De Jong, E. D. (1999). Autonomous concept formation. In Proceedings of the Sixteenth International Joint Conference on Artificial Intelligence IJCAI '99, pages 344--349, San Francisco, CA. Morgan Kaufmann. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Formalizing Distinguishability: A Contexts Based Case - Ramos, Barreiro (Correct)
....a new language combining its own with these guessings or hypothesis. These LC expressions about B represent A s own subjective view of its relation with B. Guessings or hypothesis are made by A based on its own intuition, partial information given by an external source, empirical observation [ Jong, 1999 ] or whatever. Since both agents may eventually interact, A may want to verify or refine its hypothesis by interacting with B in an appropriate way. Summarizing, A is facing a three step process: 1) making hypothesis about B s language, 2) writing down LC expressions about B using A s own ....
E.D. De Jong. Autonomous concept formation. In T. Dean, editor, Proc. 16 th International Joint Conference on Artificial Intelligence, IJCAI'99, pages 344--349, San Francisco, CA, 1999.
Grounding Language About Actions: Mobile Robots Playing Follow Me.. - Vogt (2000) (3 citations) (Correct)
....phase. This would be more realistic since we humans continuously test newly learned behaviors and adjust our capabilities according to the outcome of such a test. Simulations of such an experiment on situation concepts have already been carried out in the context of the evolution of communication (De Jong, 1999). Acknowledgments The author would like to thank Professor Luc Steels for the opportunity to work in his fine laboratory and for his supervision on this research. The members of the AI Lab in Brussels are thanked for the technical support and useful suggestions. Special thanks are for Professor ....
De Jong, E. D. (1999). Autonomous concept formation.
Grounding a Lexicon in a Coordination Task on Mobile Robots - Vogt (Correct)
....phase. This would be more realistic since we humans continuously test newly learned behaviors and adjust our capabilities according to the outcome of such a test. Simulations of such an experiment on situation concepts have already been carried out in the context of the evolution of communication (Jong, 1999). Acknowledgements The author would like to thank Professor Luc Steels for the opportunity to work in his fine laboratory and for his supervision on this research. The members of the AI Lab in Brussels are thanked for the technical support and useful suggestions. Special thanks are for Professor ....
Jong, E. D. D. (1999). Autonomous concept formation.
Attractors in the Development of Communication - de Jong (2000) Self-citation (De jong) (Correct)
....this section, an algorithm will be briefly described that, when used by individual agents, leads from an initially random situation to a state where agents refer to the situations in their environment with the same words. The stochastic algorithm is a slightly adapted version of that introduced in (De Jong, 1999), and is described in my Ph.D. thesis (De Jong, 2000) Here, the behavior of the stochastic system is analyzed by relating it to a deterministic version. 1.1 Internal Structure of the Agents Each agent has a set of meanings. These meanings represent situations in the environment that are ....
....Internal Structure of the Agents Each agent has a set of meanings. These meanings represent situations in the environment that are important to distinguish, and the current meaning can be determined directly from sensor data. They have been developed by an autonomous concept formation method, see (De Jong, 1999), and may in principle differ from agent to agent. Here however, the conditions of the experiments have been chosen such that all agents develop ideal conceptual systems, and because of this the concepts of the different agents are identical. For each meaning, an agent has a set of associations ....
De Jong, E. D. (1999). Autonomous concept formation. In Proceedings of the Sixteenth International Joint Conference on Artificial Intelligence IJCAI '99, pages 344--349, San Francisco, CA. Morgan Kaufmann.
Analyzing the Evolution of Communication from a Dynamical Systems .. - de Jong Self-citation (De jong) (Correct)
....forms (words, signals) may come about. In this paper, we will be investigating the case where these connections are learned by a group of agents in a shared environment. The concepts themselves are formed by the agents based on experience within the environment using autonomous concept formation [de Jong, 1999]. Although the fact that simple local adaptive behavior of agents causes a system of communication to emerge is interesting in itself, a deeper understanding of this process is valuable. Since the complete system of agents, concepts and forms is a complex adaptive system (see e.g. Steels, 1996a] ....
....of Situation Concepts Based on the relationship between sensor information and success, agents learn to distinguish different situations. Knowing in which situation it is allows an agent to choose optimal actions. Situation concepts are formed using autonomous concept formation, described in [de Jong, 1999]. Briefly, the method as it is used here identifies regions in the sensor action space where success is constant. Since concepts are formed based on experience, the number of concepts an agent will develop is not known in advance, and may differ among agents. Adaptation of Form Meaning ....
[Article contains additional citation context not shown here]
de Jong, E. D. (1999). Autonomous concept formation. In Proceedings of the Sixteenth International Joint Conference on Artificial Intelligence IJCAI'99.
Analyzing the Evolution of Communication from a Dynamical Systems .. - de Jong Self-citation (De jong) (Correct)
....that concepts are not given from the beginning, but are formed by agents themselves based on experience within the environment. More specifically, the concepts involved represent different situations that are useful for an agent to distinguish between. These situation concepts were defined in [de Jong, 1999b] as patterns in the interaction history between an agent and its environment that, when present, allow one to predict some aspect of the subsequent interactions. Some examples of situation concepts are the presence of sunshine after rain (which announces the possible advent of a rainbow) the ....
....is that agents develop concepts based on their experiences in the environment, and use these concepts to communicate information about the environment to other agents. This can be achieved using autonomous concept formation. 1. 3 Autonomous Concept Formation Autonomous concept formation [de Jong, 1999b] is a class of generalization methods that, based on observed relations between an agent s interaction history with the environment and some aspect of the future (the success in this case) classify all possible interaction histories into situations. It requires of a generalization method that it ....
[Article contains additional citation context not shown here]
de Jong, E. D. (1999b). Autonomous concept formation. In Proceedings of the Sixteenth International Joint Conference on Artificial Intelligence IJCAI'99.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC