There is a growing body of research on multi-agent systems bootstrapping a communication system. Most studies are based on simulation, but recently there has been an increased interest in the properties and formal analysis of these systems. Although

An exploratory study of a Map Task dialogue indicates that dialogue participants coordinate on an ad-hoc vocabulary and associated concepts (meanings) to enable information exchange, and that ad-hoc vocabularies can be cobbled together from a heterogeneous mix of “micro-vocabularies ” borrowed from various other (a priori unrelated) domains. To account for these observations, we sketch a basic framework for formalising the process of coordination of semantic systems in dialogue, and relate this framework to some interactional processes of semantic coordination in dialogue, such as feedback, negotiation and accommodation. 1 Vocabulary in a Map Task dialogue In the Map Task corpus 1, a GIVER explains a route, provided on the giver’s map, to a FOLLOWER who has a similar (but slightly different) map but with no route marked. A map contains landmarks portrayed as labelled line drawings. In a route-giving task like that recorded in the Map Task corpus, expressions referring to landmarks, compass directions etc. can be a priori expected as a kind of “prototype ” devices for talking about maps. A typical utterance may look as follows 2: GIVER: right a camera shop, right, head due south... from that just... down for about twelve centimetres, have you got a parked van at the bottom?

How is information transmitted in a group? A multi-agent connectionist model is proposed that combines features of standard recurrent models to simulate the process of information uptake, integration and memorization within individual agents, with novel aspects that simulate the communication of beliefs and opinions between agents. A crucial aspect in belief updating based on information from other agents is the trust in the information provided, implemented as the consistency with the receiving agents’ existing beliefs. Trust leads to a selective propagation and thus filtering out of less reliable information, and implements Grice’s (1975) maxims of quality and quantity in communication. By studying these communicative aspects within the framework of standard models of information processing, the unique contribution of communicative mechanisms beyond intra-personal factors was explored in simulations of key phenomena involving persuasive communication and polarization, lexical acquisition, spreading of stereotypes and rumors, and a lack of sharing unique information in group decisions.

Abstract. Language evolution takes place at two levels: the level of language strategies, which are ways in which a particular subarea of meaning and function is structured and expressed, and the level of concrete linguistic choices for the meanings, words, or grammatical constructions that instantiate a particular language strategy. It is now reasonably well understood how a shared language strategy enables a population of agents to self-organise a shared language system. But the origins and evolution of strategies has so far been explored less. This paper proposes that linguistic selection, i.e. selection driven by communicative success and cognitive effort, is relevant and shows a concrete case study for the domain of colour on how different language strategies may cooperate and compete for dominance in a population. 1 Language Strategies and Language Systems Human languages are complex adaptive systems that

Abstract — Colour naming games are idealised communicative interactions within a population of artificial agents in which a speaker uses a single colour term to draw the attention of a hearer to a particular object in a shared context. Through a series of such games, a colour lexicon can be developed that is sufficiently shared to allow for successful communication, even when the agents start out without any predefined categories. In previous models of colour naming games, the shared context was typically artificially generated from a set of colour stimuli and both agents in the interaction perceive this environment in an identical way. In this paper, we investigate the dynamics of the colour naming game in a robotic setup in which humanoid robots perceive a set of colourful objects from their own perspective. We compare the resulting colour ontologies to those found in human languages and show how these ontologies reflect

Abstract not found

We introduce an approach for physically-grounded natural language interpretation by robots which reacts appropriately to unanticipated physical changes in the environment and dynamically assimilates new information pertinent to ongoing tasks. At the core of the approach is a model of object schemas that enables a robot to encode beliefs about physical objects in its environment using collections of coupled processes responsible for sensorimotor interaction. These interaction processes run concurrently in order to ensure responsiveness to the environment, while coordinating sensorimotor expectations, action planning, and language use. The model has been implemented on a robot that manipulates objects on a tabletop in response to verbal input. The implementation responds to verbal requests such as “Group the green block and the red apple, ” while adapting in real-time to unexpected physical collisions and taking opportunistic advantage of any new information it may receive through perceptual and linguistic channels.

Open-ended language communication remains an enormous challenge for autonomous robots. This paper argues that the notion of a language strategy is the appropriate vehicle for addressing this challenge. A language strategy packages all the procedures that are necessary for playing a language game. We present a specific example of a language strategy for playing an Action Game in which one robot asks another robot to take on a body posture (such as stand or sit), and show how it effectively allows a population of agents to self-organise a perceptually grounded ontology and a lexicon from scratch, without any human intervention. Next, we show how a new language strategy can arise by

Although most previous model-based research has not moved beyond first-order semantics, human languages are clearly capable of expressing second-order semantics: the meanings expressed in a sentence do not only consist of conjunctions of first-order predicates but also predicates that take other predicates as an argument. In this paper we report on multi-agent language game experiments in which agents handle second-order semantics. We focus our discussion on how this type of research is able to provide fundamental insights in how properties of humanlanguage-like properties could once have emerged. For recursion, this might have happened as a side-effect of agents trying to reuse previously learned language structure as much as possible. 1.

Abstract. Artificial agents trying to achieve communicative goals in situated interactions in the real-world need powerful computational systems for conceptualizing their environment. In order to provide embodied artificial systems with rich semantics reminiscent of human language complexity, agents need ways of both conceptualizing complex compositional semantic structure and actively reconstructing semantic structure, due to uncertainty and ambiguity in transmission. Furthermore, the systems must be open-ended and adaptive and allow agents to adjust their semantic inventories in order to reach their goals. This paper presents recent progress in modeling open-ended, grounded semantics through a unified software system that addresses these problems. 1