Zilliz - Overview of Generative models in ML

Learn the following topics:

What is Generative AI?
Use cases of generative AI
Large Language Models, Neural Networks, and Parameters
GAN (Generative Adversarial Network)
Diffusion models
Multimodal models

1. Overview of Generative models in ML
-Aishit Dharwal

3. Generative AI

4. Generative AI
A branch of AI that focuses on creating new data.
It utilizes machine learning algorithms to learn the
underlying patterns and relationships within a
dataset and then uses that knowledge to generate
entirely new content that resembles the training
data.

5. Generative AI
A branch of AI that focuses on creating new data.
It utilizes machine learning algorithms to learn the
underlying patterns and relationships within a
dataset and then uses that knowledge to generate
entirely new content that resembles the training
data.

6. ●Image Generation: Creating photorealistic
images, editing existing photos, and
generating artistic variations.

7. ●Image Generation: Creating photorealistic
images, editing existing photos, and
generating artistic variations.
●Text Generation: Writing different creative
text formats like poems, scripts, or even code
in a coherent and stylistic manner.

8. ●Image Generation: Creating photorealistic
images, editing existing photos, and
generating artistic variations.
●Text Generation: Writing different creative
text formats like poems, scripts, or even code
in a coherent and stylistic manner.
●Video Generation: Producing realistic videos
from scratch or editing existing videos by
adding or removing elements.

9. ●Image Generation: Creating photorealistic
images, editing existing photos, and
generating artistic variations.
●Text Generation: Writing different creative
text formats like poems, scripts, or even code
in a coherent and stylistic manner.
●Video Generation: Producing realistic videos
from scratch or editing existing videos by
adding or removing elements.
●Music Generation: Composing novel musical
pieces in various styles.

10. Oreilly Media

11. LLM

12. LLM
A large language model (LLM) is a deep learning algorithm
that can perform a variety of natural language processing
(NLP) tasks. Large language models use transformer
models and are trained using massive datasets — hence,
large. This enables them to recognize, translate, predict, or
generate text or other content.

13. LLM
A large language model (LLM) is a deep learning algorithm
that can perform a variety of natural language processing
(NLP) tasks. Large language models use transformer
models and are trained using massive datasets — hence,
large. This enables them to recognize, translate, predict, or
generate text or other content.
GPT 4 is one of the most versatile LLMs. It is multimodal - vision and text - can
write code as well

14. https://www.linkedin.com/pulse/demystifying-llms-quick-summary-andrej-karpat
hys-intro-lye-jia-jun-r70jc/
Next word prediction task
forces the model to learn a
lot about the world.

15. https://www.marktechpost.com/2023/10/15/this-artificial-intelligence-survey-research-provides-a-comprehensive-
overview-of-large-language-models-applied-to-the-healthcare-domain/

16. GAN - Generative Adversarial Network

17. GAN - Generative Adversarial Network
GANs can generate high-quality synthetic data, making
them valuable for tasks like image generation, data
augmentation, and more.

18. GAN - Generative Adversarial Network
GANs can generate high-quality synthetic data, making
them valuable for tasks like image generation, data
augmentation, and more.
In a GAN, two neural networks contest with each other in
the form of a zero-sum game, where one agent's gain is
another agent's loss.

19. https://towardsdatascience.com/decoding-the-basic-math-in-gan-simpl
ified-version-6fb6b079793

20. https://developers.google.com/machine-learning/gan/gan_structure

21. https://developers.google.com/machine-learning/gan/gan_structure

22. Diffusion Models

23. Diffusion Models
Diffusion models are generative models that can create new data samples by starting
with basic data and gradually transforming it into more complex and realistic data.
Used for tasks like image generation, data denoising, and probabilistic data
generation.

24. Diffusion Models
Diffusion models are generative models that can create new data samples by starting
with basic data and gradually transforming it into more complex and realistic data.
Used for tasks like image generation, data denoising, and probabilistic data
generation.
Inspired by non-equilibrium thermodynamics: Add randomness and reverse it
successfully!
Add noise to original image and learn how to reverse the process.

25. Diffusion Models
Diffusion models are generative models that can create new data samples by starting
with basic data and gradually transforming it into more complex and realistic data.
Used for tasks like image generation, data denoising, and probabilistic data
generation.
Inspired by non-equilibrium thermodynamics: Add randomness and reverse it
successfully!
Add noise to original image and learn how to reverse the process.
1.Stable Diffusion - uses latent diffusion model
2.DALL-E

26. https://en.wikipedia.org/wiki/Diffusion_model

27. Multimodal Models

28. Multimodal Models
A Multimodal Generative model draws outputs from a
combination of multiple data types to provide
responses as insights, content, and more.
Capable of processing information from different
modalities, including images, videos, and text.

29. Assembly AI

30. Assembly AI

31. Assembly AI

32. Feel free to send in your
questions in the
comments/DM on Linkedin