Agentic Ai and Agentic RAG

## Details

Join us for an online event dedicated to exploring principles, tools and applications in the domain of Agentic AI and Agentic RAG (Retrieval Augmented Generation). Whether you're a beginner looking to dive into Agentic AI, this event will have something for everyone. We will discuss the role of Agentic AI and Agentic RAG in today's world and the importance of various applications. Generative AI uses LLMs to generate text, images, videos along with other contents based on prompts. Agentic AI takes actions making decisions, interacting with the environment, utilizing LLMs in order to reach certain objectives. Agentic AI operates with a level of autonomy and intelligence to add new dimensions. Agentic RAG (Retrieval-Augmented Generation) is a framework which augments capabilities of traditional RAG by using AI agents to control and manage retrieval and generation tasks in a dynamic way. AI agents in Agentic RAG make intelligent decisions regarding retrievals from data sources along with decisions about correct responses to generate accurate results.

Google Meet Link: https://meet.google.com/qho-avbm-qgz
(Time in Pacific Daylight Time ‎(UTC-7)‎, California (USA) time)
12:00 PM - 12:05 PM Introduction
12:05 PM - 12: 50 PM Session 1

Title: Agentic RAG, Agentic reasoning and Formalisms for Autonomous AI agents

Speaker: Dr. Shyam S Sarkar, Big Data Science Meetup Organizer

Abstract: Retrieval-Augmented Generation (RAG) evolved as a solution for enhancing LLMs by integrating real-time data retrieval to provide contextually relevant up-to-date responses. But RAG systems are
constrained by static workflows lacking in the adaptability for multi-step reasoning and complex task management. The solutions with Agentic Retrieval-Augmented Generation (Agentic RAG) eliminates
these limitations by embedding autonomous AI agents into the RAG pipeline. Agentic RAG systems are capable of providing agentic designs for reflection, multi-agent collaborations, planning with adaptive collaborations and complex workflows. In this talk there will be discussions about different architectures based on agent collaborations in key applications such as healthcare, finance etc. This presentation also includes topics on AI agents automating the web scraping process which allows for more flexible and adaptive data extraction from the web compared to traditional scraping methods. Formalisms for Agentic AI can provide structured and mathematical ways to represent and
reason about the creation, behavior and interactions of autonomous AI agents. Understanding, designing, verifying, and controlling agentic systems will need formalisms to provide a rigorous and well-defined foundation for these increasingly sophisticated systems.

12:50 PM - 12:58 PM Q/A