《luclucsteelssteels》由会员分享,可在线阅读,更多相关《luclucsteelssteels(6页珍藏版)》请在金锄头文库上搜索。
1、Luc SteelsLuc SteelsLuc SteelsLuc SteelsSONY Computer Science Laboratory, Paris, France; and University of Brussels (VUB AI Laboratory), Belgium Corresponding author: Luc Steels (e-mail: steelsarti.vub.ac.be)The computational and robotic synthesis ofThe computational and robotic synthesis ofThe comp
2、utational and robotic synthesis ofThe computational and robotic synthesis of language evolution is emerging as a newlanguage evolution is emerging as a newlanguage evolution is emerging as a newlanguage evolution is emerging as a new exciting field of research. The objective is toexciting field of r
3、esearch. The objective is toexciting field of research. The objective is toexciting field of research. The objective is to come up with precise operational models ofcome up with precise operational models ofcome up with precise operational models ofcome up with precise operational models of how comm
4、unities of agents, equipped with ahow communities of agents, equipped with ahow communities of agents, equipped with ahow communities of agents, equipped with a cognitive apparatus, a sensori-motor system,cognitive apparatus, a sensori-motor system,cognitive apparatus, a sensori-motor system,cogniti
5、ve apparatus, a sensori-motor system, and a body, can arrive at shared groundedand a body, can arrive at shared groundedand a body, can arrive at shared groundedand a body, can arrive at shared grounded communication systems. Such systems maycommunication systems. Such systems maycommunication syste
6、ms. Such systems maycommunication systems. Such systems may have similar characteristics to animalhave similar characteristics to animalhave similar characteristics to animalhave similar characteristics to animal communication or human language. Apart fromcommunication or human language. Apart fromc
7、ommunication or human language. Apart fromcommunication or human language. Apart from its technological interest in building novelits technological interest in building novelits technological interest in building novelits technological interest in building novel applications in the domain of humanap
8、plications in the domain of humanapplications in the domain of humanapplications in the domain of humanrobot orrobot orrobot orrobot or robotrobotrobotrobotrobot interaction, trobot interaction, trobot interaction, trobot interaction, this research is ofhis research is ofhis research is ofhis resear
9、ch is of interest to the many disciplines concerned withinterest to the many disciplines concerned withinterest to the many disciplines concerned withinterest to the many disciplines concerned with the origins and evolution of language andthe origins and evolution of language andthe origins and evol
10、ution of language andthe origins and evolution of language and munication.Artificial Life (AL) developed in the early 1990s as a field investigating principles of living systems by building artificial systems that model some of their key aspects. For example, researchers have built operational model
11、s of pattern formation in biological systems (as observed on shells or in nest structures of insect societies) that generate similar structures to those found in nature 1. These models thus provide a causal insight in how these phenomena arise. The notion of a complex adaptive system, in which many
12、independent elements dynamically coordinate their activity through self-organization, has emerged as a key concept. Several other system concepts from biology, such as evolution by natural selection, structural coupling, order arising out of chaos, and network dynamics, could be operationalized and
13、shown to be relevant for understanding a wide range of natural phenomena. The interest of this work is not only theoretical. Genetic evolution, as captured in AL models known as genetic algorithms, have made it possible to solve a large set of engineering problems by using the search power implicit
14、in selectionist processes 2. Also in robotics, the AL approach, which insists on a bottom-up emergence of complexity and the role of collective dynamics in behaviour, has had an important impact and is largely responsible for the impressive jump forwards in autonomousrobots accomplished during the p
15、ast decade 3. More recently, AL models have begun to impact on research in biomolecular information processing 4.Language as a complex adaptive systemLanguage as a complex adaptive systemLanguage as a complex adaptive systemLanguage as a complex adaptive system Within the same methodological framewo
16、rk and pool of ideas, a few researchers began to build computer simulations and robotic experiments in which artificial communication systems emerge, invented and learned by artificial agents. The objective is again twofold. The first goal is scientific: to make precise models of how certain key properties of language-like communication systems might in principle originate, and how such communication systems might continue to evolve a