Let's develop our own AI agent

Hemant Rakesh (~hemant56)


Description:

Introduction

The term "Artificial Intelligence" is often used to describe machines (or computers) that mimic "cognitive" functions that humans associate with the human mind, such as "learning" and "problem solving". In this workshop I will be demonstrating the concept of reinforcement learning, which is one of the subcategories of AI. And currently the hottest topics in today's AI generation.

You may have noticed that computers can now automatically learn to play ATARI games (from raw game pixels!), they are beating world champions at Go, simulated quadrupeds are learning to run and leap, and robots are learning how to perform complex manipulation tasks that defy explicit programming.

In my opinion the best opensource resource to understand and build AI is OpenAi's gym, which include many pre-built environments as well AI agents suitable for the environments. I will be talking about how one can build their own AI agent given the environment with the help of some "mathmagic" and OpenAI

Basic outline of the talk

  • Introduction to AI - [4-5 minutes]
  • What is OpenAI ? - [4-5 minutes]
  • Success of OpenAI - [8-10 minutes]
  • Key concepts - [2 minutes]
  • Building an AI agent - [10-12 minutes]

    • Behind the scenes math
    • Usinig OpenAI
    • Demos
  • Q/A - [2 minutes]

Who is this talk for?

  • Developers who want to know what AI is, and why it's growing steadily
  • The curious folk who want to know how to build one on their own

Key takeaways

  • A new perspective of AI development
  • Use of Opensource resources to revolutionize the field of Artificial Intelligence
  • Notes/Material to get started with your own AI agent

Prerequisites:

Preferred

  1. Fundamentals of Probability theory
  2. Fundamentals of Statistics
  3. Fundamentals of Differential Calculus

Recommended

  1. Basic Python scripting
  2. What is Artificial Intelligence ?

Content URLs:

A few samples are in the following link:

https://github.com/Hemantr05/Reinforcement-Learning

Speaker Info:

Hemant Rakesh a final year CSE student at NMIT, Bangalore; has keen interests in deep learning and is a reinforcement learning enthusiast. With prior experience in this field he loves to share his skills and knowledge to the community as he believes - " together we grow ". He also heads the machine learning club at nmit and has authored quite a few AI based blogs on Medium.

Hemant is also a research intern currently working on projects with Biomedical Engineering(10^-6 - 10^9) and Electronic System, Indian Institute of Science, Bangalore. He has a rich experience in computer vision, deep learning, reinforcement learning, neural computing and medical imaging and EEG based computation. Also, he has experience in building software tools in python for Image and neural analysis.

Id: 1374
Section: Data Science, Machine Learning and AI
Type: Talks
Target Audience: Advanced
Last Updated: