AI Essentials: Beginner’s Guide to ML, DL, NN, Generative AI #AIlearning

Introduction

Artificial intelligence, the art of creating computer systems that mimic human intelligence, is fascinating. To truly grasp AI’s potential, it’s essential to understand the various techniques and concepts used in AI, such as machine learning, deep learning, neural networks, and Generative AI.

This blog helps you better understand these AI concepts: machine learning, deep learning, neural networks, and Generative AI.

Artificial Intelligence

AI describes computer systems that can apply reasoning to subjects that previously required human intelligence. AI plays a significant role in helping machines learn, improve their work, and become experts at particular tasks.

Machine Learning

ML is a subset of AI that enables systems to use algorithms to predict and classify given data in response to ever-changing data, like how you learn from experience.

ML algorithms have three general ways of learning from the data to get better results and provide better predictions.

♦ Supervised learning

♦ Unsupervised learning

♦ Reinforcement learning

Deep Learning

Deep Learning is a compelling type of machine learning inspired by the human brain’s operation of neural networks to process large amounts of data and make complex decisions.

🎉Neural Network

A neural network uses electronic circuitry inspired by how neurons communicate in the human brain.

In the brain, cells called neurons have a cell body at one end where the nucleus resides and a long axon leading to a set of branching terminals at the other end. Neurons communicate by receiving signals into the axon, altering those signals, and then transmitting them through the terminals to other neurons.

A human brain has about 100 billion neurons. Each neuron is connected to up to 10,000 other neurons.

In a neural network, a building block called a perceptron acts as the equivalent of a single neuron. A perceptron has an input layer, one or more hidden layers, and an output layer. A signal enters the input layer, and the hidden layers run algorithms on it. Then, the result is passed to the output layer.

The operation of a neural network is pure mathematics. The network isn’t “thinking”; it is calculating. But it’s using those calculations to create an output humans can interpret as an answer or a recommendation.

🎉Example scenario

Joe will ask the neural network. Should I order a Pizza?

This neural network has already been trained. It has learned many things about Joe’s past pizza-buying behavior, so it’s ready to suggest whether Joe will likely order one today. Today, the neural network is going to invoke three things it has observed:

1. Joe usually orders ahead.
2. Joe is trying to lose weight, so lately, he’s ordered pizza less often.
3. Joe’s email includes a coupon for a free pizza.

Now Joe asks, “Am I likely to order a pizza today?”

Ordering ahead?

One node in the hidden layer looks up Joe’s buying pattern and observes that Joe always orders ahead. So, the node assigns ordering ahead a value of 1. (For clarity, one (1) and zero (0) values are used.)

The system has learned something else about Joe’s ordering: Joe gives more weight to some factors than others. So, the system gives each factor a weight based on Joe’s past behavior. The more Joe tends to do something, the higher the weight.

Joe always orders ahead, so the system gives today’s order-ahead factor a weight of 3. Multiplying the initial value (1) by the weight (3) gives the node an output value of 3. So far, the total value of the ordering factor is:

1 x 3 = 3

Getting Salad instead?

The next node in the hidden layer registers that Joe has been ordering a salad about half the time lately. So, the node gives the salad option an initial value of 0.5.

The system also knows that Joe is not consistent about avoiding pizza in his diet, so it gives the option to order a salad with a weight of 0.5.

The initial value (0.5) multiplied by the weight (0.5) is 0.25.

0.5 x 0.5 = 0.25

Free coupon?

The third node in the hidden layer knows that Joe has a coupon for a free pizza. So, it gives the coupon a factor of 1. Finally, the system observes that Joe always uses free coupons as fast as they arrive. So, the system multiplies the coupon factor by a weight of 5:

1 x 5 = 5

Now that these algorithms are run, the system advances results from all three hidden layer nodes to the output node.

Weighting the answer

The hidden layer has one more step to perform: using an activation algorithm. This algorithm will “fire” only when the total weight from Joe’s preferences algorithms reaches a certain threshold.

Suppose the threshold for pizza is reached when the total weight is greater than or equal to 6. The final node adds all three-factor values, resulting in 8.25.

3 + 0.25 + 5 = 8.25

That’s greater than the threshold of 6, so…

…it’s pizza for supper tonight!

🎉Trial and error

A neural network learns by continuously adjusting itself through trial and error. Once it has ingested or learned a certain amount of data, it stores it in its “body of information,” called its corpus. To learn, the neural network constantly tests new data or the results of its calculations against its corpus.

If the network determines that the new data or results don’t match the already established patterns, it modifies those patterns for a better fit. Sometimes, the network rapidly tests hundreds or thousands of modifications and adjusts to improve a single match. Then, the network tests to determine if the match is improving. So, step by step, the machine learns.

