Welcome

Passionately curious about Data, Databases and Systems Complexity. Data is ubiquitous, the database universe is dichotomous (structured and unstructured), expanding and complex. Find my Database Research at SQLToolkit.co.uk . Microsoft Data Platform MVP

"The important thing is not to stop questioning. Curiosity has its own reason for existing" Einstein



Thursday 15 February 2024

Fundamentals of Generative AI

 Generative AI is a branch of artificial intelligence that focuses on creating new content or data from scratch, such as images, text, music, or code. Generative AI models learn from existing data and use it to generate novel and realistic outputs that are not part of the original data. Some of the applications of generative AI include:

Image synthesis: Generative AI can create realistic images of faces, landscapes, animals, or objects that do not exist in the real world. 

Text generation: Generative AI can produce natural language texts on various topics, such as stories, poems, essays, or code. 

Music composition: Generative AI can compose original music in different genres, styles, and moods. 

Data augmentation: Generative AI can enhance or expand existing data sets by creating new samples that are similar but not identical to the original ones. This can help improve the performance and robustness of machine learning models. For example, generative AI can create new images of handwritten digits or new sentences of natural language.

The main challenge of generative AI is to ensure that the generated outputs are both diverse and realistic, meaning that they cover a wide range of possibilities and resemble the real data. To achieve this, generative AI models often use two types of techniques:

Probabilistic models: These are models that learn the probability distribution of the data and sample from it to generate new outputs. For example, variational autoencoders (VAEs) are probabilistic models that encode the data into a latent space and decode it back into the original space, adding some noise in the process to create variations.

Adversarial models: These are models that consist of two components: a generator and a discriminator. The generator tries to create outputs that fool the discriminator, while the discriminator tries to distinguish between real and fake outputs. The two components compete with each other and improve over time. For example, generative adversarial networks (GANs) are adversarial models that use neural networks as the generator and the discriminator.

Generative AI is a fascinating and rapidly evolving field of artificial intelligence that has many potential benefits and applications for society. However, it also poses some ethical and social risks, such as misuse, deception, or bias. Therefore, it is important to develop and use generative AI models responsibly and transparently, with respect for human values and rights. 

To learn more about the Fundamental of Generative AI , Microsoft Learn has a great course

It also covers what is the Azure OpenAI service. This being a Microsoft's cloud solution for deploying, customizing, and hosting large language models. There is a brief overview of  CoPilot.



No comments:

Post a Comment

Note: only a member of this blog may post a comment.