Genomics at the Nexus of AI, Computer Vision, and Machine Learning

22PacMan-RL: A Game-Changing Approach to Drug Development Through Reinforcement Learning

Abhishek Goud Amkamgari^1*, Harshita Sharma², Rashmi Verma³ and Bhawna Kaliraman³

¹Department of Computer Science and Engineering, Neil Gogte Institute of Technology, Hyderabad, India

² Department of Computer Science and Engineering, DPG Institute of Technology and Management, Gurugram, India

³Department of AIML/DS, DPG Institute of Technology and Management, Gurugram, India

Abstract

Drug development, an intricate labyrinth of experimentation, often resembles a high-stakes gamble. This study presents PacMan-RL, a reinforcement learning method inspired by the game Pac-Man, for drug development. PacMan-RL navigates the chemical universe to find target proteins while avoiding undesired traits like toxicity. This dynamic approach promises more effective and safer medication candidates. This document provides the conceptual framework of PacMan-RL, including the definition of the agent, the action space, the state representation, and the reward mechanisms. This potent technique can transform medication design, resulting in expedited and more effective creation of efficacious drugs without adverse effects.

Keywords: PacMan-RL, drug development, reinforcement learning, game-changing approach, Pac-Man development

22.1 Introduction

The process of discovering new drugs is time-consuming and costly, frequently hindered by the need for experimentation and trial and a lack of extensive study of the ...