Mentat: A Data-Driven Agent-Based Simulation of Social Values Evolution Samer Hassan Luis Antunes Juan Pav ó n Universidad Complutense de Madrid University of Surrey Universidade de Lisboa
Samer Hassan MABS Objectives of the Mentat ABM Case Study of Data-Driven ABM approach Study the evolution of the Spanish society in the period Framework for the application of different AI techniques
Samer Hassan MABS Contents Methodological approach The Sociological Problem Mentat: Architecture Mentat: Social Dynamics Mentat: Results Future work
Samer Hassan MABS Heading towards Data-Driven ABM Learning from Microsimulation: Minimizing random initialisation Feeding the simulation with representative survey samples Explicit rules can be problematic Empirical probability equations to determine changes in the micro behaviour Injecting more data into ABM From other sources (e.g. qualitative) In other stages (e.g. design)
Samer Hassan MABS Classical Logic of Simulation
Samer Hassan MABS Proposal for Data-Driven ABM
Samer Hassan MABS Methodological aspects for Data-driven ABM Microsimulation concepts Initialisation with survey data Empirically grounded probability equations Design fed with data Qualitative info, equations Life cycle, micro-processes Validation with different empirical data ‘Deepening KISS’ for exploring the model space
Samer Hassan MABS Contents Methodological approach The Sociological Problem Mentat: Architecture Mentat: Social Dynamics Mentat: Results Future work
Samer Hassan MABS The Problem Aim: simulate the process of change in social values in a period in a society Plenty of factors involved To which extent the demographic dynamics can explain the mental change? Inertia of generational change
Samer Hassan MABS The Problem Input Data loaded: EVS-1980 Quantitative periodical info Representative sample of Spain Allows Empirical Validation Intra-generational: Agent characteristics remain constant Macro aggregation evolves
Samer Hassan MABS Contents Methodological approach The Sociological Problem Mentat: Architecture Mentat: Social Dynamics Mentat: Results Future work
Samer Hassan MABS Mentat: architecture Agent: Mental State attributes Life cycle patterns Demographic micro-evolution: Couples Reproduction Inheritance
Samer Hassan MABS Mentat: architecture World: 3000 agents Grid 100x100 Demographic model 8 indep. parameters Social Network: Communication with Moore Neighbourhood Friends network Family network
Samer Hassan MABS Contents Methodological approach The Sociological Problem Mentat: Architecture Mentat: Social Dynamics Mentat: Results Future work
Samer Hassan MABS Understanding Friendship Dynamics “Meeting” & “Mating”: strangers => acquaintances => friends => partner “Meeting”: depends on opportunities alone space & time “Mating”: depends on both opportunities & attraction Proximity principle: ‘the more similar two individuals are, the stronger their chances of becoming friends’ Features channel individual preferences Homogeneous friendship choices
Samer Hassan MABS Mentat: Social Dynamics Meeting Agents randomly distributed in space Mating Similarity operator => Friendship Matchmaking Couple chosen among “ candidates ” Quantity? The more friends, the more couples Quality? Couples should be similar
Samer Hassan MABS Be Fuzzy, my Friend Similar, Friend: fuzzy concepts Fuzzification Improves accuracy of similarity Improves realism of friendship Improves quality of couples But friendship develops through time: Dynamic evolution! Hypothesis: Logistic function
Samer Hassan MABS Fuzzy Friendship Evolution
Samer Hassan MABS Contents Methodological approach The Sociological Problem Mentat: Architecture Mentat: Social Dynamics Mentat: Results Future work
Samer Hassan MABS Results
Samer Hassan MABS Results It may arise new sociological assumptions: In the prediction of social trends in Spain, Demographic Dynamics probably have, attending to the results, a key importance
Samer Hassan MABS Contents Methodological approach The Sociological Problem Mentat: Architecture Mentat: Social Dynamics Mentat: Results Future work
Samer Hassan MABS Future Work Mentat as a stage-based modular framework Enabling/Disabling modules for exploration ceteris paribus Explore the application of other AI techniques: NLP: biography of a representative individual Complementary output in natural language Events tracing -> XML -> NL DM: clustering over the input and output Helpful in design and validation
Samer Hassan MABS Thanks for your attention! Samer Hassan University of Surrey Universidade de Lisboa Universidad Complutense de Madrid
Samer Hassan MABS Contents License This presentation is licensed under a Creative Commons Attribution You are free to copy, modify and distribute it as long as the original work and author are cited