Novedad bibliográficaInfoling 12.15 (2018)

Autor/a: Steels, Luc L.
Título: The Talking Heads Experiment
Subtítulo: Origins of words and meanings
Año de publicación: 2016
Lugar de edición: Berlin
Editorial: Language Science Press
Descripción

The Talking Heads Experiment, conducted in the years 1999-2001, was the first large-scale experiment in which open populations of situated embodied agents created for the first time ever a new shared vocabulary by playing language games about real world scenes in front of them. The agents could teleport to different physical sites in the world through the Internet. Sites, in Antwerp, Brussels, Paris, Tokyo, London, Cambridge and several other locations were linked into the network. Humans could interact with the robotic agents either on site or remotely through the Internet and thus influence the evolving ontologies and languages of the artificial agents.

 

The present book describes in detail the motivation, the cognitive mechanisms used by the agents, the various installations of the Talking Heads, the experimental results that were obtained, and the interaction with humans. It also provides a perspective on what happened in the field after these initial groundbreaking experiments. The book is invaluable reading for anyone interested in the history of agent-based models of language evolution and the future of Artificial Intelligence.

 

About the author

 

Luc Steels is a pioneer in the use of agent-based models for exploring issues in the origins and evolution of language. He is currently ICREA research professor at the Institut de Biologia Evolutiva (UPF/CSIC) in Barcelona (Spain). Steels studied linguistics at the University of Antwerp and computer science (with specialisation in Artificial Intelligence) at MIT. He founded the Artificial Intelligence Laboratory at the Free University of Brussels (VUB) in 1983 and the Sony Computer Science Laboratory in 1996. Both laboratories have been at the forefront of scientific research into language evolution. Steels published several key articles on the computational modeling of language evolution and proposed a theory which maps concepts from evolutionary biology to the cultural level. Steels edited several books, including 'Experiments in Cultural Language Evolution' published by John Benjamins in 2011.

 

Más información y otros formatos de edición: http://langsci-press.org/catalog/book/91

Temática: Ciencia cognitiva, Lingüística cognitiva, Lingüística computacional, Pragmática, Semántica, Sintaxis, Sociolingüística

Índice

Preface 

 

I The 1999 Talking Heads book 

 

1 Introduction 
1.1 The Talking Heads experiment 
1.2 The main hypotheses 
1.3 A bottom-up approach to artificial intelligence 
1.4 History of the project 
1.5 Beyond Turing 
1.6 The book 

 

2 Preview 
2.1 The main components 
2.1.1 Teleporting 
2.1.2 The robots 
2.1.3 The agents 
2.1.4 Interactivity 
2.2 The Guessing Game 
2.2.1 Rules of the game 
2.2.2 Nature of the game 
2.2.3 The semiotic square 
2.2.4 Processes involved in language communication 
2.2.5 Knowledge sources and competences 
2.3 Perception and categorisation 
2.3.1 Scene and topic selection 
2.3.2 Sensory channels 
2.3.3 Making distinctions 
2.4 Lexicalisation
2.4.1 Same meaning, same referent 
2.4.2 A new word 
2.4.3 Competition between words 
2.4.4 Disambiguation 
2.4.5 Same meaning, different referent 
2.4.6 Situated grounded semantics 
2.5 The origins of grammar 
2.6 Conclusions 

 

3 Perception 
3.1 What sensors sense 
3.1.1 Artificial sensors and actuators
3.1.2 Natural sensing 
3.1.3 Behaviours 
3.2 Segmentation 
3.2.1 Feature extraction 
3.2.2 Divergent perception 
3.2.3 The sieve architecture 
3.3 Sensory channels
3.3.1 Example channels 
3.3.2 Conceptual spaces 
3.3.3 Perceptual constancy 
3.3.4 Transformations 
3.3.5 Scaling 
3.3.6 Saliency 
3.4 Methodology 
3.4.1 Putting up scaffolds 
3.4.2 Idealisation and realism 
3.5 The geom world 
3.6 Conclusions 

 

4 The Discrimination Game 
4.1 The paradoxes of meaning 
4.1.1 The empiricist’s stance 
4.1.2 The rationalist’s stance
4.1.3 Arguments for and against rationalism 
4.1.4 Arguments for and against empiricism 
4.2 Selectionism 
4.2.1 Principles of selectionism 
4.2.2 Selectionist cognitive systems 
4.2.3 The tree metaphor 
4.2.4 Deriving new sensory channels 
4.2.5 Comparing approaches 
4.3 Discrimination trees 
4.3.1 Making distinctions 
4.3.2 Categorisers 
4.3.3 The Discrimination Game 
4.3.4 The Pachinko machine 
4.3.5 Competition between categories 
4.3.6 Variations on discrimination 
4.3.7 The Discrimination Game in action 
4.3.8 The importance of scaling and saliency 
4.3.9 Combinations of categories 
4.3.10 A real world scene 
4.4 An ecology of distinctions 
4.4.1 Growth dynamics 
4.4.2 Pruning dynamics 
4.4.3 Average discriminatory success and repertoire size 
4.4.4 Adaptivity in categorisation 
4.4.5 Real world scenes 
4.5 Conclusions 

