Simulationist Storytelling and Personality in AI

This set of vocabulary flashcards covers key concepts from simulationist storytelling lectures, focusing on the Versu and Facade systems, personality modeling through utility functions, and AI evaluation metrics.

Versu

A simulationist storytelling system developed by Evans and Short designed for text-only social drama featuring flexible AI characters and player roles.

Social Practice Objects

Objects that provide agents with possible actions and the constraints that must be satisfied to undertake those actions within the Versu system.

Affordances

A term from human-computer interaction literature used in Versu to describe the available options or possible actions an agent can take.

Preconditions

Requirements that must be satisfied for an operation to be available to an agent, similar to the operators in the STRIPS system.

Utility Function

The model used in Versu to represent individual personality by defining the value or utility an agent places on specific actions or social outcomes.

Richard Evans

One of the authors of the paper on Versu and a former author of the popular life simulation game, The Sims.

Regency Period

The historical setting for Versu, modeled after the life dramas of wealthy English people around the year $1812$.

Facade

An interactive drama system used as a primary comparator for Versu, featuring full interactive graphics and a model of joint behaviors.

Beats

Theatrical scripting units used in Facade to coordinate sets of actions and maintain continuous story state values.

Local Agency

An evaluation metric where a system is considered successful if player actions cause an immediate, context-specific, and meaningful reaction.

Global Agency

An evaluation metric referring to the user's ability to control the overall shape of a story and lead it toward one of various specific endings.

Physical Symbol System Hypothesis

A theoretical framework suggesting that personality is a manifestation of values that shape an agent's search for satisfying operator sequences.

STRIPS system

An AI planning system exemplified by Shaky the robot where move operators add new location facts and remove facts about the room being left.

Back-channel feedback

Non-verbal communication cues such as nodding and frowning used by AI agents like Auto Tutor to represent part of their discourse.

Authoring Effort

A criterion used by Evans and Short to evaluate their framework, claiming it allows for more efficient creation of interactive experiences compared to others.