Artificial Intelligence in Construction

Translator | Cui Hao

Artificial intelligence (AI) is expected to completely change us The world , although there is no such dramatic thing as robots with sentience taking over society occur. But, the computer algorithmcan Perform complex tasks effectivelyThere is no doubt that this is intelligence.

For the field of construction, in many aspects of , artificial intelligence will serve as the perfect tool to build this field. considers artistic elements when designing buildings and outdoor spaces Design,At this time, a lot of is needed Intelligent operation.

If you break down the architectural design into numbers, Then the computercomputer can participate. The geometric shape of the space, reasonableutilizationbuilding materials , The quantity of building materials used, wind direction, load-bearing, and even foot traffic are all Can we use the artificial intelligencematurefields. Of course,

artificial intelligenceas a design program In part, computers have done some of this work in preliminary ways. However, the promise of artificial intelligence is that computers will figure out solutions to #design problems in ##ThereforerequireslimitedmanpowerIntervention.

At the time of writing, artificial intelligence in construction is moving forward through research and experimentation. Also take into account the complexity and expense of many construction projects. Despite this,

its development is still Excited ,The research results packageIncludes many commercial products from startups and major technology companies. Generative Design

Machine Learning (ML) is

The Construction Industry is a major focus of AIapplication. However, ML is not the same as AI, but a subset of a more important discipline. Using Machine Learning (ML

), is to create and train the algorithm, and then through continuous modeling and testing, after Keep tryingto incremental improvements method to complete the task. Thisa processis similar to human learning by rote learning, just machine learning (ML) Can be done faster. Machine Learning (ML)

One of the frameworks is GAN or Generative Adversarial Network, which is a method that uses two Unsupervised learning of ​​artificial neural network​​ #model.

The basic idea of ​​​​this model is , a network called a generator is created according to the requirements Content- For example, an image of a human face. Then, anotherneural network called the discriminatorjudgesThe output of the generatoris consistent with the created content (face image)whetherConsistent.

Two neural networksarea kind of competition Relationship, the generator is trying to trick identification, while the identifier defends to avoid Misled by the generator.

In the example, the final generator would be Create increasingly realistic

## cannot be distinguished and ultimately fool the discriminator.

## Apartment building data set | Image source: Stanislas Chayo ##Stanislas Chaillou is one of many people trying to use GANs to create architectural designs. #For his master’s thesis work at Harvard, Chaillou used GANs to create and improve floor plan designs. At work, he discovered that space and function not only affect but also affect the ## design will also affect the style of the floor plan generated by

to

. "It seems that style irreversibly permeates the essence of the generative processhit …There will be no agnostic AI for architecture,” Chaillou said in an article on Towards Data Science

arrive############. "Instead, each model or algorithm will have its style, personality and know-how."#########

It is an attractive concept, for example, you can have some models that like modern architectural design styles, such as Frank Gehry and Zaha Hadid ), a school influenced by the Bauhaus（Bauhaus）school influence, another fusion of multiple styles.

Chaillou co-founded ​​Rayon​​, this is a collaboration software company specializing in graphic space design.

GAN is just a style of generative design where the idea is to slowly improve a computer-generated design over time. You can use an unsupervised learning model (such as a GAN) or combined with CollaborationWay.

Another example of generative design is ​​Delve ​​, it is a real estate development tool from Google’s Sidewalk Labs. Delve can generate hundreds of designs in minutes.

#Each design considers a variety of retail, residential, parking and public space requirements. These designs also have a detailed cost model to estimate the cost of a given design.

Web Applications

The beauty of the state of the technology is that it doesn’t necessarily need to be on-site The computing power to complete some advanced work. Importantly, web applications are becoming more sophisticated on the front end, while blurring the distinction between local software and cloud software boundaries.

Some web apps are looking to revolutionize the construction industry. Delve is such an exampleFinch 3D The same is true. Finch is used in the early stages of a projectFinch’s AI capabilities can generate multiple options based on design requirements.

Another company developed based on artificial intelligenceHigharc is designed to be created through an iterative process 3D models and plansto complete the of the automated house #design. Higharc founder and CEO Marc Minor told the Financial Times in 2020: “While a buyer or builder sees a simple 3D models, but behind the scenes there are complex algorithms constantly determining critical details that often require hours of manual work.”

Acquired by Autodesk in late 2020 AI architecture startup Spacemaker. This cloud-based software helps teams analyze and design real estate websites using design tools like Delve.

#Spacemaker can analyze up to 100 city blocks for criteria: zoning, views, daylight, noise, wind, Roads, traffic, heat islands, parking, etc.”

Enter Omniverse​

Whether it is local or in the cloud, AI projects are It requires a lot of computer power, and the graphics card

can provide just that. ##​

​

NVIDIA’s Omniverse

​​ The collaboration and simulation platform is a "scalable, multi-GPU real-time reference development platform for 3D simulation and design collaboration." Omniverse can accommodate a small or large number of users and single or multiple GPUs. Omniverse is based on the capabilities of NVIDIA’s RTX GPUs and Pixar’s universal scene description software.

Omniverse is more than just an architecture tool. Instead, it is used to generate realistic 3D designs and simulations for fields such as architecture, engineering, animation, or industrial design.

Architects can integrate Omniverse with tools like Autodesk 3ds Max, Rhino and Trimble Sketchup to transform designs into 3D simulations.

Currently, many AI jobs in Omniverse require products to be input into Omniverse

. Generally speaking, these products All generated and designed within the workflow. However, the future promises some exciting developments. For example, NVIDIA’s Toronto-based AI research lab created a tool called ​​GANverse 3D​​Tool that takes a 2D image and converts it into a 3D model. It’s easy to see how AI toolscanchangethe construction industryOr any industry that focuses on 3D modeling .

Computer Vision​

Many art projects alsowill be artificial The application of intelligence in buildings is influenced by , such as London's AI Build's Deadalus Pavilion.

##Credit：AI Build

This small structure is designed using NVIDIA GPUCome out, uses computer vision and deep learning to improve the capabilities of 3D# models created by industrial robots, including The ability to print structural elements with speed and accuracy.

Artificial Intelligence and the Future of Architecture

Artificial Intelligence isThe future of many industriesbrings hope, and construction is no exception. Although the construction industry is still in the stage of applying artificial intelligence #Early stages, but artificial intelligence shows great potential in enhancing architectural design. However, although there areartificial intelligence-driven Tools

assist in completing the design work,

human beings still need to# on the design results. ##Improvements and approvals. While construction tools are getting more advanced, they won’t replace human architects anytime soon. Translator Introduction Cui Hao, 51CTO community editor, senior architect, has 18 years of experience in software development and architecture, 10 Years of distributed architecture experience. Formerly a technical expert at HP. He is willing to share and has written many popular technical articles with more than 600,000 reads. Author of "Principles and Practice of Distributed Architecture". Original title: Artificial Intelligence in Architecture, author: Kevin Vu

##Link: https://dzone.com/articles/artificial-intelligence-in-architecture

##

The above is the detailed content of Artificial Intelligence in Construction. For more information, please follow other related articles on the PHP Chinese website!