Generative AI and Agent AI

Generative AI Models

Generative AI in Applications

• Image and video generation.
• Natural language processing (NLP) is a key area, focusing on generating new documents and content.

Defining Natural Language Processing

• NLP involves generating human-like, meaningful, and contextually appropriate language.

Generative AI Models

• Generative Adversarial Networks (GANs): Two modules compete.
• Variational Autoencoders.

Examples Slides

• Examples include Atom 10 to warm transformer and GaN super resolution.

Key Concept Training Data

• Utilizing generative AI models to translate phrases across languages.

Deep Learning and Translation

• Deep learning, particularly transformer architectures, excels at tasks like text generation, summarization, translation, and question answering.

Agent AI

• Agent AI: Autonomous entities that can sense the environment, analyze data, and act.
• Agent AI is important in AI system designed to act autonomously.

Generative AI vs. Agent AI

• Generative AI focuses on content generation from learned patterns.
• Agent AI is autonomous, proactive, and goal-driven, often built on top of LLMs, adding capabilities like memory.

AI Models

• LLMs provide language understanding for agent AI.
• Agent AI uses LLMs to interpret, reason, and plan in natural language.

Future of AI

• Agent AI is a gateway towards AI solutions in complicated cases.

LLMs & Agent AI

• LLMs can be converted to agent AI by enabling planning and autonomous behavior.

Agentic AI

• Is an AI system that act as agents and are capable of making decisions, setting goals, interacting with the environment.
• Agents can plan, reason, and learn from experience (reinforcement learning).