There is growing interest in “Generative AI” recently. With technologies like “Ghat GPT” making their mark, a revolution that could completely transform the work environment is becoming a reality. Even big tech companies like Apple, Microsoft, Google, and Meta are pushing forward with AI technology development. Just the mention of “AI” can cause stock prices to soar or plummet based on who is leading or lagging in this field.
Apple, which was once an icon of innovation, is a prime example. It continually innovated by building the Apple ecosystem with products like Mac, iPhone, iPad, and services like iCloud. However, recently it has been perceived as lagging behind in AI compared to Google or Microsoft. The company’s decision to abandon the Apple Car, which it had been planning and preparing for over a decade, led to a significant drop in its stock price.
With discussions around “Generative AI” popping up everywhere, there’s a natural growing interest in “AI.” While the term “AI” has become commonplace in everyday language, there’s still a lingering sense of not fully understanding what AI is or how it works.
Fortunately, Coursera has released a 20-minute video titled “Introduction to Generative AI” provided by Google, which helps shed light on these topics. Although it’s a short video, it covers a lot of ground and is quite accessible for beginners. It creates an effective atmosphere for delivering a vast amount of information about “Generative AI” in a brief period. This text is based on a free post on Coursera and serves as a review.
“Coursera: Introduction to Generative AI”
AI is currently used in many fields. “Chat GPT” is a prime example, an AI that can generate something based on the input text. However, this technology has evolved beyond simple text generation to being able to generate images, videos, and even audio. While there are still some shortcomings, it’s becoming more sophisticated over time.
The course will cover the following four topics:
- Define Generative AI
- Explain How Generative AI Works
- Describe Generative AI Model Types
- Describe Generative Applications
“What is Generative AI?”
Generative AI is a type of artificial intelligence technology that can produce various types of content, including text, imagery, audio, and synthetic data.
“Difference Between AI and Machine Learning”
AI is a concept that includes machine learning (ML), and deep learning is a subset of ML. Deep learning is a method within ML that uses artificial neural networks to process more complex patterns than traditional machine learning.
AI is a discipline, like how physics is a discipline of science. AI is a branch of computer science that deals with the creation of intelligent agents, and are systems that can reason, learn, and act autonomously.
AI has to do with the theory and methods to build machines that think and act like humans.
- AI
- ML
- Unsupervised Learning
- Supervised Learning
- Reinforcement Learning
- Deep Learning
- ML
Machine Learning is a program or system that trains a model from input data. The trained model can make useful predictions from new (never-before-seen) data drawn from the same one used to train the model. This means that Machine Learning gives the computer the ability to learn without explicit programming. So what do these Machine Learning models look like? Two of the most common classes of machine learning models are unsupervised and supervised ML models.
“Supervised ML Learning VS Unsupervised ML Learning”
- Unsupervised ML Models: Unsupervised problems are all about discovery, about looking at the raw data, and seeing if it naturally falls into groups.
- Supervised ML Models: The model learns from past examples to predict future values. They require labeled data.
“Deep Learning”
While machine learning is a broad field that encompasses many different techniques, deep learning is a type of machine learning that uses artificial neural networks, allowing them to process more complex patterns than machine learning. Artificial neural networks are inspired by the human brain.
Like your brain, they are made up of many interconnected nodes, or neurons, that can learn to perform tasks by processing data and making predictions. Deep learning models typically have many layers of neurons, which allows them to learn more complex patterns than traditional machine learning models. Neural networks can use both labeled and unlabeled data.
The labeled data helps the neural network to learn the basic concepts of the task, while the unlabeled data helps the neural network to generalize to new examples.
“Generative AI”
Gen AI is a subset of deep learning, which means it uses Artificial Neural Networks, can process both labeled and unlabeled data, using supervised, unsupervised, and semi-supervised methods. LLMs are also a subset of Deep Learning.
