Summary of Behavioural Cloning in Vizdoom, by Ryan Spick et al.
Behavioural Cloning in VizDoom
by Ryan Spick, Timothy Bradley, Ayush Raina, Pierluigi Vito Amadori, Guy Moss
First submitted to arxiv on: 8 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper presents methods for training autonomous agents to play “Doom 2” through Imitation Learning (IL), utilizing only pixel data as input. The researchers compare Reinforcement Learning (RL) to IL in terms of humanness, analyzing camera movement and trajectory data. They also explore behavioral cloning, examining individual models’ ability to learn varying behavioral traits, such as aggressive or passive play styles. The trained agents exhibit human-like behavior, outperforming traditional AIs. While RL approaches may achieve better performance, IL methods provide stronger human-like behavioral traits. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper shows how computers can be taught to play a popular video game called “Doom 2” by watching and imitating real players. The researchers want to make these computer programs behave more like humans, so they test different ways of doing this. They found that the imitation learning method works well, making the computer agents behave aggressively or passively just like human players do. This is a step towards creating more realistic video game characters. |
Keywords
* Artificial intelligence * Reinforcement learning