 

5 The Naming Game 
5.1 Inventing a lexicon 
5.1.1 Representing lexical associations 
5.1.2 Updating the score 
5.1.3 Constructing and acquiring words 
5.1.4 The Naming Game in action 
5.1.5 Characterising the lexicon 
5.1.6 Monitoring 
5.1.7 Measuring lexical coherence 
5.2 Scaling up 
5.2.1 Coping with new meanings 
5.2.2 Lexicon acquisition by virgin agents 
5.2.3 Preservation in changing populations 
5.3 Self-organisation 
5.3.1 Winner-take-all processes 
5.3.2 Collective behaviour and self-organisation
5.3.3 Increasing-returns economics 
5.3.4 Lessons from nature 
5.4 Lexical dynamics 
5.4.1 Spatially distributed naming games 
5.4.2 Language contact 
5.5 Conclusions 

 

6 The Guessing Game 
6.1 Defining the Guessing Game 
6.1.1 Example of a coupled game
6.1.2 Input-output coupling 
6.1.3 Updating the scores 
6.1.4 Repair processes 
6.2 Synonymy
6.3 Ambiguity 
6.3.1 How words may still get the same meaning 
6.3.2 How words get different meanings 
6.3.3 Competition between word meanings 
6.3.4 Lexical and ontological development 
6.4 Scaling up 
6.4.1 Increasing the population size 
6.4.2 Lexicon acquisition by new agents 
6.5 Conclusions

 

7 Grounding 
7.1 A first grounding experiment 
7.1.1 Integrating perception and action 
7.1.2 Concept acquisition 
7.1.3 Generalisation without learning 
7.1.4 The influence of the environment 
7.1.5 Coping with perceptual anomalies 
7.2 Semiotic dynamics 
7.2.1 Tracking language evolution 
7.2.2 Semiotic landscapes
7.2.3 Competition diagrams 
7.2.4 RMF coherence 
7.3 The ideal language 
7.3.1 Total coherence 
7.3.2 Communicative success despite incoherence 
7.4 Damping synonymy and ambiguity 
7.4.1 The story of fepi 
7.4.2 The story of xu 
7.4.3 The entry of O3
7.5 Rousseau’s paradox 
7.5.1 Universality versus relativism 
7.5.2 Ontological coherence
7.6 Conclusions 

 

II Installations and experiments 

 

8 The first series (1999) 
8.1 The Laboratorium exhibition 
8.2 The installation 
8.3 Start up of the experiment 
8.4 Results of the experiment 
8.5 Conclusions 

 

9 The second series (2000–2001) 
9.1 The N01SE exhibition 
9.1.1 The exhibition 
9.1.2 Installation at Kettle’s Yard in Cambridge 
9.1.3 Installation at the Wellcome Gallery in London 
9.2 Iconoclasm 
9.3 Installation at the Palais de la Découverte in Paris 
9.4 The portable Talking Heads 
9.5 Look into the Box
9.6 Conclusions 

 

III Beyond the Talking Heads 

 

10 Beyond the Talking Heads experiment 
10.1 Experiments with the aibo robots 
10.1.1 aibo’s first words 
10.1.2 The Perspective Reversal experiment 
10.2 Scaling up to grammar 
10.2.1 Early syntax experiments 
10.2.2 The Case Grammar experiments 
10.3 Conclusions

 

11 Language strategies for humanoid robots 
11.1 The Proper Naming Game 
11.1.1 Challenges 
11.1.2 Semiotic networks
11.2 Action Games 
11.3 The Colour Description Game 
11.3.1 Compositional procedural semantics with IRL 
11.3.2 Building blocks for natural language semantics 
11.3.3 Strategies for colour 
11.3.4 Translation to grammar 
11.3.5 Influence of embodiment 
11.4 Conclusion

 

12 Language evolution 
12.1 Culture-driven language evolution 
12.2 Fitness landscapes 
12.2.1 Fitness landscapes of language systems 
12.2.2 The fitness landscape of language strategies 
12.3 Selection and alignment of language strategies
12.4 Generation of new strategies 
12.4.1 A meta-strategy for generating new conceptualisation strategies 
12.4.2 Meta-strategies for generating new lexicogrammatical strategies 
12.5 Conclusions 

 

Bibliography 
Index 
Name index 
Subject index 


Colección: Computational Models of Language Evolution
Formato: PDF
Págs.: 374
ISBN-13: 9783944675428

Remitente: Infoling  <infolingantispaminfoling.org>
Fecha: 9 de diciembre de 2018

Información publicada en Infoling: http://www.infoling.org/informacion/NB1975.html



Con la ayuda de:
Editorial Arco Libros

© Infoling 1996-2019. Reservados todos los derechos.
ISSN: 1576-3404
Logo image by Hay Kranen / CC-BY