S.P.A.N Project ANN

צםרקרק.jpg
press to zoom
press to zoom
press to zoom
press to zoom
press to zoom
press to zoom
press to zoom
press to zoom
press to zoom
press to zoom
press to zoom
press to zoom

INTRODUCTION

MACHINE LEARNING And AI FOR HIGHER COMPLEXITY DESIGN METHODS

DYNAMIC OCCUPATIONS SIMULATIONS

This project presents a search method reflecting on theories of social behavior. This project's aim is to introduce and explore how the interactions of heterogeneous individuals impact the wider behavior of social/ spatial systems. In this project, we introduce agent-based modeling and its utility for studying hybrid and spatial systems and aim to create models capable of capturing and incorporating human behavior. The implementation of ABM opens new possibilities to simulate and analyze the processes derived from human expansion. This projects focus on interior spaces and office floors as the key generators for the agent's behavior. This prototype aims to simulate and develop 4 sub-models, where planners, developers, and architects may create their specific structures and functions of the agent’s occupation. In the more advanced phases of the project, we use machine learning and AI tools to get more complex agents and more “real” behaviors that are unique to humans. The models are highly flexible and can be applied to alternative and areas for further study. They can also be used in different scales and numbers.

PROJECT TITLE

S.P.A.N Project ANN

LOCATION

Barcelona- Tel Aviv

YEAR

2018

DESIGN

EXPERIMENTS TO DEVELOP MODELS FOR A RULED SURFACES, AND MINIMAL SURFACES PREDICTION AS A GEOMETRY DEFINERS

a numerical investigation and prediction of new structural bottom-up development Project ANN0.1 is based on neural network processes, similar to those inside the human brain. Neural networks have thousands of layers that receive information, process it, and then pass it on to a higher level. I succeeded in developing an algorithm that can be self-thought using unsupervised analysis. Moreover, the algorithm also uses machine-learning processes as a way to let the computer analyze its own mistakes and learn how to improve itself during the process. One of the key elements in the project is that all the different shapes are being made without any human involvement. The computer itself learns how to develop points into lines and surfaces that can later be 3D printed. This newly proposed algorithm coefficient is based on particle swarm optimization.

© 2020  All works copyright by StudioCHIMERA