Game Playing Using AI on Apache Spark

Using AI to play games is often perceived as an early step towards achieving general machine intelligence, as the ability to reason and make decisions based on sensed information is an essential part of general intelligence. Games are good playgrounds for experimenting with intelligent agents as the goals and action rules are often well-defined and abstract. People have been interested in using AI to play games for quite a while. Recent development of deep neural networks allowed visual information in games to be processed effectively and directly used for decision making of agents. And the area of deep reinforcement learning and meta-learning are also being explored in this aspect.

In this presentation we will share experiences from our attempts in using AI on Spark for game playing. The talk will include demos and some details of the experiments, and our learnings, for example, whether Spark is a good fit for implementing game related AI, which parts needs to be improved, and the chances of Spark in the area of AI game playing.

 

Try Databricks
See More Spark + AI Summit in San Francisco 2019 Videos


« back
About Yuhao Yang

Yuhao Yang is a software engineer at Intel, where he provides implementation, consulting, and tuning advice on the Hadoop ecosystem to industry partners. Yuhao’s area of focus is distributed machine learning, especially large-scale analytical applications and infrastructure on Spark. He’s also an active contributor to Spark MLlib (50+ patches), has delivered the implementation of online LDA, QR decomposition, and several transformers of Spark feature engineering, and has provided improvements on some important algorithms.