[SOLD OUT] Mastering the New World of ChatGPT, GPT-4, DALL-E 2, and Multi-Modal Models

Since ChatGPT, everybody became aware of the power of large AI models, yet not many people have an overview of this field, and how to get started.

This workshop provides a closer look at the state of the art of large AI models that you can use.

This is a rapidly evolving field and the workshop will aim at talking about the available state of the art.

Some of the items covered will be:

Large-Language Models:
– What are Large Language Models and why are they new?
– Overview of the brief history of Large Language Models
– Differences between OpenAI’s models: Base-model, Instruct Models, Chat Model
– Other relevant Large Language Models
– Large Language models you can run on your own infrastructure
– Use cases such as Text generation, Summarization, Entity extraction, Classification, Correction, Translation, Style-transfer (rewriting), Code generation, Semantic search, Clustering of information
– Prompt-engineering and Meta-prompts
– Working with program code, e.g. Codex and GitHub CoPilot
– Using OpenAI’s APIs directly and in Azure
– Integrating with Large Language Models
– Build your own Chat Bot

Image-generation:
– What are the principles behind image generation models?
– What services exist
– OpenAI’s DALL-E 2
– Image generation, In-Painting, Out-painting
– Prompt engineering for image generation
– Using APIs to control DALL-E 2
– A closer look at Stable Diffusion, the advantages of direct model access
– Fine-tuning and adapting Image Generation to specific requirements (e.g. generate images that look like you)
– Exploring advanced concepts such as Latent Space, and changing other advanced parameters
– Generating animations/movies

Multi-Modal models:
– Overview, e.g. Florence, GPT-4
– What are the known capabilities?
– How are Multi-Modal models used?

– Azure:
– What pre-built large AI services exist?
– How do those Services relate to specialized Azure Services, e.g. Translation or Summarization, Computer Vision, Custom Vision, Form Recognition, Face Recognition, and more

Large AI Models Everywhere
– in Bing
– in M365
– in Power Platform

More advanced integration
– Integrate your data-sources with large AI models

Discussion – Impact and Responsible AI:
– Where are we headed?
– What risks exists?
– How can risks me mitigated?

This is not a complete list and it will be updated as new capabilities appear in the market.

This workshop will start introductory but cover more advanced topics.

No deep-learning experience required.

Note that for many topics, the cloud providers and vendors require “Responsible AI” approval processes, those parts will be mostly demonstrated by the presenter and probably cannot be tried out live by the audience.

Share this on...