Generative AI
Generative AI For Beginners
Topic of the course
- Definition and Overview of Generative AI
- How Generative AI differs from traditional AI approaches
- Key AI subfields involved (Deep Learning, Neural Networks)
- Real-world applications of Generative AI (Text, Image, Music Generation)
- Definitions and Key Differences
- Role of Probabilistic Models
- Use Cases of Generative Models (GANs, VAEs)
- Use Cases of Discriminative Models (CNNs, SVMs)
- Practical Scenarios: When to use Generative vs. Discriminative AI
- Definitions and Key Differences
- Role of Probabilistic Models
- Use Cases of Generative Models (GANs, VAEs)
- Use Cases of Discriminative Models (CNNs, SVMs)
- Practical Scenarios: When to use Generative vs. Discriminative AI
- Introduction to GANs: What they are and how they work
- Components of GANs: Generator and Discriminator
- Training Process and Loss Functions
- Applications: Image Synthesis, Deepfakes, Data Augmentation
- Common GAN Architectures: DCGAN, StyleGAN
- Challenges with GANs: Mode Collapse, Training Instabilities
- Hands-on Demo: Training a simple GAN
- Introduction to VAEs and How They Differ from GANs
- Structure of a VAE: Encoder, Decoder, and Latent Space
- Understanding Variational Inference
- Applications of VAEs: Anomaly Detection, Image and Text Generation
- Hands-on Example: Implementing a Basic VAE
- Introduction to Prompt Engineering - Crafting Effective Prompts for AI Models - Examples of Different Prompt Structures - Using Prompts to Guide AI Output (Creative Tasks, Coding, Language Translation) - Challenges in Prompt Engineering (Ambiguity, Misinterpretation) - Hands-on Exercise: Writing Prompts for Specific Tasks
- Overview of OpenAI API: Capabilities and Use Cases
- Moderation:
- How the Moderation API Works
- Using AI to Detect Harmful or Inappropriate Content
- Chat Completion:
- Building Conversational Agents
- Best Practices for Using GPT for Conversations
- Fine-tuning:
- Process of Fine-tuning Pre-trained Models for Custom Tasks
- Fine-tuning Techniques and Common Mistakes
- Hands-on Demo: Fine-tuning GPT on a Custom Dataset
- Function Calling:
- What Function Calling is and Use Cases
- Integrating Function Calling into Business Applications
- Example: Using Function Calling to Retrieve Real-time Data
- What is LangChain? Overview and Key Features
- Building AI Applications using LangChain
- Understanding LLM (Large Language Models) Chains
- Integrating External Tools with LangChain (Search Engines, Databases)
- Hands-on Demo: Setting Up a Simple LangChain Project
- Use Cases: Building Chatbots, AI-Assisted Tools
- What is Vertex AI: Overview and Features
- Vertex AI vs. Other AI Platforms (Azure ML, AWS Sagemaker)
- Building and Deploying Models using Vertex AI
- Managed Pipelines and Workflow Automation in Vertex AI
- Integrating Vertex AI with Other Google Cloud Services
- High-level Introduction to Vertex AI Co-pilot
- Example Use Case: Deploying a Pre-trained Model with Vertex AI
- Introduction to AI Ethics: Why It Matters
- Ethical Concerns with Generative AI (Bias, Misinformation, Misuse)
- Addressing Bias in Generative Models
- Implementing Fairness and Transparency in AI Models
- Regulatory Considerations: GDPR, AI-specific Laws
- Real-world Examples of AI Misuse and Lessons Learned
- Use Cases of Generative AI in Various Industries:
- Healthcare: Medical Image Generation
- Marketing: Text and Ad Generation
- Finance: Fraud Detection, Synthetic Data Generation
- Entertainment: Content Creation (Movies, Games)
- Hands-on Exercise: Exploring a Use Case and Developing an AI Solution
- Objective: Craft Effective Prompts for Language Models
- Steps:
- Introduction to Task
- Writing Prompts for Specific Outputs (e.g., Product Descriptions, Text Summarization)
- Evaluating Output: Correctness, Coherence, Creativity
- Adjusting Prompts to Improve Results
- Insights and Conclusion: Learning to Fine-tune Prompts for Better AI Results