural language. How could a robot understand instructions expressed in English? How could a robot learn from instructions? Crangle and Suppes begin to answer these questions through a theoretical approach to language and learning for robots ad by experimental work with robots.

The authors develop the notion of an instructable robot- one which derives its intelligence in part from interaction with humans. Since verbal interaction with a robot requires a natural language semantics, the authors propose a natural-model semantics which they then apply to the interpretation of robot commands. Two experimental projects are described which provide natural-language interfaces to robotic aids for the physically disabled. The authors discuss the specific challenges posed by the interpretations of "stop" commands and the interpretation of spatial prepositions.

The authors also examine the use of explicit verbal instruction to teach a robot new procedures; propose ways a robot can learn from corrective commands containing qualitative expressions, and discuss the machine-learning of a natural language use to instruct a robot in the performance of simple physical tasks. Two chapters focus on probabilistic techniques in learning.

Colleen Crangle is a researcher at CSLI. Patrick Suppes (1922–2014) was Lucie Stern Professor Emeritus of Philosophy at Stanford.

List of Tables
List of Figures
Preface
Acknowledgements

Part I Theory

1 Instructible Robots
1.1 Robots that learn from verbal instruction
1.2 Tasks and Types of robot learning
1.3 Autonomy and learning
1.4 Challenges for verbal instruction

2 Natural Models for the Interpretation of Commands
2.1 Grammars and parsing
2.2 Natural-model semantics

3 Context-fixing Semantics and Model Structures
3.1 Context-fixing semantics
3.2 Syntactic structures and model structures

4 Models for Arithmetic Instruction
4.1 The set of models
4.2 The lexicon and the grammar
4.3 Examples of context fixing
4.4 Closing remarks

Part II Language Performance

5 Verbal Commands to a Mobile Robot
5.1 Background
5.2 The class of models and procedures
5.3 The lexicon and the grammar
5.4 Command execution

6 Verbal Commands to a Robotic Arm
6.1 Introduction
6.2 The robot subsystem
6.3 The natural-language subsystem
6.4 Concluding remarks: future work

8 Extended Models: Geometric Semantics
8.1 Introduction
8.2 Some relevant geometries for prepositions
8.3 The set of geometric models
8.4 Detailed examples
8.5 Concluding remarks

Part III Learning

9 Discourse on Arithmetic Instruction
9.1 Background
9.2 Grammar and parsers
9.3 Semantics and translation
9.4 Interpretation and execution
9.5 Learning and instruction

10 Robot Learning from Corrective Instruction
10.1 Introduction
10.2 What the robot learns
10.3 Interpretation, response, and learning
10.4 A sample instruction session
10.5 Concluding remarks

11 Learning Natural Language from Robot Task Descriptions
11.1 Learning theory
11.2 Leanring principles
11.3 Internal representation
11.4 Mean learning curves
11.5 Grammars constructed from learning
11.6 Comments on related work

References
Index

2/1/94

ISBN (Paperback): 1881526194 (9781881526193)
ISBN (Cloth): 1881526208 (9781881526209)

Language and Learning for Robots

Contents

Part I Theory

Part II Language Performance

Part III Learning