Software AlphAI Neural Network Virtual Laboratory

Simple Network

Understand in Details

On simple learning scenarios, understand and teach step by step how artificial neural networks work.

Draw connections inside the network yourself and observe the consequences on the robot’s movements. Thanks to multiple displays (activity values, thickness of connections, colors for learning, etc.), explain visually the learning equationsthat automatically optimize these connections.

Complex networks

Scale Up

Activate the camera and the intermediate layers of neurons and see how the simple principles detailed above allow the learning of complex behaviors (obstacle avoidance, pushing a ball, etc.) after scaling up to thousands of neurons and hundreds of thousands of connections. You have in front of you all the calculations made, but they are so numerous that we still miss something, isn’t it fascinating?

Other algorithms

Decide like your neighours

Discover – still very visually – other Machine Learning algorithms. In particular, the K-nearest  neighbors algorithm, which is part of several high school curricula, is particularly intuitive and  adapted to introduce supervised learning. Then compare its results to those of neural networks, on the same training database

Beginner

Let us Guide You

Select the beginner display mode to see only the most common settings. Most importantly, load demo configurations to effortlessly run pre-set supervised or reinforcement learning scenarios.

Intermediate

Start your own learning process

Test your own learning by selecting the sensors used as inputs, the allowed actions, the reward mode for reinforcement learning. Finally, select and configure the AI algorithm, for example to test what the intermediate neurons serve to…

Expert

For Professionals

Use the AlphAI platform to debug your learning model, test the most advanced parameters, study the properties of individual neurons, etc.

"Classical" Programming

Learn to program with AlphAI

A dedicated Python library allows you to take control of the robot from the computer with simple instructionsto direct the motors, retrieve sensor data, and program paths, maze exit,  etc., without necessarily requiring AI. It is also possible to run Python code directly on the Raspberry Pi for faster reaction times.

Program the AI

A Modular approach

Choose the “student code algorithm option and code the one of your choice in the popping-up Python window: the AlphAI platform takes care of all the input-output machinery and you focus on the most interesting! Even easier than AI, program your own reward function to make AlphAI learn new behaviors. Our curricula provide examples exercises to practice basic programming skills while coding an AI.

Other systems

AI Everywhere? 

We are working on making the AlphAI software compatible with other robots and automated systems, so that they can be driven by machine learning. You will also be able to carry out your own projects to develop new materials to be controlled with AlphAI.

The AlphAI device offers a platform for teaching artificial intelligence in different ways and at different levels. Its approach is very concrete, based on the realization of mini learning scenarios.

A teaching with AlphAI will take these scenarios, or others, and run them according to the age of the audience, either in the form of a teaching by a presenter, or in the form of a workshop involving manipulation, where several groups will each manipulate a different robot.

Supervised learning, image recognition

 
Supervised learning, robot racing

Reinforcement learning in a mini-neural network

Reinforcement learning with camera and more complex network 

Reinforcement learning: Following a balloon