Effects of Anticipation in Individually Motivated Behaviour on Survival and Control in a Multi-Agent Scenario with Resource Constraints

Guckelsberger, Christian; Polani, Daniel

doi:10.3390/e16063357

Cited by 8 publications

(6 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Far beyond a simplistic search for "better-performing" models, this again serves as a cognitive aid to build understanding and intuition about how aspects of different models function and behave in different situations. Guckelsberger and Polani (2014), for instance, compared the impact of different variants of empowerment, a popular computational IM (Salge, Glackin, & Polani, 2014), on the social interaction dynamics of a triad of AI agents under resource scarcity and limited observability. In their scenario, an agent's empowerment calculation and action selection requires the agent to anticipate long-run consequences of other agents' behaviour.…”

Section: Advancing Psychology Via Computational Simulationsmentioning

confidence: 99%

Advancing Self-Determination Theory With Computational Intrinsic Motivation: The Case of Competence

Lintunen,

Ady,

Guckelsberger

et al. 2024

Preprint

View full text Add to dashboard Cite

Computational models are powerful tools to formalise psychological theories, render them severely testable, and embed them into digital applications. Yet they have seen little uptake in Self-Determination Theory (SDT). Here we demonstrate how SDT's conception of competence and optimal challenge can benefit from computational modelling informed by AI research on computational intrinsic motivation (IM). We surface underspecification in present verbal definitions, challenging the construct validity of common operationalisations and impeding computational implementations. We identify four separate verbal facets of competence and match them to distinct computational IM formalisms. These leverage formal accounts of novelty, diversity, or progress to drive skill acquisition, yielding optimal challenge-seeking behaviour and other competence dynamics postulated by SDT. We argue that these accounts specifically and computational IM more widely highlight inconsistencies within, and can concretise SDT's verbal articulations. We gently introduce motivation researchers to computational modelling and IM, complement model intuition with formal detail, and provide practical pointers for their use in psychological research and application.

show abstract

Section: Advancing Psychology Via Computational Simulationsmentioning

confidence: 99%

Advancing Self-Determination Theory With Computational Intrinsic Motivation: The Case of Competence

Lintunen,

Ady,

Guckelsberger

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Capdepuy et al [6,7] investigated the interactions between agents that individually attempt to maximize empowerment in a shared environment and found that interesting structures can emerge. Similarly, Guckelsberger and Polani [17] evolved agents for empowerment maximization in a resourcecentric environment which resulted in agents exhibiting biologically plausible behaviors, such as greed and parsimony. Lastly, Clements and Polani [9] introduce the idea of team empowerment.…”

Section: Related Workmentioning

confidence: 99%

Empowered Neural Cellular Automata

Grasso,

Bongard

2022

Preprint

View full text Add to dashboard Cite

Information-theoretic fitness functions are becoming increasingly popular to produce generally useful, task-independent behaviors. One such universal function, dubbed empowerment, measures the amount of control an agent exerts on its environment via its sensorimotor system. Specifically, empowerment attempts to maximize the mutual information between an agent's actions and its received sensor states at a later point in time. Traditionally, empowerment has been applied to a conventional sensorimotor apparatus, such as a robot. Here, we expand the approach to a distributed, multi-agent sensorimotor system embodied by a neural cellular automaton (NCA). We show that the addition of empowerment as a secondary objective in the evolution of NCA to perform the task of morphogenesis, growing and maintaining a pre-specified shape, results in higher fitness compared to evolving for morphogenesis alone. Results suggest there may be a synergistic relationship between morphogenesis and empowerment. That is, indirectly selecting for coordination between neighboring cells over the duration of development is beneficial to the developmental process itself. Such a finding may have applications in developmental biology by providing potential mechanisms of communication between cells during growth from a single cell to a multicellular, target morphology. Source code for the experiments in this paper can be found at: https://github.com/caitlingrasso/empowered-nca. CCS CONCEPTS• Computing methodologies → Artificial life.

show abstract

“…Furthermore, we do not need to know about the game's complete mechanics or parameters, but only about how the player's actions affect what they can perceive. Existing studies demonstrate that empowerment is not restricted to single-player games [16,18], and can deal with uncertainty and dynamic game worlds [42].…”

Section: Aggregating Quantities Into Experience Predictionmentioning

confidence: 99%

Predicting Player Experience without the Player.

Guckelsberger

Salge

Gow

et al. 2017

Proceedings of the Annual Symposium on Computer-Human Interaction in Play

Self Cite

View full text Add to dashboard Cite

A key challenge of procedural content generation (PCG) is to evoke a certain player experience (PX), when we have no direct control over the content which gives rise to that experience. We argue that neither the rigorous methods to assess PX in HCI, nor specialised methods in PCG are sufficient, because they rely on a human in the loop. We propose to address this shortcoming by means of computational models of intrinsic motivation and AI game-playing agents. We hypothesise that our approach could be used to automatically predict PX across games and content types without relying on a human player or designer. We conduct an exploratory study in level generation based on empowerment, a specific model of intrinsic motivation. Based on a thematic analysis, we find that empowerment can be used to create levels with qualitatively different PX. We relate the identified experiences to established theories of PX in HCI and game design, and discuss next steps.

show abstract

Effects of Anticipation in Individually Motivated Behaviour on Survival and Control in a Multi-Agent Scenario with Resource Constraints

Cited by 8 publications

References 42 publications

Advancing Self-Determination Theory With Computational Intrinsic Motivation: The Case of Competence

Advancing Self-Determination Theory With Computational Intrinsic Motivation: The Case of Competence

Empowered Neural Cellular Automata

Predicting Player Experience without the Player.

Contact Info

Product

Resources

About