Example of machine learning class library implemented in front-end Javascript

The entire system of the Web has developed significantly in recent years, although the use cases of Javascript and node.js are still far from comparable to Java/Python. But it is enough to be applied to many machine learning environments. And the biggest advantage is - one browser can do it all for you!

This article mainly introduces to you the machine learning class library implemented by front-end Javascript. The editor thinks it is quite good. Now I will share it with you and give you a reference. I hope it can help you.

Although machine learning libraries based on Javascript are still very early and many are still under development, they can indeed provide a relatively early experience. In today's article, we will select several cool machine learning and AI-related web applications to let everyone experience it for the first time~~

Brain

brain is a A library that allows you to quickly and easily create neural networks and train them based on input/output. Although a CDN browser version can directly load the class library into the web page, because this training process takes up a lot of resources, the class library is run in a Node.js environment. This class library contains a very mini online demonstration that can be used to train the contrast recognition of colors

Deep playground

This educational The web application allows you to play with neural networks and explore its different components. Having a very well-designed UI allows you to control the input data, number of neurons, algorithm used, etc. Various related metrics will affect the final result. Of course there's a lot to learn behind the scenes, the code is open source, uses a custom machine learning language (typescript), and has very good documentation

FlappyLearning

This is a machine learning Javasript class library implemented with 800 lines of code, which implements a demo of the machine learning flappy bird game. AI techniques are used in this class library: Neuroevolution, which applies neural system algorithms from the journal "Nature" to dynamically learn from the success and failure of each iteration. The demo is very simple to run, just use the browser to open index.html

You can see in the screenshot that after 20 generations of learning, this bird, in the screenshot I took time, and still haven’t died yet!

Synaptic

Probably one of the most actively maintained projects, Synaptic is a node.js and browser library. This library is designed to be architecture-agnostic. Allows developers to create any type of neural network. With little built-in architecture, it enables fast testing and algorithm comparison. It also contains a very complete description of neural networks, some practical demonstrations, and many other related tutorials to introduce how machine learning works

Land Lines

land lines is a very interesting Chrome Web experiment that looks for satellite images of the Earth to find similar user graffiti. This app has no server calls, runs completely in the browser, uses webGL and machine learning, and has a good experience on the mobile side.

ConvNetJS

Although no longer actively maintained, ConvNetJS is the most advanced library for Javascript machine learning. It was first developed at Stanford University and later became very famous on Github, with many community-developed features and tutorials. Runs directly in the browser, supports multiple learning techniques, is low-level, and is very suitable for larger experiences in neural networks

Thing Translator

This is a web experiment demonstration that allows you to use your mobile phone to identify real-life items and name them in different languages. This app is implemented using web technology and two machine learning APIs from Google, including:

Cloud Vision (image recognition) and Translate API (language translation)

Neurojs

An AI system framework based on "reinforcement learning". It's a pity that this project does not have proper documentation, but there is a demonstration of self-driving with a written description of many relevant details. This class library is pure Javascript and can be compiled and packaged using webpack or babel

Machine_learning

Another Javascript library that allows us to set up/train neural networks. It is very simple to install using node.js and the client, has a very clean API, and is very adaptable to developers of different skill levels. This library contains a large number of demos, including many popular algorithms, to help you understand the core and language learning principles

DeepForge

This is a user-friendly deep learning development environment that allows you to design neural networks using a simple graphical interface. It supports training models on remote machines and has built-in version control. This project is based on Node.js and MongoDB and runs in the browser. , the installation process is very similar to most web development processes

Related recommendations:

Top Ten Machine Learning Needs to Know Algorithm

How PHP implements the Naive Bayes algorithm for machine learning

PHP Machine Learning Library Example tutorial of php-ml

The above is the detailed content of Example of machine learning class library implemented in front-end Javascript. For more information, please follow other related articles on the PHP Chinese website!