How mobile robots can self-organise a vocabulary
Paul Vogt
How mobile robots can self-organise a vocabulary
Free
Description
Contents
Reviews

One of the hardest problems in science is the symbol grounding problem, a question that has intrigued philosophers and linguists for more than a century. With the rise of artificial intelligence, the question has become very actual, especially within the field of robotics. The problem is that an agent, be it a robot or a human, perceives the world in analogue signals. Yet humans have the ability to categorise the world in symbols that they, for instance, may use for language. This book presents a series of experiments in which two robots try to solve the symbol grounding problem. The experiments are based on the language game paradigm, and involve real mobile robots that are able to develop a grounded lexicon about the objects that they can detect in their world. Crucially, neither the lexicon nor the ontology of the robots has been preprogrammed, so the experiments demonstrate how a population of embodied language users can develop their own vocabularies from scratch.

Language
English
ISBN
978-3-944675-43-5
Preface
Acknowledgments
1 Introduction
1.1 Symbol grounding problem
1.1.1 Language of thought
1.1.2 Understanding Chinese
1.1.3 Symbol grounding: philosophical or technical?
1.1.4 Grounding symbols in language
1.1.5 Physical grounding hypothesis
1.1.6 Physical symbol grounding
1.2 Language origins
1.2.1 Computational approaches to language evolution
1.2.2 Steels' approach
1.3 Language acquisition
1.4 Setting up the goals
1.5 Contributions
1.6 The book's outline
2 The sensorimotor component
2.1 The environment
2.2 The robots
2.2.1 The sensors and actuators
2.2.2 Sensor-motor board II
2.3 The Process Description Language
2.4 Cognitive architecture in PDL
2.5 Summary
3 Language games
3.1 Introduction
3.2 The language game scenario
3.3 PDL implementation
3.4 Grounded language games
3.4.1 Sensing, segmentation and feature extraction
3.4.1.1 Sensing
3.4.1.2 Segmentation
3.4.1.3 Feature extraction
3.4.2 Discrimination games
3.4.2.1 The prototype method
3.4.2.2 An example
3.4.2.3 Binary subspace method
3.4.2.4 Summary
3.4.3 Lexicon formation
3.4.3.1 Different language games
3.4.3.2 The lexicon
3.4.3.3 Joint attention
3.4.3.4 The speaker's production
3.4.3.5 The hearer's understanding
3.4.3.6 Feedback
3.4.3.7 Adaptation
3.5 Coupling categorisation and naming
4 Experimental results
4.1 Measures and methodology
4.1.1 Measures
4.1.2 Statistical testing
4.1.3 On-board versus off-board
4.2 Sensory data
4.3 The basic experiment
4.3.1 The global evolution
4.3.2 The ontological development
4.3.3 Competition diagrams
4.3.4 The lexicon
4.3.4.1 Semiotic landscape
4.3.4.2 Lexical and ontological growth
4.3.5 More language games
4.4 Summary
5 Varying methods and parameters
5.1 Impact from categorisation
5.1.1 The experiments
5.1.2 The results
5.1.3 Discussion
5.2 Impact from physical conditions and interactions
5.2.1 The experiments
5.2.2 The results
5.2.3 Discussion
5.3 Different language games
5.3.1 The experiments
5.3.2 The results
5.3.3 Discussion
5.4 The observational game
5.4.1 The experiments
5.4.2 The results
5.4.3 Discussion
5.5 Word-form creation
5.5.1 The experiments
5.5.2 The results
5.5.3 Discussion
5.6 Varying the learning rate
5.6.1 The experiments
5.6.2 The results
5.6.3 Discussion
5.7 Word-form adoption
5.7.1 The experiments
5.7.2 The results
5.7.3 Discussion
5.8 Summary
6 The optimal games
6.1 The guessing game
6.1.1 The experiments
6.1.2 The results
6.1.3 Discussion
6.1.3.1 One-to-many relations between form and meaning
6.1.3.2 Polysemy and lexical dynamics
6.2 The observational game
6.2.1 The experiment
6.2.2 The results
6.2.3 Discussion
6.3 Summary
7 Discussion
7.1 The symbol grounding problem solved?
7.1.1 Iconisation
7.1.2 Discrimination
7.1.3 Identification
7.1.4 Conclusions
7.2 No negative feedback evidence?
7.3 Situated embodiment
7.4 A behaviour-based cognitive architecture
7.5 The Talking Heads
7.5.1 The differences
7.5.2 The discussion
7.5.3 Summary
7.6 Future directions
7.7 Conclusions
Appendix A: Glossary
Appendix B: PDL code
Appendix C: Sensory data distribution
Appendix D: Lexicon and ontology
References
Indexes
Name index
Subject index
The book hasn't received reviews yet.
You May Also Like
Church Cooperation in Community Life
Free
Paul L. (Paul Leroy) Vogt
Church Cooperation in Community Life