🎉Deep Neural Network

Multiple groups of multilayer perceptrons are arranged differently, which extends machine learning in the deep learning ecosystem.

A perceptron requires more brainpower in the form of deep learning. Deep learning relies on multiple layers of nodes (even groups of perceptrons with multiple layers of nodes!) to finish the work reasonably.

A deep neural network (DNN) is an advanced AI system that uses many hidden layers whose algorithms pass the results of sophisticated calculations. DNN layers can be arranged in groups or elaborate blocks of groups for greater power. DNNs can be doubled in competing teams that judge and learn from each other’s mistakes without human intervention, creating powerful reinforcement learning.

Generative AI

Generative AI is a type of artificial intelligence that creates new, original content that people have never seen before.

Most AI systems are discriminative AI models, which predict and classify data. However, generative AI models are deep-learning AI systems that use algorithms to generate content based on a submitted prompt.

For example, a discriminative model could tell a bicycle from a truck, and a generative model could generate a new image that looks like a bicycle. So, generative AI’s distinction from other AI systems is its ability to produce new and considered creative content, such as images, videos, music, synthetic data, essays, answers to questions, and more.

🎉Image generating tool

Here’s a quick example of how easily generative AI can create new images. A person typed “mountains, water, sky” in the text prompt of the image-generating tool called Free Image Generator. The tool instantly created the following four images!

The overall generative AI process:

💡First, a person feeds the AI a large amount of data, including images, sounds, text, and numbers.

💡Then, the AI analyzes this data, looking for patterns and relationships between the different pieces of information. The neural network is trained on a dataset of examples of the type of output it is intended to generate, such as images or text. During training, the neural network learns to identify patterns and relationships in the input data and use them to generate new outputs similar to the examples it was trained on but not identical to.

💡Next, the AI uses what it has learned to create something new. The neural network generates new outputs by inputting a random seed value. The seed value serves as the starting point for the generation process. The neural network processes the seed value and generates a new output based on the patterns and relationships learned during training.

For example, if someone gave the AI a set of images of dogs, it might use its knowledge of different dog breeds to create an image of a new dog that doesn’t exist in real life.

Generative AI can also complete more complex tasks, like writing stories or composing music. In these cases, the AI analyzes patterns in language or music to create something entirely new.

Types of generative AI models

Let’s explore the three primary types of generative AI models:

📀Variational autoencoder (VAE)

Consider variational autoencoder (VAE) models as skilled artists who can look at a painting, quickly sketch a simplified version, and then recreate a new painting using only that simplified sketch as a reference. The artist captures the essential elements of the painting and then uses them to create a new work of art.

VAEs use a similar process. The “encoder” network compresses the input data into a lower-dimensional representation, and the “decoder” network reconstructs the original data from this compressed representation. This allows VAEs to capture the underlying structure and patterns in the data, which can then generate new, similar data.

📀Generative adversarial network (GAN)

Consider the generative adversarial network (GAN) model as a competition between a skilled forger (the generator) and a talented art critic (the discriminator). The forger creates fake paintings, while the critic tries to determine whether each painting is genuine or a forgery. As the forger improves their technique, the critic becomes more discerning, and this cycle continues until the forger can create near-perfect forgeries.

In GANs, the generator creates new data while the discriminator evaluates the quality of the generated data. The generator tries to create realistic data to fool the discriminator, while the discriminator learns to better distinguish between real and generated data. This competition leads to the generator creating increasingly realistic content.

📀Autoregressive

Imagine an autoregressive model as a skilled storyteller who listens to the beginning of a story and then continues it by predicting what comes next based on the words and events that have occurred so far. The storyteller uses their knowledge of language, grammar, and storytelling conventions to create a coherent and engaging continuation of the story.

Autoregressive models generate new content by predicting the next element in a sequence based on the previous elements. They are particularly well-suited for generating text because they can model the conditional probabilities of words and characters in a sentence.

Examples of generative AI systems and applications

🤖ChatGPT, an AI chatbot from OpenAI, follows instructions in the prompt and provides a detailed response.

🤖DALL-E from OpenAI, Midjourney, and Adobe Firefly can generate realistic digital images based on text prompts.

🤖Music Generation tools include Amper and MuseNet.

Uses of generative AI in industries

Generative AI is impacting today’s world across many industries, including sports, entertainment, healthcare, retail, banking, manufacturing, engineering, security, media, agriculture, and more, and it continues to expand.

Conclusion

As we conclude our exploration of the 🎯AI essentials, it’s important to remember that this is just the beginning of your journey🚀 into the fascinating world of artificial intelligence.

We’ve covered the core concepts of machine learning, deep learning, neural networks, and Generative AI, giving you a solid foundation to build. Armed with this knowledge📚, you’re now better equipped to understand the inner workings of the AI systems that are transforming industries and shaping the future🌎.