Running an Artificial Neural Networks on an Arduino Uno -06/12/23

 In this lesson, we used our Arduino to run an Artificial Neural Network which uses Machine Learning Algorithms. This was the perfect project to learn about machine learning and the basics of artificial intelligence. The program was designed to interpret the segments displayed on an LED numerical display to solve the corresponding binary number.

An artificial neural network is a segment of code which learns how to respond to inputs based on examples sets of inputs and outputs. They are very powerful tools and use in autonomous vehicles and facial recognition system and pattern recognition. A good example of it is in handwriting recognition.


One of the key principles, in an artificial neural network is that the network needs to be trained. So, first the program starts with the untrained values. We need to run the serial monitor to see this values.


After that, the program runs in training cycle to solve the answer which is needed and the picture is attached below.


There is a problem with this which can be solved by XOR.


XOR problem is one of the logical operations when performed on binary inputs that yield output for different combinations of input, and for the same combinations of input no output is produced. The outputs produced by the XOR logic are not linearly separable in the hyperplane. 

From the above truth table, it can be say that XOR produces an output for different states of inputs and for the same states of inputs it does not produce any output. 

In order to add an additional layer of input for the program to run through the artificial neural network, hidden nodes were added as a variable in example code to solve the XOR problem.

Additionally, the program itself keeps running through training cycles until the error rate is reduced to the predetermined level. Every time, the AI is trained the number of training cycles will be vary.

This is important historically because, in early days of neural network the researchers were exploring the idea to solve the complex problems using artificial neural network but due to XOR problem they were not able to do it early. This was holding back the development of AI in early days. Solving XOR problem was not just about XOR itself, it demonstrated the power of neural network. It paved the way for neural networks to tackle real world problems like image recognition, natural language processing and more.

The XOR problem played a crucial role in shaping the field of neural networks, emphasizing the need for deeper architectures and non linear activation functions. Its historical importance lies in sparking advancements that led to the neural networks we use today.

Comments

Post a Comment

Popular posts from this blog

Main Project- MIDI Keyboard Light Displayer

Image
 For our final project, our group carefully picked an interesting topic which is to display the different colours from DMX light box using the MIDI keyboard with arduino. It's main purpose is to like capture the vibe while playing the keyboard. While playing the keyboard it would generate a different types of colour from DMX light box. Each keyboard button would generate a different colour like red, green, blue and so on.            We did our project in class only, as we were not able to do it at home and also we had a little time so, we were not able to make it perfect. Our original plan was to make like to add the sound in keyboard but due to little time we were not able to make it. However, we were able to complete the project at the last moment. And also, we were able to show our project in the STEM FAIR which was the program conducted by university. Many people saw our project and they were happy about it.  Overall, this project was little t...
Read more

Building a MIDI Instrument

Image
MIDI (Musical Instrument Digital Interface) is a standard used for transmitting musical data between electronic devices. It allows instruments, computers and other devices to communicate with each other and to control various aspects of music such as notes, pitch, velocity and more. Information on the standard set by the MMA (MIDI Manufactures Association) can be found on this link MIDI.org – Expanding, promoting, and protecting MIDI technology for the benefit of artists and musicians around the world. There are also other Standards like; 1. USB(United Serial Bus) which is a standard interface used for connecting peripherals to computers and other devices. It allows for plug and play functionality and supports data transfer. 2. JPEG(Joint Photographic Experts Group) is a standard image compression format commonly used for storing and transferring digital images. 3. TCP/ IP(Transmission Control Protocol/ Internet Protocol) is a set of communication protocols used for connecting devices ...
Read more

Accessibility Tasks

Image
Activity 1 - Real Users For the following pictures decide what existing or fantasy input/output devices would benefit them. Remember people have five senses sight, smell, taste, touch, and hearing, so you might want to create devices for each of them – there is also the possibility of direct brain input. Blind People - For blind people keyboard with braille letters, and the speaker when they type the keyboard, and the system which converts words to speech would be best for them. Old People - For old people, they would need something like the monitor to display the text at larger size, a good speaker and guidance on screen. Parents and Kids - For them, they would need modern devices like iPads, laptops, smart phones, gps to track the kids, etc. Business man - Smart phones, laptops, headphones, touch pad, etc. Foreign people - For foreign people, they would need keyboard that supports their language, translator devices or apps Activity 2 1)      What do you need to know o...
Read more