LLMs are also a subset of Deep Learning. Deep learning models (or machine learning models in general) can be divided into two types – generative and discriminative.
- Discriminative Model
- Generative Model
“Discriminative Model”
A discriminative model is a type of model that is used to classify or predict labels for data points. Discriminative models are typically trained on a dataset of labeled data points, and they learn the relationship between the features of the data points and the labels. Once a discriminative model is trained, it can be used to predict the label for new data points.
“Generative Model”
A generative model generates new data instances based on a learned probability distribution of existing data. Generative models generate new content.
To summarize: Generative models can generate new data instances and Discriminative models discriminate between different kinds of data instances.
“Generative AI: Configuration”
The Generative AI process can take training code, labeled data, and unlabeled data of all data types and build a “foundation model”. The foundation model can then generate new content. It can generate text, code, images, audio, video, and more.
In traditional programming, defining a “cat” was done through defining it in code, but in generative AI, users can create content directly.
“A New Definition of Generative AI”
GenAI is a type of Artificial Intelligence that creates new content based on what it has learned from existing content. The process of learning from existing content is called training and results in the creation of a statistical model. When given a prompt, GenAI uses this statistical model to predict what an expected response might be–and this generates new content.
The power of Generative AI comes from the use of Transformers. Transformers produced the 2018 revolution in Natural Language Processing. At a high-level, a Transformer model consists of an encoder and decoder. The encoder encodes the input sequence and passes it to the decoder, which learns how to decode the representations for a relevant task.
“Hallucinations”
Sometimes Transformers runs into issues though. In Transformers, Hallucinations are words or phrases that are generated by the model that are often nonsensical or grammatically incorrect.
Hallucinations can be caused by a number of factors, like when: The model is not trained on enough data, The model is trained on noisy or dirty data, The model is not given enough context, or The model is not given enough constraints.
“Prompt”
A prompt is a short piece of text that is given to the LLM as input, and it can be used to control the output of the model in a variety of ways. Prompt design is the process of creating a prompt that will generate the desired output from a large language model (LLM).
Generative AI depends a lot on the training data that you have fed into it. It analyzes the patterns and structures of the input data, and thus “learns.”
“Model Types”
- Text to Text
- Text to Image
- Text to Video, Text to 3D
- Text to Task
- Foundation Model
“Foundation Model”
A foundation model, which is a large AI model pre-trained on a vast quantity of data “designed to be adapted” (or fine-tuned) to a wide range of downstream tasks, such as sentiment analysis, image captioning, and object recognition. Foundation models have the potential to revolutionize many industries, including healthcare, finance, and customer service. They can even be used to detect fraud and provide personalized customer support.
“Vertex AI”
Vertex AI provides a Model Garden including Foundation Models. Language Foundation Models include the PaLM API for chat and text. Vision Foundation Models include stable diffusion, proven effective for generating high-quality images from text descriptions.
Vertex AI Studio provides various tools and resources for developers to create and deploy generative AI models. It includes pre-trained model libraries, model fine-tuning tools, production deployment tools, and a community forum for sharing ideas and collaboration.
“PaLM API”
The PaLM API allows testing and experimenting with Google’s large-scale language models and Gen AI tools. Developers can integrate MakerSuite with PaLM API for rapid prototyping, accessing APIs through a graphical user interface. The suite includes training, deployment, and monitoring tools for models.
“Gemini”
Gemini is a multimodal AI model that understands nuances in images, audio, and programming code—enabling complex tasks previously deemed impossible. With advanced architecture, Gemini is flexible, scalable, and suitable for various applications. The Model Garden is continually updated with new models.
This course is a short 20-minute lecture introducing what Generative AI is and even showcasing “Generative AI” that users can experience firsthand. It’s a great introductory video for understanding AI concepts and is still available for free on the Coursera site.
“Introduction to Generative AI”
- Topic: Computer Science, Artificial Intelligence
- Coursera: http://app.ac/GACROu503